Turbine

Published on June 2016 | Categories: Documents | Downloads: 78 | Comments: 0 | Views: 1973
of 386
Download PDF   Embed   Report

Comments

Content

more information - www.cambridge.org/9780521764056

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 emissions 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 practical issues that must be addressed in developing different types of gas turbines that emit pollutants at acceptable levels. 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, and received three patents. He chaired the Combustion and Fuels Committee of the International Gas Turbine Institute of the American Society of Mechanical Engineers (ASME). He is also on the Propellants and Combustion Technical Committee of the American Institute of Aeronautics and Astronautics (AIAA), and he previously served 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 of Energy, NASA, General Accounting Office, and Department of Defense. Lieuwen 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, 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 the ASME Westinghouse Silver Medal. Other recognitions include ASME best paper awards, the Sigma Xi Young Faculty Award, and the NSF CAREER award. 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, combustion 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 published 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 (2001–9) 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.

Cambridge Aerospace Series Editors: Wei Shyy and Vigor Yang 1. J. M. Rolfe and K. J. Staples (eds.): Flight Simulation 2. P. Berlin: The Geostationary Applications Satellite 3. M. J. T. Smith: Aircraft Noise 4. N. X. Vinh: Flight Mechanics of High-Performance Aircraft 5. W. A. Mair and D. L. Birdsall: Aircraft Performance 6. M. J. Abzug and E. E. Larrabee: Airplane Stability and Control 7. M. J. Sidi: Spacecraft Dynamics and Control 8. J. D. Anderson: A History of Aerodynamics 9. A. M. Cruise, J. A. Bowles, C. V. Goodall, and T. J. Patrick: Principles of Space Instrument Design 10. G. A. Khoury (ed.): Airship Technology, Second Edition 11. J. P. Fielding: Introduction to Aircraft Design 12. J. G. Leishman: Principles of Helicopter Aerodynamics, Second Edition 13. J. Katz and A. Plotkin: Low-Speed Aerodynamics, Second Edition 14. M. J. Abzug and E. E. Larrabee: Airplane Stability and Control: A History of the Technologies that Made Aviation Possible, Second Edition 15. D. H. Hodges and G. A. Pierce: Introduction to Structural Dynamics and Aeroelasticity, Second Edition 16. W. Fehse: Automatic Rendezvous and Docking of Spacecraft 17. R. D. Flack: Fundamentals of Jet Propulsion with Applications 18. E. A. Baskharone: Principles of Turbomachinery in Air-Breathing Engines 19. D. D. Knight: Numerical Methods for High-Speed Flows 20. C. A. Wagner, T. Hüttl, 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 22. W. Shyy, Y. Lian, H. Liu, J. Tang, and D. Viieru: Aerodynamics of Low Reynolds Number Flyers 23. J. H. Saleh: Analyses for Durability and System Design Lifetime 24. B. K. Donaldson: Analysis of Aircraft Structures, Second Edition 25. C. Segal: The Scramjet Engine: Processes and Characteristics 26. J. F. Doyle: Guided Explorations of the Mechanics of Solids and Structures 27. A. K. Kundu: Aircraft Design 28. M. I. Friswell, J. E. T. Penny, S. D. Garvey, and A. W. Lees: Dynamics of Rotating Machines 29. B. A. Conway (ed.): Spacecraft Trajectory Optimization 30. R. J. Adrian and J. Westerweel: Particle Image Velocimetry 31. G. A. Flandro, H. M. McMahon, and R. L. Roach: Basic Aerodynamics 32. H. Babinsky and J. K. Harvey: Shock Wave–Boundary-Layer Interactions 33. C. K. W. Tam: Computational Aeroacoustics: A Wave Number Approach 34. A. Filippone: Advanced Aircraft Flight Performance 35. I. Chopra and J. Sirohi: Smart Structures Theory 36. W. Johnson: Rotorcraft Aeromechanics 37. W. Shyy, H. Aono, C. K. Kang, and H. Liu: An Introduction to Flapping Wing Aerodynamics 38. T. C. Lieuwen and V. Yang (eds.): Gas Turbine Emissions

Gas Turbine Emissions
Edited by TIMOtHY C. LIEUWEN
Georgia Institute of Technology

VIgOR YANg
Georgia Institute of Technology

cambridge university press
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo, Delhi, Mexico City Cambridge University Press 32 Avenue of the Americas, New York, NY 10013-2473, USA www.cambridge.org Information on this title: www.cambridge.org/9780521764056 © Timothy C. Lieuwen and Vigor Yang 2013 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 2013 Printed in the United States of America A catalog record for this publication is available from the British Library. Library of Congress Cataloging in Publication data Lieuwen, Timothy C. Gas turbine emissions / Timothy C. Lieuwen, Vigor Yang. pagesâ•… cm. – (Cambridge aerospace series; 38) Includes bibliographical references and index. ISBN 978-0-521-76405-6 (hardback) 1.╇ Gas-turbines – Environmental aspects.â•… 2.╇ Gas-turbines – Combustion. 3.╇ Combustion gases – Environmental aspects.╇ I.╇ Yang, Vigor.â•… II.╇ Title. TJ778.L524â•… 2013 621.43′3–dc23â•…â•…â•… 2012051616 ISBN 978-0-521-76405-6 Hardback 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

List of Contributors Foreword by Alan H. Epstein Preface
Part 1  Overview and Key Issues

page ix xi xv

1 Aero Gas Turbine Combustion: Metrics, Constraints, and System Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Randal G. McKinney and James B. Hoke

2 Ground-Based Gas Turbine Combustion: Metrics, Constraints, and System Interactions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
Vincent McDonell and Manfred Klein

3 Overview of Worldwide Aircraft Regulatory Framework. . . . . . . . . . . . . . 81
Willard Dodds

4 Overview of Worldwide Ground-Based Regulatory Framework. . . . . . . . 95
Manfred Klein
Part 2  Fundamentals and Modeling: Production and Control

5 Particulate Formation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Meredith B. Colket III

6 Gaseous Aerosol Precursors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Richard C. Miake-Lye

7 NOx and CO Formation and Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Ponnuthurai Gokulakrishnan and Michael S. Klassen

8 Emissions from Oxyfueled or High-Exhaust Gas Recirculation Turbines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Alberto Amato, Jerry M. Seitzman, and Timothy C. Lieuwen
vii

viii

Contents
Part 3  Case Studies and Specific Technologies: Pollutant Trends and Key Drivers

9 Partially Premixed and Premixed Aero Engine Combustors. . . . . . . . . . 237
Christoph Hassa

10 Industrial Combustors: Conventional, Non-premixed, and Dry Low Emissions (DLN). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
Thomas Sattelmayer, Adnan Eroglu, Michael Koenig, Werner Krebs, and Geoff Myers

Index

363

Contributors

Alberto Amato, Georgia Institute of Technology, Atlanta, Georgia, U.S.A. Meredith B. Colket III, United Technologies Research Center, East Hartford, Connecticut, U.S.A. Willard Dodds, General Electric Aviation Company, Cincinnati, Ohio, U.S.A. Alan H. Epstein, Pratt & Whitney Company, East Hartford, Connecticut, U.S.A. Adnan Eroglu, Alstom Power, Inc., Baden, Switzerland Ponnuthurai Gokulakrishnan, Combustion Science & Engineering, Inc., Columbia, Maryland, U.S.A. Christoph Hassa, German Aerospace Center, DLR, Linder Hoehe, Cologne, Germany James B. Hoke, Pratt & Whitney Company, East Hartford, Connecticut, U.S.A. Michael S. Klassen, Combustion Science & Engineering, Inc., Columbia, Maryland, U.S.A. Manfred Klein, National Research Council, Ottawa, Ontario, Canada Michael Koenig, Siemens Energy Inc., Orlando, Florida, U.S.A. Werner Krebs, Siemens AG, Fossil Power Generation Division, Muelheim an der Ruhr, Germany Timothy C. Lieuwen, Georgia Institute of Technology, Atlanta, Georgia, U.S.A. Vincent McDonell, University of California, Irvine, California, U.S.A. Randal G. McKinney, Pratt & Whitney Company, East Hartford, Connecticut, U.S.A. Richard C. Miake-Lye, Aerodyne Research, Inc., Billerica, Massachusetts, U.S.A. Geoff Myers, GE Energy Company, Greenville, South Carolina, U.S.A. Thomas Sattelmayer, Technische Universität München, Garching, München, Germany Jerry M. Seitzman, Georgia Institute of Technology, Atlanta, Georgia, U.S.A.
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 concern regarding their emissions has grown even faster. Now, the electric power generation 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 marketing 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 independent, 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 effective 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 aircraft 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 increasing 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 conflicting 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 emissions have also become a major diplomatic challenge. Aviation is the most international 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 purview of more than one regulator. It is critical to the efficient operation of the world’s 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 worldwide 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 challenge generates business uncertainty and risk, with financial penalties large enough to destroy a business. Manufacturers of ground-based engines are often contractually 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 operated, 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 associated 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 challenge – 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 world’s 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 conversion 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 emissions, 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

and can be controlled by appropriate combustor designs. Changes to engine cycles and pressure ratio to increase fuel efficiency, however, generally promote the production of emissions such as NOx, and, thus, maintaining safe, reliable, low-emission aircraft engines is an increasingly important issue. The present volume compiles the key scientific and technological knowledge associated with gas turbine emissions into a single authoritative source. The book consists of three parts. The first part provides an overview of major issues relating to gas turbine combustion, including design approaches and constraints, at both the component and system levels, within the context of emissions. It also addresses approaches to meeting regulatory requirements. Important considerations for design optimization are discussed across all metrics of significance for gas turbine operation, including cost, safety, and reliability. The second part addresses fundamental issues associated with pollutant formation, characterization, modeling, and prediction. This part treats aerosol soot precursors, soot, NOx, and CO. In addition, it includes a chapter on emissions from gas turbines with significant levels of exhaust gas recirculation, or whose exhaust will be used for enhanced oil recovery or sequestered in geologic formations; in these cases, the emissions-related concerns are quite different. The third part of this book presents case studies from manufacturers and technology developers, emphasizing the system-level and practical issues that must be addressed in developing different types of gas turbines that emit pollutants at acceptable levels. It is our hope that this book will provide a valuable resource to workers in this field, as a foundation both for scientists researching various aspects of gas turbine emissions and for technology developers who translate this fundamental knowledge into products. This book would not have been possible without assistance from many individuals. Peter Gordon encouraged this project and supported us throughout. Our assistant Glenda Duncan was a tremendous help . . . a great help in the numerous tasks associated with preparing the text. We owe a great debt of gratitude to Jong-Chan Kim for his enormous effort in editing figures and ensuring that the illustrations are of the highest quality. Dilip Sundaram deserves special appreciation for indexing the book.

PART 1

OVERVIEW AND KEY ISSUES

1 Aero Gas Turbine Combustion: Metrics, Constraints, and System Interactions
Randal G. McKinney and James B. Hoke

1.1 Introduction
The aircraft gas turbine engine is a complex machine using advanced technology from many engineering disciplines such as aerodynamics, materials science, combustion, 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 industry’s capability to design combustors 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.

1.2  Overview of Selected Aircraft and Engine Requirements and their Relation to Combustor Requirements
Aircraft gas turbine engines have been used in many different sizes of aircraft since their introduction in the 1940s. Small aircraft such as single-engine turboprops use engines of low shaft horsepower, which are of small physical size. Business jets and smaller passenger aircraft may use turbojets or turbofans with thrust in the range of several thousand pounds, usually with two engines per aircraft. The other extreme includes four-engine aircraft with turbofan engine thrusts as high as seventy thousand pounds and very large twin-engine aircraft with thrust per engine in the one hundred thousand pound class. These thrust designs are also physically very large, with fan diameters over 100 inches. In all of these applications, the engine system imposes a common set of requirements upon the combustor, as summarized in Table 1.1.
3

4

Aero GT Combustion
Table 1.1.  Engine system-level requirements and supporting combustor characteristics
Engine requirement Optimize fuel consumption Meet emissions requirements Wide range of thrust Ground and altitude starting Turbine durability Overhaul and repair cost Combustor characteristic High combustion efficiency and low combustor pressure loss Minimize emissions and smoke Good combustion stability over entire operating range Easy to ignite and propagate flame Good combustor exit temperature distribution Meet required combustor life by managing metal temperatures and stresses

Stability

Durability

Emissions

Altitude relight and starting

Exit temperature

Figure 1.1.  Combustor performance requirements are interrelated.

As shown in Figure 1.1, these requirements are interdependent. Years of design and development within the industry have produced successive designs that improve upon all of the requirements concurrently. Although emissions are the focus of this text, each of these other requirements interacts with the emissions constraints and will be introduced briefly.

1.3  Combustor Effects on Engine Fuel Consumption
Gas turbine engines are Brayton cycle devices. An ideal version of such a cycle comprises isentropic compression, addition of heat at constant pressure, and isentropic expansion through the turbine. Figure 1.2 is a simplified schematic of the effect of such a cycle on the pressures and temperatures in the engine. In real engines, all of the processes incur some loss of performance versus the ideal, manifested as a stagnation pressure loss in the combustor. Combustion systems incur pressure losses because of flow diffusion and turning, jet mixing, and Rayleigh losses during heat addition (Lefebvre and Ballal, 2010). However, at most power conditions, the efficiency with which the fuel chemical energy is converted into thermal energy is very high, typically greater than 99.9 percent. “Low” levels of 98 to 99.5 percent can be seen at low-power levels. In general, though, the combustion system is a small parasitic effect on overall fuel consumption.

1.4  Fundamentals of Emissions Formation
Fan flow Core flow Fan Thrust

5

Core Power to operate fan + some thrust

Core flow Pressure Temperature Compressor

Combustor

Turbine Pressure Temperature

Figure 1.2.  Summary of component characteristics.

1.4  Fundamentals of Emissions Formation
The pollutants emitted by engines that are of most interest are carbon monoxide (CO), unburned hydrocarbons (UHC), nitric oxides (NOx), and particulate matter (PM or smoke). At low-power conditions, the inlet combustor pressure and temperature are relatively low, and reaction rates for kerosene-type fuels are low. Liquid fuel must be atomized, evaporated, and combusted, with sufficient residence time at high enough temperatures to convert the fuel into CO2. If the flow field permits fuel vapor to exit the combustor without any reaction, or, if partially reacted to species of lower molecular weights, UHC will be present. If a portion of the flow field subjects a reacting mixture to a premature decrease in temperature via mixing with cold airstreams, these incomplete or quenched reactions lead to the production of CO, as detailed in Chapter 7. At high power conditions, high air pressures and temperatures lead to fast reactions, with the result that CO and UHC are nearly zero. At these elevated temperatures, emissions of NOx and PM become more prevalent. NOx can be formed through several processes, but the dominant pathway is thermal NOx, as described by the extended Zeldovich mechanism, also detailed in Chapter 7. O2 = 2O N 2 +O= NO+N N + O2 = NO + O N + OH = NO+ H The formation rate is exponentially related to the temperature in the flame, peaking near stoichiometric conditions. Thermal NOx emissions can be reduced by limiting the time the flow spends at the high temperature and/or by reducing the maximum temperatures seen in the flame via stoichiometry control. Other NOx formation

6

Aero GT Combustion

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 temperatures, the formation of small particles of carbon can occur. These carbon particles 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, maximum NOx and PM increase with engine power and are typically maximized at ­ power. Chapters 5 and 7 discuss these emissions formation processes in more detail.

1.5 Effect of Range of Thrust and Starting Conditions on the Combustor
Flight gas turbine engines must provide a range of thrust and thrust response to power the aircraft mission. Missions vary depending on the aircraft application. Commercial aircraft and military transports have similar missions. Military fighters and other specialized aircraft can have very different missions because their use is not exclusively for the transport of payload between two points. Design requirements are also very different for commercial and military applications. Military fighter engines are often designed for maximized thrust developed per unit weight so that the maneuverability of the aircraft is maximized. Military fighter engines also fly at a wide range of thrust throughout the flight envelope and must undergo frequent rapid thrust transients. Typically, commercial engines are designed for maximum fuel efficiency per unit thrust. They fly at high altitude to achieve the best fuel efficiency and often do not have to endure the aggressive and numerous thrust transients of military fighter engines. Engine combustors must operate stably and efficiently over the full range of operating conditions, and must reliably relight if an engine shutdown or flameout should occur in flight. 1.5.1  Engine Mission Characteristics 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
60 50 EI or Smoke Number 40 30 20 10 0 0 20000 40000 60000 Thrust 80000 100000 HC CO NOx Smoke

7

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 altitude 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 occasionally 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 (−40ºF) 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
(a)

Aero GT Combustion

(b)

Figure 1.4.  (a) Can-annular combustor (Pratt & Whitney JT8D-200); (b) RQL annular combustor (IAE V2500).

and ignition. This requires favorable combustion fuel-air stoichiometry to meet requirements at all operating conditions. Two principal approaches have been used to achieve stoichiometry control in the industry. The first, fixed geometry without fuel staging, is the most common approach and is in the large majority of engines in service. These systems have all fuel injectors operating at all conditions. The second approach controls local fuel-air ratio through fuel staging. In these systems, not all fuel injectors operate at low power. This enables more active control of the local combustion fuel-air ratio. 1.5.2  Fixed-Geometry Rich-Quench-Lean (RQL) Combustors 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 (Figure 1.4b). 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 atomization 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 providing 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
Steady state fuel-air ratio

9

Combustor inlet condition

Temperature

Pressure

Transient decel fuel-air ratio Take-off Idle Thrust Take-off

Idle

Thrust

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 distribution 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 stoichiometric 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

Mix

ing

air

CO HC

io bust Com ear at n 1 Φ=

n

” hes enc “Qu ction rea

Gas temperature

CO consumed Threshold temperature Turbine inlet

Compressor exit Gas residence time in combustor

Figure 1.6.  Combustor at low power.

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 temperatures. 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 combustion 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

Rich n io bust m o c 2 Φ~

ion bust Com = 1 at Φ gen xy as o ded is ad

NOx Smoke

Rapid NOx formation Gas temperature Threshold temperature Compressor exit Gas residence time in combustor Turbine inlet

Figure 1.7.  Combustor at high power.

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 reduction of over 50 percent when compared to early annular combustors. They are also shorter and have lower volumes to reduce residence times. Reduced-length combustors 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 demonstrated 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 durability. 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 critical. 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 durability challenge. Because modern gas turbines operate at high temperatures and pressures, peak gas temperatures can exceed 4200ºF. Metallic liners have a practical temperature limit of <2000ºF for designs that meet typical durability life requirements. 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 temperature 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 therefore 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 hottest 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. 1.5.3  Fuel-Staged Combustors Having discussed RQL approaches, we next consider fuel-staged combustors, which have seen limited use in commercial aircraft service. First-generation designs were introduced in the 1990s, and updated designs are scheduled for release in future engines. The overall approach in a fuel-staged combustor is to control the combustion stoichiometry through use of fuel injection in multiple locations. Where the fixed-geometry RQL combustor injected fuel and air as uniformly as possible in the front end of the combustor, the staged combustor deliberately provides for multiple airflow and fuel flow zones. The objective is to achieve fuel-lean combustion conditions for NOx reduction at high power. The fuel-lean conditions keep gas temperatures low and virtually eliminate the highest temperatures associated with stoichiometric conditions that exist in the RQL design. The lack of fuel-rich and stoichiometric combustion creates two immediate benefits when compared to an RQL design. The first is that the fuel-lean flame produces very low levels of soot emissions. This means that carbon particulate emissions have the potential to be lower from fuel-staged combustors. Significant future efforts are required to characterize the full range of particulates emitted from both types of combustors (see Chapter 5). The second benefit is that the staged lean combustor requires less film cooling air for the liner. Because the lean reaction produces less soot, it is less luminous, resulting in reduced radiation heat load on the liner. Additionally, because the peak gas temperatures are lower, the convective heat loading is reduced. These factors allow for reduced liner cooling flux. This air can in turn be used for emissions control or to improve combustor exit temperature uniformity. In a fuel-staged combustor, a large amount of airflow is mixed with the fuel at the injection point, so that fuel-lean conditions are achieved at high power with all fuel injectors flowing. The large amount of airflow and fuel-lean conditions pose a stability challenge at low power due to the fuel-air ratio lapse that occurs between high power and low power. To mitigate the stability risk, some of the fuel injectors are turned off at low power. This allows for the control of the combustion stoichiometry at idle to ensure high combustion efficiency. The zone that operates at low power is referred to as the pilot zone, and the high-power fuel injectors are referred to as the main zone. A difficult challenge for staged combustor designs is the transition between operating with only the pilot at low power and all fuel injectors at high-power conditions. The transition often occurs at mid-power conditions such as approach thrust where fuel-air ratio, pressure, and temperature are not as high as

14

Aero GT Combustion

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 conditions. 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 (400ºF). This air is in contact with the main stage fuel injectors containing the stagnant fuel. To prevent fuel coking, cooling 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 fluctuations 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 acoustic damping as the combustor is modified for lean-staged operation (Lieuwen and Yang, 2005). 1.5.4  Ignition and Engine Starting 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 combustion 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 redundancy 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 significant 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 200ºF. 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 120ºF. 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 starting 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 0ºF are typical for an engine that windmills until cold. These conditions inhibit the atomization 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 ignition. Two types of fuel injectors are typically used: pressure atomizing and airblast

16
x103ft Successful ignition No ignition 30

Aero GT Combustion

40

Altitude

20

10

0

0

0.2

0.4

0.6 Mach number

0.8

1.0

1.2

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 atomization uses the energy of the airflow to produce the spray. The fuel is typically delivered 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 stabilization 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 Air Fuel

Airblast fuel injector

Secondary fuel

Primary fuel

Duplex fuel injector in air swirler

Figure 1.10.  Fuel injector types.

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 quantity 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 direction 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 ignition 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.

1.6  Turbine and Combustor Durability Considerations
The combustor has a significant impact on turbine durability and consequently impacts engine performance. The temperature distribution at the combustor exit affects the cooling airflow required to protect the airfoils and platforms in the turbines. This cooling airflow, in turn, reduces the engine performance by diverting flow from the mainstream so that it is not used to produce work. The cooling also causes mixing losses if it is introduced as low momentum film on the airfoils. Combustor film cooling is required for aero engines because the metallic liners are exposed to the high-temperature combustion process. Combustor cooling itself does not affect engine performance because it is added upstream of the turbines. However,

18

Aero GT Combustion

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 performance. 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: radial profile factor = (Tra − Te )/(Te − Ti ) 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: pattern factor = (Tstr − Te )/(Te − Ti ) 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 factor 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 temperatures 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 lowest 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 therefore 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
Maximum pattern factor Radial pattern factor Temperature factor

19

Radial profile factor

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 processes to achieve uniform exhaust temperatures are quite different. In swirling flow mixers, multiple airstreams are often used to create shear layers that promote mixing. 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 influencing 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 where the momentum ratio J is given by: J = (ρj U j2 )/(ρg 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 arrangement 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 average 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
P

h

Figure 1.12.  Jet mixing in a duct.

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 uniformity 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 determine 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 temperature above metal temperature in some areas of the combustor. The radiation flux is given by: q r = 0.5(1+ε w )(ε g Tg 4 − α g Tw 4 ) where εw is the wall emissivity, εg is the gaseous emissivity, αg is the gas absorptivity 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
Combustor liner history

21

Sheet metal liner Film Louver lips Double pass liner

Louver lips with backside cooling Floatwall liner Film

Figure 1.13.  Liner cooling designs.

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 critical 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 temperature is dependent on the local gas temperature and film effectiveness. Film effectiveness 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 (Figure 1.13). 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 operating 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) combustors 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 cooling technologies, the cooling flow allocation has not limited the achievement of profile 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 combustion temperature. Future aircraft engine cycles will require improved thermal and propulsive efficiency to meet aggressive fuel burn goals and address CO2 emission concerns. Current cycles have sufficient differential between the coolant and target metal temperatures 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.

1.7 Summary
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
REfERENcES

23

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: 31–70. 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 infrastructure 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, constraints, and much wider possible overall system interactions relative to the combustion 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 (“mechanical 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 components, 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, developments for advanced central plant designs couple the gas turbine with an integrated gasifier combined cycle (IGCC) in configurations that may also be designed to isolate 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 intermediate 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 consumption on the customer side of the meter is less than what the gas turbine produces, 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 optimization 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  Key Differentiators between Aero and Ground-based Gas Turbines
The need for advanced combustion technology in gas turbines is driven by a number of factors, including market need, regulatory pressure, performance, and reliability. The relative importance of these factors differs for stationary and aviation gas turbines. While regulatory pressures and associated emissions requirements are arguably a principal driver for stationary power generation gas turbines, emissions are a “secondary consideration” for aviation gas turbines owing to the more important operability and safety requirements. These drivers and constraints are discussed in this section.

26

Ground-Based Gas Turbine Combustion

2.2.1  Emissions In terms of regulatory pressure, legislation involving criteria pollutants continues to bring challenges to the gas turbine industry. While post engine treatment is capable of providing regulated levels of criteria pollutants (e.g., particulate, carbon monoxide, oxides of nitrogen), many regions, especially those with poor air quality, require increasingly stringent emission limits for operation (more details are provided in Section 2.4.8.1). It is usually preferred to avoid formation of pollutants in the combustion system rather than implementing post engine cleanup in order to circumvent the additional capital and maintenance costs of cleanup equipment. As a result, great interest in low NOx combustion systems exists as a means to minimize ground-level ozone. While worldwide pressure in this regard differs (e.g., developing nations such as China and India may not prioritize emissions requirements in the same way as other nations), the increasing recognition of the impact of pollutants on air quality and quality of life will likely increase the priority of pollutant mitigation in all regions. To this end, resources from sources such as The World Bank do have emission requirements, so even in developing regions seeking assistance, emissions is a driver. Regardless of regulatory pressures, those original equipment manufacturers (OEMs) that can offer the lowest emissions systems will have an edge ­ in markets with tight regulations but also in their image as environmentally friendly (i.e., “green”). In some regions, the ability to reduce pollutant emissions below prescribed limits may well translate directly into income in the form of emission credits. Several NOx, SOx, and CO2 trading markets exist and have proven successful in helping to reduce regional pollutant emission levels. Although a major motivation for using advanced combustion technology in gas turbines is generally associated with NOx emissions reduction, it is helpful to summarize the current emissions issues for gas turbines in general. Table 2.1 summarizes the current emissions drivers for aircraft and power generation gas turbines. 2.2.2  Operational Considerations Comparing operational considerations for aero and ground-based gas turbines reveals a number of key differences that impact the combustion system requirements. The most obvious difference is associated with the weight constraints. While aero engines must achieve high thrust-to-weight ratios, weight constraints are secondary for ground-based engines. As a result, the overall layout and architecture of the ground-based engine has more flexibility. This allows, for example, considerations for cycle enhancements such as reheat, recuperation, and/or intercooling (e.g., Kharchenko, 1998). These cycle enhancements can impact the conditions required for the combustion system. For example, recuperation will result in a significantly higher combustor inlet temperature, yet with relatively modest pressure ratios. In a reheat scenario, additional fuel may be added after partial expansion of hot gases from the combustor, utilizing a second combustor fed with a high-temperature, vitiated oxidant stream. The implications of

2.2  Key Differentiators between Aero and Ground-based Gas Turbines
Table 2.1.  Primary emissions drivers for gas turbine engines
Species Aircraft engine Landing/takeoff Soot HC (VOC, ROG, NMHC) CO NOx (NO, NO2) SOx (SO2, SO3, sulfates) CO2 H2O
✗ An issue ☒ High Priority

27

Power generation Cruise ☒ ☒ ☒ ☒ ☒ Distillate ✗ ✗ ✗ ☒ ✗ ☒ ✗ Gas ✗ ✗ ☒ ☒ ✗

✗ ✗ ✗ ☒ ☒ ✗

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., Alstom’s GT-24 and GT-26 reheat sequential combustion system, GE LMS-100 intercooled industrial engine; Solar Mercury 50 recuperated engine). ­ Ground-based systems also allow inclusion of water addition (e.g., inlet fogging 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 additional 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 involving water. Pratt’s effort involved the adaptation of an FT4000 turbine (EPRI, 1993 and ongoing work with the U.S. Department of Energy). Hitachi has carried out several 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 combustion 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 compositions. 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 accommodate market and political forces. The desire to take advantage of carbon mitigation 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 technology 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.

2.3  Gas Turbine–Grid Interaction
A key factor associated with ground-based turbines for power generation is their interaction with the electrical grid. Electricity cannot be stored conveniently; as a result, its generation must tightly align with its use. Sudden changes in demand must be met very rapidly. Gas turbines, owing to their ability to quickly ramp load, provide significant support to the grid in meeting this demand. In addition, gas turbines are a key source of base load power generation in most parts of the world. In such cases, utilities generally play a critical role in defining requirements, as they may own and operate the generation equipment or control “the grid” for conveying electricity and/or fuels. Because of the importance of energy and economy, utilities are often highly regulated to ensure fairness to the rate-paying customers and also to ensure an environment whereby the utility can mitigate economic risk associated with the significant investment required for new power generation or associated infrastructure, and still realize a reasonable return on investments. Clearly, the business of providing power is institutionalized and highly conservative in most cases. Within this context, the role of gas turbines and their interaction with the grid is evolving relatively rapidly. In all cases, whether large or small scale, the interconnection of the gas turbine with the local power distribution grid is critical. In general, gas turbines must be operated in coordination with other power generation sources (e.g., hydro, coal-fired boilers/steam turbines, nuclear power, etc.). As a result, the gas turbine needs to operate in harmony with the power demand from the grid in order to maintain the grid frequency. This results in a need for the engine to respond to a 50 percent load change in as little as a few seconds (e.g., Walsh and Fletcher, 2004). For large-scale, baseload-type operation, gas turbines offer perhaps the best source of flexible power. In other words, the inherent design and operation of a gas turbine lends it to adapting to changes in power output. As a result, gas turbines are a valuable “dispatchable” resource grid operators can look to for short-term or load-following power. Of course, when configured to maximize efficiency (e.g., when combined with other cycles such as steam turbines, or in conjunction with a gasifier), some flexibility in transient response is generally sacrificed. When operated in synchronous speed (i.e., a free-power turbine driving a generator at fixed speed), if a step increase in load is required, the power turbine speed initially drops. This requires the control system to increase fuel flow to generate additional gas turbine power until power generated and the load output equilibrate. If the grid load drops suddenly because of a grid-connected device failure, this can impact the engine dramatically

2.4  Plant-Level Requirements, Metrics, and Trade-offs

29

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 important 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 viability 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.

2.4  Plant-Level Requirements, Metrics, and Trade-offs
This section discusses various interactions between components and the overall power plant. Stationary gas turbine energy systems come in several arrangements and can be applied to most industrial and commercial sectors of the economy. Mechanical and electrical power, as well as thermal energy via capture of exhaust heat, is commonly produced from simple, combined cycle and cogeneration systems fueled by a variety of sources. The emissions from the plant can broadly be described as air pollutants and toxic emissions that affect regional health and ecosystems and global climate change through greenhouse gas emissions. In addition, air emissions must be balanced with

30

Ground-Based Gas Turbine Combustion

water and energy security considerations. Pollution prevention and energy conservation 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 creating 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 considerations 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 technology 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 manner. 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 concurrently 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 certain 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. 2.4.1  Trade-offs for Peaking Engine Applications 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 implementing air and fuel control strategies. Large gas turbines have lower NOx emission 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

Air Generator

Natural gas

Hot exhaust Gas turbine

Heat recovery steam generator

High pressure steam Transformer yard (to electric transmission company) Condensed steam (warm water) Warm moist air Low pressure steam Generator Steam turbine Hot water Cold water Condenser

Cooling tower

Figure 2.1.  Combined cycle plant (ENMAX, Calgary).

high-pressure ratios, NOx versus CO emissions divergence (discussed further in Section 2.5.3), bleed air losses in transient operation, and cold weather challenges when more power is available. 2.4.2  Trade-offs for Combined Cycle Plants 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 “greenfield” 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, underscoring 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, condensers, and off-gases. Large gas turbine systems can be built very quickly, and consume 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 impressive combustion technology advancements by the major frame engine OEMs. At the same time, the relatively smaller near-term economic benefit for low-emission system development for smaller systems makes implementation of more efficient, distributed systems more difficult. Moreover, large GTCC units, despite their advanced low-emissions combustion systems, must sometimes also have backend selective catalytic reduction (SCR) systems to meet ultralow NOx levels, resulting in an increase in the system’s 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. 2.4.3  Trade-offs for Repowering Applications 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 industry’s emissions of GHGs and other pollutants is sourced from these coal- and oil-fired plants, using boilers and condensing 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 techniques 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. 2.4.4  Trade-offs for Combined Heat and Power Combined heat and power (CHP) or cogeneration is the simultaneous production of electricity and thermal energy from the same fuel source in a process. Figure 2.3 illustrates the concept. Generally, 40 to 60 percent of all energy produced from traditional power generation boiler systems is lost as waste heat in condensers and in the exhaust stack. Using CHP concepts, the heat energy produced in generating power is recovered and used for industrial processes or for municipal district energy. Note that the world’s very first commercial power plant was built in 1882 in New York as

2.4  Plant-Level Requirements, Metrics, and Trade-offs
Steam turbine Steam LP Existing steam turbine IP

33

Cond. Gas turbine HRSG Gas fuel

Figure 2.2  Repowering utility boiler with a GT/HRSG system.

Fuel

MWe

Power generation technology

Heat recovery

Heating load

Absorption chiller

Cooling load

Figure 2.3.  Combined heat and power.

a cogeneration facility to produce industrial steam, with by-product electricity for local street lighting (Pearl St., NY, T. Edison). Small and medium-sized gas turbines (and other equipment) are well suited for onsite energy to match the thermal loads, but NOx emissions may be somewhat higher than from those of large engines. Plant sizing and location are critical elements in matching thermal and electrical outputs. CHP is one of the most important thermal energy technology concepts for energy conservation, with air pollution and GHG reductions. Waste heat utilization for absorption chilling can also help to reduce use of traditional refrigerants (e.g., CFCs) associated with ozone depletion, or can be used for GT power augmentation on hot days by cooling the inlet air. Energy security and process reliability can be achieved with onsite generation of both heat and power to avoid local power disruption effects and transmission losses. Like combined cycle plants, CHP systems may employ HRSG systems to recover high-value heat and, with duct burners, add flexibility to steam and power production. Duct burning will result in a decrease in the HRSG exhaust temperature, thereby improving system thermal efficiency. However, the duct burners used in the HRSG may produce additional NOx at the burner tip, depending on the air conditions at

34

Ground-Based Gas Turbine Combustion
Steam power E 35 Energy output (E) GT Combined Cycle Cogen E (45) H Auxiliary losses 15 48 Condenser, Stack loss GT CHP E (33) 85 H Energy output (E,H) 24 Auxiliary losses 10

65

Energy output (E,H)

14 Stack loss

Auxiliary losses

Figure 2.4.  Cycle efficiency trade-offs.

that section, illustrating another combustion technology where low-emissions strategies can play a role in pollutant prevention. As a result, a trade-off between higher pollutant formation rates and overall efficiency exists in terms of the absolute mass rate of emissions. One of the challenges of the cogeneration industry is to quantify the energy equivalence and value of these various energy products, especially the low-grade energy often wasted. Electricity generation generally has the highest value, whereas warm air or water may have little value, depending on the application. Heat to power (H:P) ratio is a design criterion related to system efficiency for two basic energy forms. Figure 2.4 illustrates three examples of energy systems that have various heat to electricity ratios (H:E), system efficiencies, and losses through condensers, stack, and auxiliaries. As a result, the ability to use CHP effectively is strongly dependent upon the application. In terms of combustion system impacts, CHP does not generally present major issues, with the one exception that many applications where CHP can be used effectively involve use of alternative fuels. As a result, fuel interchangeability in the design of these systems is an important consideration. 2.4.5  Trade-offs for District Energy District energy systems (DES) use captured heat from industrial processes to distribute 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.

2.4.6  Trade-offs for IGCC Applications 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 solution 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 hydrogen. Specific challenges for syngas combustion include (Lieuwen et al., 2010): • Flashback, autoignition, and explosion limits, flame sensing; • Combustor dynamics and vibration;

36
Air Air separation O2

Ground-Based Gas Turbine Combustion

N2 Water Water shift reaction

CO2

To pipeline

Gasifier

CO H2

Gas cleanup

H2

Gas turbine combined cycle

MW

Coal Petcoke Biomass

Slag

Sulphur

N2 injection, dilution

Figure 2.5.  Integrated gasification combined cycle (IGCC) system.

• Ability to have a combustor capable of operating with a range of fuels, including pure H2, CO/H2 syngas variations, methane for startup, and consideration for diluent addition; and • Operational flexibility, turndown, and transient ambient effects. 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. 2.4.7  Trade-offs for Pipeline Compressors 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 turbines, causing problematic shutdowns with stops and starts, increased methane emissions, 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 system, the high GWP poses issues. A common conclusion is that most fugitive CH4

2.4  Plant-Level Requirements, Metrics, and Trade-offs

37

23 MW

OTSG

6 MW HP steam 7 MW

26 MW

OTSG

20 MW steam turbine

26 MW Pipeline

OTSG

7 MW Air cooled condenser

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 station, often associated with unseated relief valve vents or unnecessary station blowdowns. 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 important tasks. Another challenge for these applications is limited need for heat. As a result, gas turbine electrical efficiency is of paramount importance. Thus, combustion systems may be impacted by need for higher firing temperatures or impacts associated 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). 2.4.8  Environmental Impacts 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.
2.4.8.1  BACT Considerations A definition of best available control technology (BACT) for pollutant emissions (such as NOx) may not be consistent with best practices with respect to other environmental issues (e.g., GHG emissions). Best practices and BACT 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, rather than backend controls and dilution, are encouraged. In determining a suite of clean energy choices, air pollution and GHG emissions from any energy or combustion system cannot occur individually, but always together as a system.

38
(a) 2.5 2 kg/MWhr 1.5 1 0.5 0 Coal (b) 1000 750 kg/MWhr 500 250 0 Coal

Ground-Based Gas Turbine Combustion

SO2

NOx

PM

Oil

Gas

GTCC

GTCHP

Bio

IGCC

Carbon Dioxide

Oil

Gas

GTCC

GTCHP

Bio

IGCC

Figure 2.7.  Comparison of emissions from various power plant configurations.

As illustrated in Figure 2.7, natural-gas-fired gas turbine cogeneration and CHP plants have a 60–75 percent GHG reduction from existing coal steam units, with emissions in the 220–300 kg/MWhr range. This is a result of switching to a lower carbon fuel (e.g., natural gas) and the effect of higher thermal efficiency, in the 50–80 percent range. Gas turbine units used in coal gasification can also have important reductions if CO2 can be captured and stored. Woodwaste and other biofuels also have fairly good overall emission characteristics. Effective consideration of GHG and efficiency issues requires basic knowledge of more than combustion and stack design, but also design and operation of major gas turbine engine components, the “demographics” of various unit types, as well as their associated equipment and the specific industrial plant applications. Best available control technology for various objectives can be based on: • System energy efficiency; • Pollution prevention of NOx and GHGs; • Collateral impacts (PM, ammonia, toxics, etc.); • Balancing of good combustion practices; • Dry low NOx combustion development; • Steam and water injection alternatives; • Understanding transient operating conditions; • Cogeneration unit sizing based on thermal loads; • Gasification potential of solid-derived fuels; and • Emissions monitoring and reporting.

2.4  Plant-Level Requirements, Metrics, and Trade-offs

39

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 cogeneration, district energy, gasification, methane leakage prevention, absorption chilling, 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 pollution, 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.
2.4.8.2  Air Pollution Reduction Technologies: Prevention and Control The details of pollutant formation chemistry in gas turbines are discussed in Chapter 7. In this section, some discussion regarding the approaches for mitigating them is provided along with implications for the overall gas turbine plant. Early efforts to mitigate NOx emissions relied on water injection into the combustor zone to lower the flame temperature for simple cycle units. The water-fuel mass ratio can range up to about 1.2 to 1, with a ratio of 1.0 achieving about 70–80 percent NOx reduction. Above a ratio of 1.1, CO emissions can climb significantly. Water injection may lower unit efficiency, plus contribute to pulsation and erosion in the combustion system, and must be carefully monitored in frequent inspections. Steam injection was commonly used in CHP and combined cycle applications before the commercial development of dry low NOx combustion. Steam injection into the combustor will increase mass flow and generate a 20 percent power output gain and a subsequent improvement in heat rate of up to 10 percent. For a given NOx reduction, steam-fuel mass ratio is about 50 percent higher than for water injection, but steam also has a less serious effect on component deterioration. The downside is the need for water management as discussed in Section 2.4.9. Selective catalytic reduction (SCR) is a backend clean up system where the exhaust gas stream in the heat recovery steam generator (HRSG) is sprayed with ammonia and sent through a catalyst bed in the HRSG. In a temperature range of 300–400°C, the ammonia in the presence of the catalyst reacts with the NOx in the exhaust to form nitrogen and water vapor. The SCR system needs the HRSG to reduce the 500–600°C exhaust to the required reaction temperature range. While practical for gas turbines in steady state combined cycle applications, cycling operations may present challenges in maintaining low ammonia slip. SCR is typically used after water/steam injection systems to reduce emissions from the 30–40 ppmv level to 5–10 ppmv. While effective for NOx removal, SCR systems may result in other issues such as:

• Ammonia “slip” from unreacted NH3 (less reaction during plant cycling operation);

40

Ground-Based Gas Turbine Combustion

Industrial Frame Combustion Systems - low pressure ratios (10–16) - simple designs, lots of mechanical space - early successes

Aero-Derived Systems - high pressure ratios (15–35) - small physical space, more complexity - improving reliability, but challenges remain

Figure 2.8.  Dry low NOx combustion systems (General Electric).

• The need to transport and handle hazardous ammonia; • Additional pressure drop inside the HRSG (flowpath restriction, additional length); • Emission of fine particulate as trace sulphur becomes ammonium bisulphate (ABS), which can also foul and corrode HRSG tubes at the low-temperature backend; • Higher HRSG exhaust temperatures (and lower efficiency) to promote ammonia reactions; and • possible N2O (another powerful GHG) emissions with catalytic systems. Alternatively, NOx reduction can be achieved by modifying the combustion process itself in DLN combustion systems, by rearranging the airflow and fuel mixture inside the combustor to minimize the occurrence of high local peak flame temperatures. More details regarding the advantages and disadvantages to this approach are discussed in Section 2.6.1, but, in brief, fuel is mixed with compressor discharge air to achieve a uniform mixture prior to entering the combustion zone. This “lean ­ premix” prevents the mixture from passing through a stoichiometric ratio in the combustor when mixing and combustion take place simultaneously. The fuel-air ratio is kept very lean to maintain low combustion temperatures and thereby minimize NOx formation, but this must be closely controlled during off-design conditions to prevent combustor pressure oscillations, flashback, and possible blowout. Figure 2.8 illustrates examples of lean premixed systems for minimization of NOx. Additional examples are shown in Section 2.7.2 and Chapter 10. Few collateral environmental impacts are associated with DLN systems. However, it is generally found that forcing lower NOx levels increases combustor operability concerns such as acoustic instabilities and blowoff and reduces the reliability and operational flexibility of the system. This is discussed in Section 2.6.1.4.

2.4  Plant-Level Requirements, Metrics, and Trade-offs

41

Combustor design is greatly influenced by how carbon monoxide emissions are handled, especially for off-design, transient, and cold ambient conditions. This raises the question of the importance of CO emissions. These emission levels are often set at the same ppm level as NOx emissions. However, CO levels are less harmful to human health and oxidize into CO2 within a day or two. Some of the operability and cost issues associated with DLN systems could be alleviated to some extent if CO emissions were de-emphasized relative to NOx and combustor operability. Fine particulate emissions (PM 10 and PM 2.5) are an important health issue and have gained increased attention in some regulatory initiatives. These emissions result from combustion of liquid fuels, from the use of ammonia-based SCR NOx controls, or, in smaller amounts, from heavier unprocessed natural gas or LNG containing some ethane, propane, and butane. Fine PM emissions from gas-fired gas turbines have been discussed (as per the US EPA AP42 rates of about 0.07 lb/MWhr). The discussion is confounded by a lack of precision in measurement of small levels combined with the presence of very fine airborne dust and VOC oils in the incoming air, of which a substantial portion bypasses the combustion system for cooling. Thus understanding the relative contribution (or possible reduction) by the air filtration system has many open questions.
2.4.8.3  Overall System Efficiency and GHG Emissions Discussions on risks of climate change have been ongoing since the 1970s, and, since the 1992 Rio Conference on Climate Change and Biodiversity, greenhouse gas emissions have become an important aspect of environmental performance. While debate remains, many energy choices for international “clean energy” implementation to minimize GHGs would be the same choices made based solely on economics. Thus choices that reduce air and water pollution tend to also increase energy efficiency and security. Most GHG emissions come from the same sources that produce high air pollution, underscoring a systems approach for energy and environmental design. An example of the simultaneous considerations in terms of fuel and overall architecture is shown in Figure 2.9.

2.4.9  Water Impacts 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 8–12°C 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
Industrial Blue Box Recycling Plastics, Rubber Sewage, Waste Treatment

Waste Heat L.P. Steam Hot Water

Woodwaste

MSW, Tires

Flyash, Gypsum Sulphur, Pet.Coke 5R’s

Energy Solutions – Synergies Cogeneration & District Energy

Waste Heat Recovery

By-Product Synergies

Sustainability Natural Gas & Hydrogen GHGs, CACs, Toxics, Water Energy Security & Conservation Energy Supply and Pricing Gas Turbines

Low Emission Combustion

CO2 Capture Polygeneration

Combined Cycles

Gasification & Biomass

Figure 2.9.  Considerations relative to integrated energy solutions.

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 sufficient 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 environment. 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 300–400 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 compression. 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 system components; increased engine maintenance; and fuel penalty. High costs can be incurred in isolated areas where the water acquisition and treatment costs are significant. Steam injection also requires extensive water treatment. During certain ambient 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  Engine-Level Metrics and Trade-offs
As mentioned previously and in Chapter 10, gas turbine engines are available for a wide range of applications. As is generally the case, trade-offs must be made when considering a specific application. Major considerations have traditionally been flexibility, efficiency, and emissions. 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 relative 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 reaction 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
1400 Air 1200 Mass flow of air (kg/hr) 1000 800 600 400 200 0 0 Fuel

Ground-Based Gas Turbine Combustion
140 120 100 80 60 40 20 0 20 Load (kW) 40 60 Mass flow of fuel (kg/hr) F/A * 100

F/A*100

Figure 2.10.  Example of total air and fuel flows for a 3.5:1 pressure ratio gas turbine engine.

Diffusion combustion

CO HC

SOOT T

[concentration]

NO Staged: Pilot + Main

Staged: Pilot

1.0 Idle Full φ

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 combustion 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 discussed in more detail in Section 2.7.3.

2.5  Engine-Level Metrics and Trade-offs

45

2.5.2  Transient Response Of increasing importance is transient response associated with gas turbines. Because of the significant penetration of intermittent renewable energy (e.g., solar and wind), balancing the grid load is becoming increasingly challenging. Gas turbines with fast transient response are well positioned to adapt the grid energy level with that of demand as described earlier. As a result, most major engine OEMs have been developing products with increasingly rapid response, including those configured for ­ combined cycle operation, traditionally viewed as only a baseload operation (Balling, 2010). The effect of these transient phenomena can directly impact the requirements of the combustion system. One might argue that these are similar to aero engines, which are ramped in load by design. Essentially, the worst problem would be associated with blowoff, where a sudden ramp down might result in overly lean conditions (e.g., Walsh and Fletcher, 2004). But since most low-emission systems today rely on very lean operation, they operate with less margin on blowoff compared to aero engines. Coupled with fuel flexibility requirements, the impact of transient response for ground-based engines can be significant in terms of operability. Coupled with increased need for transient response for power generation to offset intermittent sources is increased attention given to emissions levels at part load or startup. Regulators are increasingly cognizant that emissions may elevate during transient operation of the gas turbine. As a result, “peaking” operation has come under scrutiny as the actual time the engine is operating at full load steady state may become a smaller fraction of the overall time in which the engine is actually running. Engine developers have responded by creating combined cycle plants that can ramp rapidly while maintaining low emissions and good part load efficiencies (e.g., GE FlexEfficiency 50 Combined Cycle Power Plant; Alstom Next Generation KA2x/GT2x products). As a result, the notional “division of labor” of the prime movers shown in Figure 2.12 will evolve, which will influence the market shares of these devices. 2.5.3  Thermal Efficiency Gas turbine engine efficiency considerations remain a significant driver. With the adoption of output-based emissions standards as well as the need to consider emissions of greenhouse gases, overall cycle efficiency is becoming increasingly important. 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 temperatures. 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 1600°C. To achieve this temperature at the turbine inlet, little or no cooling is available for the combustor itself. Also, the 1600°C temperature itself is

46

Ground-Based Gas Turbine Combustion
Daily load profile (schematic) Peak load: pumped storage, SCPP / aero derivative / etc. Regulation load CCPP Product requirements: � Low electricity production costs � Short startup times � High starting reliability � Good part load behavior

Intermediate load: Predominantly CCPP

Electricity production

Renewables replace base load units because of must feed-in obligations, but must be backed up for wind / sun shortfall 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 aeroderivative and combined cycle plants) (from Balling, 2010).

Air

Fuel

Exhaust

Combustor Compressor HP turbine

Power 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 illustration of this cycle is shown in Figure 2.13. In this cycle, the compression ratio is essentially 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 recuperated 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
Exhaust

47

Recuperator

Air

Fuel

Combustor Compressor HP turbine Power turbine Generator

Figure 2.14.  Recuperated cycle.

Water in Air

Inter cooler

Exhaust

Recuperator

Fuel

LP

Combustor Power HP HP turbine turbine compressors

Generator

Figure 2.15.  Intercooled/recuperated cycle.

in relatively high combustor inlet temperatures compared to a simple cycle engine. Hence design trades may be needed to achieve low emissions because of constraints on premixer residence times and so forth. Another alternative that has also seen developmental success is the intercooled recuperated cycle (ICR). Examples of this cycle can be found in marine engines (e.g., Rolls Royce WR-21). Others have utilized just the intercooling element (e.g., GE LMS-100, GE LM6000 SPRINT). In the case of the SPRINT, water is also sprayed into the air at the intercooler stage to improve overall efficiency. The overall layout is shown in Figure 2.15. The intercooling increases the density of the air following the LP compression stage, which allows less compression work to be done in the HP stage, effectively increasing the cycle efficiency. With spray addition, more overall mass is expanded in the turbine versus the combined compression stage, hence more output work is done for the same amount of fuel consumed. In this case, the cycle modification again impacts the inlet conditions to the combustor. The use of water in the case of the SPRINT engine will likely impact the NOx chemistry. The intercooler itself will moderate temperature relative to the recuperated cycle alone. To illustrate the trades of these cycle conditions for a fixed turbine inlet temperature, Figure 2.16 presents overall theoretical efficiency and specific power as a function of cycle type. The merits of the ICR are evident as a similar package size (i.e., specific power) can realize significantly higher overall efficiency.

48
0.50

Ground-Based Gas Turbine Combustion

ICR 0.45 Overall efficiency 0.40 0.35 0.30 0.25 0.20 1.5x Simple cycle 2x 2.5x Specific power 3x 3.5x Recuperated Intercooled

Figure 2.16.  Comparison of idealize efficiency versus specific power for fixed turbine inlet temperature.

Economizer Exhaust Inter cooler After cooler Sat ura tor

Recuperator

Air

Fuel

HP LP compressors

Combustor

Power HP turbine turbine

Generator

Figure 2.17.  Humid air turbine cycle.

Further cycle adaptations include those involving water addition. As mentioned previously, one implementation of the intercooled cycle uses water to augment power output and likely help mitigate NOx formation. Other examples are shown in Figures 2.17 and 2.18. In Figure 2.17, the so-called Humid Air Turbine example is shown. It is obviously more complex relative to the other cycles, but has the attribute of potentially much higher overall efficiency. As discussed previously in Section 2.2.2, Hitachi has carried out several 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. A final twist on cycle concepts is shown in Figure 2.18. In this case, the combustor is replaced with a high-temperature fuel cell (e.g., solid oxide fuel cell  – SOFC) that produces DC power directly. With the enthalpy remaining in the fuel

2.6  Combustor-specific Metrics and Trade-offs
Exhaust

49

Recuperator H2O DC power

Fuel Air

SOFC Fuel

Combustor Compressor HP turbine

Power turbine

Generator

Figure 2.18.  Fuel cell/gas turbine hybrid cycle.

cell exhaust, additional turbine work can be accomplished, effectively making additional electricity from the waste heat. The fuel cell, being a thermochemical process, produces essentially no pollutants. However, to accommodate start up and potential transient modes, a combustor is generally included in the overall process.

2.6  Combustor-specific Metrics and Trade-offs
Operability and emissions are primary challenges associated with lean combustion. Since emissions is a principal driver motivating the use of lean combustion, the associated challenges involve achieving low emissions while maintaining stability, avoiding autoignition and flashback, and achieving sufficient turndown to cover the range of conditions needed to fulfill the operating map of the engine. In addition to the content herein, the reader is referred to other recent reviews of this subject (Richards et al., 2001; Lieuwen et al., 2008). While lean combustion strategies have evolved as the main approach for reducing emissions, it is apparent that low temperature (and associated low NOx formation rates) can be achieved under fuel-rich conditions as well as lean. This is overviewed in Chapter 7. Indeed, the ability of rich combustion strategies to mitigate operability concerns (e.g., stability limits, ­ oscillations) has seen it evolve as a major combustion strategy for aero engines as discussed in Chapter 1. In addition, in light of fuel flexibility, certain NO formation routes such as that from fuel-bound nitrogen can be overcome specifically using rich ­ strategies. But for ground-based turbines operating on the current fuel space of interest, lean strategies have evolved as the current state-of-the-art approach (McDonell, 2008). The challenges associated with lean combustion are illustrated in Figure 2.19 in the context of a typical combustor “stability loop.” For a given inlet pressure and temperature, the fuel-air ratio for a given mass flow through the combustor can be increased or decreased to a point where the combustor can no longer sustain the reaction. Limits in fuel-air ratio can be found on both rich and lean sides of stoichiometric at which the reaction will no longer be stable. This locus of conditions at which the reaction is no longer sustained is shown as the static stability limit in Figure 2.19. This loop may change as temperature, fuel composition, and pressure

50
0.03

Ground-Based Gas Turbine Combustion

Constant T3, P3 Constant fuel composition

0.02 Fuel/Air ratio Flash back Stable burning 0.01

High frequency oscillations

Static stability limit 0 0 0.25 0.50 Air mass flow (kg/s) 0.75 1.00

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 correspond to operating conditions where combustion oscillations are problematic. Each of these operability issues is discussed in the next sections. 2.6.1  Operability and Transient Combustion Phenomenon At the combustor level, operability can be severely impacted by blowoff, flashback, autoignition, and combustion dynamics. The key combustor operability concerns are summarized here.
2.6.1.1  Blowoff Blowoff refers to the dynamic process of flame detachment and extinction. In general, 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 Dämkohler 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 Dämkohler 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 composition 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 boundary, 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 Dämkholer (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): ST = SL + 5.3u′ SL (Liu and Lenze,1988 ) ST u′ = 1.52 SL SL (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



These expressions represent attempts to establish an equation that could be used for various mixtures and flame geometries; however, as stated before, discrepancies

52
100 90 80 70 ST,LD (m/s) 60 50 40 30 20 10 0 0

Ground-Based Gas Turbine Combustion

Damkohler, 1947 Liu and Lenze, 1988 Bradley, 1992 Littlejohn, 2008

10

20 u' (m/s)

30

40

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 sensitivity 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 consumption 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 definition 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 turbine 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 velocity lead to increased ST, and the addition of diluents does not have a marked effect on 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, α: gf ∝ SL α

2.6  Combustor-specific Metrics and Trade-offs
30

53

CH4, Eq.Ratio = 0.7 CH4, Eq. Ratio = 0.9 CH4, Eq. Ratio = 0.98 Littlejohn, 2008 Linear (Littlejohn, 2008)

25

20 Based on Local Displacement Speed ST/SL 15

10

Kido (2002). Based on Global Consumption Rate

5

0

0

5

10

15 u’/SL

20

25

30

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 quenching distance with respect to the laminar sublayer (Wohl, 1952; Schäfer 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 number of works (e.g., Kröner 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 compressor 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, condensed 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).



τ = 0.508e

( 3377 / T )

P

− 0.9



(2.1)

As indicated, all a designer must know temperature and pressure, which are functions of the cycle. This expression suggests that, at thirty bar and 700 K inlet temperature, 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 suggests 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
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
Engine Pres sure atm GE 9H * Solar Taurus 65 Solar Taurus 90 Solar Mercury 50 ** GE LM 6000 Siemens V-94.3A* Siemens V-94.2* Capstone C60 ** Alstom G24/26 EV Burner * Alstom G24/26 SEV Burner *** 23 15 12.3 9.9 35 17.7 12 4.2 30 15 Inlet Temp K 705 670 644 880 798 665 600 833 815 1300 τ (msec) CH4 2036 6205 13232 346 289 5835 38988 1477 264 1.65 τ (msec) C2H6 6213 33277 112220 123 251 34082 786278 859 188 0.07 τ (msec) C3H8 2421 11264 33876 82 134 11293 191174 506 105 0.104 τ (msec) H2 85 153 221 59 35 141 336 140 35 0.003

55

*  Inlet estimated from ideal gas isentropic compression **  Recuperated Engine Cycle ***  Reheat burner, used CHEMKIN and Galway Mechanism to calculate τ with φ = 0.6 τ  for alkanes calculated from correlations in Beerer et al. (2011) τ  for hydrogen calculated from correlations in Peschke and Spadaccini (1985)

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 ignition time. Hence the influence of hydrodynamic effects may be important in these systems.
2.6.1.4  Combustion Dynamics Combustion instabilities are periodic oscillations in the combustion chamber driven by interactions between unsteady heat release and acoustic waves (Lieuwen and Yang, 2005). Under certain operation conditions, these oscillations can achieve destructive levels and, therefore, effectively place constraints on where the system can be operated continuously. These instabilities can occur when unsteady pressure and heat release oscillations are in phase (Rayleigh, 1945), which leads to heat release disturbances pumping energy into the acoustic field. Heat release oscillations occur in combustion chambers because of the inherent sensitivity of the combustor system  – which can include the fuel delivery system, the flame itself, and inherent fluid mechanic instabilities – to imposed disturbances. To illustrate, consider two mechanisms particularly significant in premixed ­ systems: fuel-air ratio oscillations and vortex shedding (Ducruix et al., 2005; Zinn and

56

Ground-Based Gas Turbine Combustion

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 vorticity 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 passage 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 natural 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 stabilization 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 stagnation point of the vortex breakdown region.

2.7  Overview of Combustion Design Architectures
(a) (b)

57

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 favorable for promotion of CO oxidation. Of increasing importance is particulate emissions, 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.

2.7  Overview of Combustion Design Architectures
Depending on the type of cycle, application, and general requirements, the overall packaging of the combustion system can vary significantly. Generally speaking, compared to aero engine requirements, the compactness of the combustion systems for ground-based applications is not such a high priority. Some overview is provided here, but Chapter 10 examines more details regarding specific examples. 2.7.1  System Packaging For ground-based power, engines are classified to some extent based on their power output. At the smallest scale (e.g., microturbines  – 25–400 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
(a) 65 kW Low Emission CHP Package (b) 1000 kW Package

Figure 2.24.  Capstone Turbine Corporation Power Generation Packages (courtesy of Capstone Turbine Corporation).
5 stage Power Turbine (PT) 2 stage Intermediate Pressure Turbine (IPT) 2 stage High Pressure Turbine (HPT) From intercooler To intercooler Exhaust diffuser

Hot end drive shaft to generator Standard Annular Combustor (SAC) High Pressure Compressor (HPC) HPC inlet collector scroll case LPC exit diffuser scroll case

Low Pressure Compressor (LPC) First 6 stages of MS6001FA

Figure 2.25.  GE LMS-100 100 MW intercooled aeroderivative engine (Reale, 2004).

Aeroderivatives evolved from adaptation of aircraft engine gas turbines. Generally, aeroderivatives are more compact and have relatively high simple cycle efficiency (owing to the higher pressure ratio designs used in aero engines). As a result, aeroderivatives are good choices for peaking packages or for plants where combined cycle or waste heat recovery doesn’t make sense either practically or economically (Figure 2.25). 2.7.2  Combustor Layouts The specific combustor architecture falls into several categories. • Can Configuration • Can Annular • Annular

2.7  Overview of Combustion Design Architectures
(a) Overall Layout (b) Scroll

59

Figure 2.26.  GE-10 layout and scroll (Gas Turbine Short Course, 2002).
(a) Combustor Layout (b) Combustor and Transition Piece

Figure 2.27.  Allison 501 can-annular design (Courtesy of Rolls Royce).

The first major category is “can” combustors. As suggested, these systems are essentially 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 combustion 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 relative to overcoming combustion issues as the “domain” that needs to be addressed is relatively compact. Any interaction of the combustion system with primary and dilution 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 symmetry 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 maintainability 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 2.28.  GE 9H gas turbine engine.

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 combustor 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 2.30.  GE LM 6000 combustor. Courtesy by GE Energy and Siemens.
(a) VX4.3A (Annular) (b) SGT-8000H (Can)

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 system, 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 feature can-type combustors. Examples include Siemens, with an early advanced annular frame engine shown in Figure 2.31 along with its updated can counterpart. 2.7.3  Fuel Staging Approaches As discussed in Section 2.5.1, staging is often required to accomplish low emissions using an overall lean approach. In some cases, staging can be accomplished within an individual injector. This can be done with a combination of a “pilot” and “main” fuel circuit. The pilot circuit can be used to enrich the reaction sitting immediately downstream of the injector. Figure 2.32 illustrates an example of a piloted injector in

62

Ground-Based Gas Turbine Combustion

(a) Piloted Injector Cross Section

(b) Reaction Structure

Figure 2.32.  Example of staging for single injector (courtesy of Solar Turbines, Incorporated).

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
Outer casting Flow sleeve

63

Primary fuel nozzles Centerbody Lean and premixing primary zone Secondary zone Dilution zone

Secondary fuel nozzle

End cover

Venturi

Figure 2.33.  GE DLN 1 fuel injection cross-section.

A

B

Outer liner

Primary zone

Dilution air Dilution zone Dilution air Turbine inlet

A

B

Inner liner Recuperator wall

Outer liner

Inner liner Igniter

Injector

Figure 2.34.  Capstone C-60 combustor.

2.8.1  Liquid Fuels 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 attractive 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 savings 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, liquid 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 contaminants 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 distillates 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 turbines. 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 liquid also used in the transportation sector. Because of the latter, it is widely available. 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 hydrocarbon 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 methodologies 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
7.61 Benzene Aromatics 9.93 23+hydrocarbons > 2% of total composition

65

100 90 80 70 Relative abundance

6.54

50 4.17 4.53 40 30 20 10 0 0 10

11.04

16.82 18.63 21.31 20.71 24.96 25.42 29.02 32.89 36.59 40.19 43.45 48.66 49.71 52.66 T 20 30 40 Time (min) 50

3.04

14.15 16.31

60

Alkanes

Volatility

60

70

80

Figure 2.35.  MS-GC analysis of typical diesel fuel #2.

temper NOx formation. These developments have led to commercialized processes such as Orimulsion®, in which bitumen (a heavy residual from tars) and water are mixed to make a liquid fuel similar to a light distillate. In more recent years, the desire to establish liquid fuels from renewable sources has led to considerable work in the area of biofuel. The aviation sector has led much of this effort as the need for sustainable, high-energy-density fuels is critical. However, for power generation, the use of biofuels has also been examined (Demirbas, 2007; Knothe, 2010; Nigam and Singh, 2011). Of particular focus has been the derivation of liquids from various oil feedstocks. In contrast to distillate fuels, biodiesels have a much simpler molecular structure (comparing Figure 2.36 to Figure 2.35). In addition, the physical properties tend to vary from those of DF2 (especially the viscosity). Table 2.3 summarizes some of the key physical properties for a soy-based biodiesel, ethanol, and DF2. In summary, the use of liquid fuels for stationary gas turbines remains an important alternative to gaseous fuels, especially in regions of the world that (1) have copious crude oil reserves, (2) have no natural gas infrastructure, or (3) face natural gas curtailment. More information on liquid fuels for gas turbines is presented in Lefebvre and Ballal (2010). 2.8.2  Gaseous Fuels
2.8.2.1  Fossil-fuel-derived Natural Gas 2.8.2.1.1  Pipeline NatUral Gas.  “Standard” pipeline natural gas will have inherent variation depending on where it is extracted and its history in the natural gas

66

Ground-Based Gas Turbine Combustion
Table 2.3.  Comparison of physical properties of fuels
B99 Molecular mass μ σ ρ L LHV g/mol kg/m-s kg/s2 kg/m3 J/kg MJ/kg 291.6 0.0057 0.031 878 215243 37.4 E100 46.07 0.0014 0.022 779 855000 26.8 DF2 198.0 0.0020 0.027 807 254000 42.3

Source:  (Bolszo and McDonell, 2009). 49.88 50.12 44.35 Volatility T 0 10 20 30 40 Time (min) 50 60 70 80

100 90 80 Relative abundance 70 60 50 40 30 20 10 0 Primarily 4 fatty acid, methyl esters

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
Table 2.4.  Concentration of fuel constituents for natural gas in the United States
Mean Methane Ethane Propane C4 + CO2+ N2 Heating value Heating value Specific gravity Wobbe number Wobbe number Air/fuel ratio Air/fuel ratio Molecular weight Critical compression ratio Methane number Lower flammability limit Hydrogen:Carbon ratio * Without peakshaving Sources:  Liss et al., 1992 Mole % Mole % Mole % Mole % Mole % MJ/m3 Btu/scf MJ/m3 Btu/scf Mass Volume g/mol 93.9 3.2 0.7 0.4 2.6 38.46 1033 0.598 49.79 1336 16.4 9.7 17.3 13.8 90 5 3.92 Minimum * 74.5 0,5 0 0 0 36,14 970 0.563 44,76 1201 13.7 9,1 16.4 12.5 73.1 4.56 3.68 Maximum * 98.1 13.3 2.6 2,1 10 41,97 1127 0.698 52,85 1418 17.1 10.6 20.2 14.2 96.2 5.25 3.97 10th Percentile 89.6 1.5 0.2 0.1 1 37.48 1006 0.576 49.59 1331 15.9 9.4 16.7 13.4 84.9 4.84 3.82 90th Percentile 96.5 4.8 1.2 0.6 4.3 39.03 1048 0.623 50.55 1357 16.8 9.9 18 14 93.5 5,07 3,95

67

Volume %

production, LNG imports, inclusion of shale-derived natural gas, and/or other changes in the infrastructure; however, the results are still likely very relevant in describing the fuel variation. One extreme scenario not included in Table 2.4 is “peak shaving,” which involves the injection of propane-air mixtures into the pipeline to maintain the energy throughput of “natural gas” in times of peak demand. This practice is analogous to peak shaving for electricity, where, for economy and reliability, local generation is brought online to offset high rates or limited supplies and is illustrated in Figure 2.37 . In the case of peak shaving with gaseous fuels, the propane content of the pipeline natural gas can exceed 20 percent (Liss et al., 1992). This practice is most common in regions where natural gas flow may be limited during high use, such as in cold winter periods in the northeastern United States. With the advent of distributed generation systems and increased need for energy reliability, the notion of incorporating backup fuels is more important and may well contribute to the nature of fuel use in the future. In some regions, notably in the South Coast Air Basin of California, regulations on backup fuel require the use of synthetic diesel. However, many power generation devices (especially smaller output devices) do not have the capability to run “seamlessly” on the normal fuel (e.g., natural gas) and the backup (“dual fuel capability”). Of course, the use of diesel fuel even for backup generation has been identified as having a potentially significant air quality impact (Ryan et al., 2002), though some studies suggest the impact is perhaps 50 percent less than originally projected by the EPA. Regardless, alternative backup fuels such as propane would seemingly be a good strategy to consider in terms of (1) ease of dual fuel operability and (2) reduced pollutant emissions. Finally, it is worth noting that pipeline natural gas can also contain carbon dioxide and nitrogen, which can affect the position of the flame in a lean premixed

68

Ground-Based Gas Turbine Combustion

combustion system (CO2 can also affect the behavior of a catalyst in a catalytic ­ ombustion system). If CO2 is present, it can affect the combustion processes c 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 probably 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 natural gases, that is, those containing hydrogen sulfide or other organo-sulfur compounds, 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 maximum 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 (0.22–0.45 grams/m3). Of course, sulfurous compounds are added to natural gas as an odorant for safety purposes.
2.8.2.1.2 UnconVentional NatUral Gas. 

In addition to the changing composition of natural gases from conventional sources, an increase in use of natural gas obtained from what can be termed nonconventional sources is occurring. The largest contributor to the national natural gas system today is the gas obtained from coal seams. This “coalbed methane” can be obtained as virtually pure methane by predraining coal seams before mining occurs. This removal of methane before mining starts is considered an essential part of mine safety because it reduces the amount of methane released during the mining operations. Methane can also be extracted (often diluted with air) from “gobs” or collections of collapsed roof rubble found in mined-out areas. Similar methane and air mixtures can be extracted from coal seams during mining operations. In this type of extraction, the gas is removed via bore holes drilled horizontally into the coal seam just ahead of the active working face. These latter two gas types usually consist of methane and air mixed! The concentration of methane in these gas streams can be controlled, if desired, to levels close to 90 percent. Such gases often meet the minimum heating values for pipeline-quality natural gas and are usually well within all contaminant level requirements. These gases, however, add air to the natural mixtures in the pipelines, and this can result in the formation of sulfur-based acid gases or condensed phase (liquid) acids if mixed with other natural gases containing sulfur compounds. “Shale gas” has also gained attention as a potential sustaining source of natural gas. The viability of these extraction methods has resulted in a significant increase in the natural gas reserves. Shale gas is mainly methane, but it is apparent that considerable amounts of ethane (approaching 20% in some cases) can exist in these

2.8  Fuels
NGL fractionation plant (ethane, propane, butane) Underground LPG Storage Oil refinery LPG

69

NGLs

nat

ura

l ga

sp

ipe

line

Liquefied petroleum gas LPG storage

LPG

NG pipeline interconnects Underground NG storage
Liquefied natural gas LNG storage

LPG ship transport

LNG

LNG/LPG-air system

LNG ship transport
Liquefied natural gas LNG storage

LPG-air system peak shaving or base load

High pressure NG transmission to load centers LPG/LPG-air system base load Future / remote natural gas consumers

Firm natural gas consumers

LPG-air systems standby or base load

Interruptible natural gas consumers
1991–2006 STANDBY SYSTEMS, INC. All rights reserved.

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.
2.8.2.1.3  High LNG Import Scenarios.  As

mentioned previously, scenarios for supplementing the natural gas supply with imported liquefied natural gas (Figure 2.37)

70

Ground-Based Gas Turbine Combustion

Table 2.5.  Fuel composition from various sources around the world
Methane mol % Typical U.S. Known GT Experience Brunei Trinidad Algeria Indonesia Nigeria LNG Source Qatar Abu Dhabi Malaysia Australia Oman 95.7 89.6 89.76 96.14 87.83 90.18 90.53 89.27 85.96 87.64 86.41 86.61 Ethane mol % 3.2 S 4.75 3.4 8.61 6.41 5.05 7.07 12.57 6.88 9.04 8.31 Propane mol % 0.7 1.5 3.2 0.39 1.18 2.38 2.95 2.5 1.33 3.98 3.6 3.32 C4+ mol % 0.4 0.9 2.29 0.07 0.32 1.03 1.47 1.16 0.14 1.5 0.95 1.76 LHV Btu/scf 949 1006 1036 940 991 1010 1017 1018 1020 1045 1036 1051 HHV Btu/scf 1052 1112 1144 1041 1099 Wobbe Index 1379 1412 1429 1374 1405 1279 1283 1419 1420 1434 1429 1438

1126 1127 1155 1145 1161

Source:  Siemens-Westinghouse Power Corporation

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  Fossil-fuel-associated Gas 2.8.2.2.1 Associated Gas.  Associated with oil extraction is often the outgasing of gases typically comprised of hydrocarbons similar to those found in natural gas. Generally speaking, these gases have more higher hydrocarbons than typical natural gas. These “associated gases” (i.e., associated with oil extraction) can offer an opportunity for use as a fuel. In contrast, gas derived in gaseous form from gas wells is known as non-associated gas. The associated gases will likely appear similar to those found in LNG scenarios as illustrated in Table 2.5. Associated gases may be flared or vented in practice, which results in substantial GHG forcing in addition to the waste of a potential fuel. In North America, associated gas represents roughly 25 percent of the potential natural gas resource (Rojey et al., 1994).
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
Table 2.6.  Refinery off-gas composition
Component H2 CO CH4 N2 C2 C3 C4 C5 LHV (Btu/ft3) Vol % (dry basis) 25.7 1.5 37.4 2.9 27.5 2.9 1.7 0.4 984

71

pressure regarding flaring. The relative emissions performance of current systems used for consuming such gases can likely be improved if advanced combustion technology can be utilized.
2.8.2.3  Coke Oven Gas Steel making generates off-gases from the coke by-product. These are similar in composition to refinery off-gases, with the exception of even higher hydrogen versus methane content (~55 percent/~25 percent on a dry basis) (Tillman and Harding, 2004). 2.8.2.4  Renewable Methane-based Fuels 2.8.2.4.1 Landfill Gas.  When a landfill is capped, landfill gas (LFG) is generated as organic portions of the municipal solid wastes (MSW) decompose.1 Traditionally, landfill gas is not controlled and the expected period over which landfill gas will be produced may range from 50 to 100 years. However, a usable landfill gas production rate that can be utilized lasts for only ten to fifteen years. A bioreactor is a controlled landfill in which water and other nutrient sources are added into the MSW to increase the landfill gas production rate. The organic portions of the MSW in a landfill, including paper and paperboard, yard wastes, and food wastes, are decomposed through anaerobic biochemical reactions. The composition of the landfill gas varies with the characteristics of the waste, age of a landfill, weather conditions, and other variables. In general, landfill gas contains 50 percent methane (CH4), 45 percent carbon dioxide (CO2), and other traces of gas such as nitrogen (N2), oxygen (O2), hydrogen sulfite (H2S), and water vapor (e.g., Tillman and Harding, 2004). However, the gas composition varies with the nature of the organic material and with time. Indeed, variation in methane levels from 35 to 65 percent are common (Tillman and Harding, 2004). In addition, while the capping process seeks to eliminate air, leaks can lead to nitrogen and oxygen levels of up to 20 percent and 2.5 percent, respectively. Contaminants are also

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, regulatory pressures require a significant reduction in criteria emissions. in landfills, organic matter in wastewater streams contains potential fuel value. While in landfills, methane is generated by the 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 (30–40°C) or thermophilic AD (50–60°C) 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 5.3–6.3 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 estimated to exceed six million. European (EU) companies are world leaders in development of AD technology.
2

2.8.2.4.2 Wastewater Treatment.  As

is also possible to consider comingling of organic matter from waste streams (e.g., food waste from restaurants or personal residences) in anaerobic digesters. This could tremendously increase the feedstock available and thus maximize the production of alternative fuel. This same material could also be submitted to landfills, but the enhanced gas production by the digester provides a much shorter time frame for the conversion of waste to fuel.

2

2.8.2.4.3 Other Applications.  It

Adapted from California Energy Commission http://www.energy.ca.gov/research/renewable/ biomass/anaerobic_digestion/index.html.

2.8  Fuels

73

2.8.2.5  Synthesis Gas In recent years, strategies for clean use of the abundant domestic coal resources for low-emission energy production have gained attention. Examples include integrated gasification combined cycle (IGCC) systems. In these cases, processing raw fuel can produce gaseous high-hydrogen fuels for direct use in gas turbine engines (Stiegel and Ramezan, 2006). In addition to pure hydrogen, syngas, or synthesis gas, mixtures primarily composed of H2 and CO can also be obtained. Other names for syngas include producer gas, town gas, blue water gas, and coal gas, dependent on formation. Syngas first emerged in the early twentieth century for the production of methane, synthesis of ammonia, development of Fischer-Tropsch synthesis, and the hydroformylation of olefins (Wender, 1996). Hydrogen and syngas can be manufactured from any hydrogen feedstock using processes such as reformation, oxidation, gasification, and pyrolysis. When looking at syngas compositions, the primary focus is on the H2-CO ratio, which depends on the production method as well as the feedstock used. Steam reformation of methane produces syngas with a H2-CO ratio of approximately 3:1, whereas the composition of syngas from gasified coal has a ratio on the order of 1:1 (Spath and Dayton, 2003). The widespread applications of syngas include its use in the synthesis of other chemicals or fuels, as well as direct use as a fuel. The major commercial uses of syngas include the manufacturing of hydrogen, the manufacturing of methanol (especially for the synthesis of methyl t-butyl ether, MTBE), the synthesis of Fischer-Tropsch liquid fuels, and the hydroformylation (oxo) reaction of olefins, with the main uses of hydrogen being similar. The direct use of hydrogen and syngas as a fuel in a gas turbine combustor is also gaining recognition as a clean and viable source of power. The same is true for vehicles. While the widespread use of hydrogen begs the question of the infrastructure requirements, scenarios whereby hydrogen or high-hydrogen-content fuels are widely available are being seriously considered and appear to have a significant potential for reduced pollutant impacts (Stephens-Romero et al., 2009). If a foothold is gained in the transportation sector, it will likely serve as a springboard for energy generation using these fuels as well. Syngas can be derived from a variety of sources. First, natural gas is a large source. Any production that uses methane as a raw fuel produces syngas as an intermediate, and the production of syngas is the only reaction that breaks down CH4 into H2 and CO with a limited amount of unfavorable CO2 (Notari, 1991). Gasification of coal or petcoke is another syngas source. Coking technologies produce a solid, carbonaceous material from the processing of heavy and extra heavy oils called petroleum coke, or petcoke (Trommer et al., 2005). With increasing use of crude oil worldwide, a resulting increase of this by-product is inevitable, introducing another issue involving its disposal. The syngas composition is a function both of the original coal composition and the gasification approach. Biomass can also be converted into syngas with pyrolysis or gasification (Ni et al., 2006). Like coal, biomass can have a wide range of compositional variation. The gasification and pyrolysis processes are shown in detail in Figure 2.38.

74

Ground-Based Gas Turbine Combustion
Oxidant Gasification Agent Direct Gasification (oxygen controlled atmosphere) Carbon Based Material Oxygen-Free Gasification Agent Indirect Gasification (oxygen free atmosphere) Gas+Tar Char Gas+Tar Char

Heat Carbon Based Inert Gas or Nothing Pyrolysis (inert atmosphere) Gas+Tar Char

Heat

Figure 2.38.  Block diagram of the gasification and pyrolysis processes (adapted from Belgiorna et al., 2003).

In the process of direct gasification, the feedstock is partially oxidized using an oxidizing gasification agent, and the temperature is self-maintained through the reactions. Indirect gasification involves the use of an oxygen-free gasification agent, and an external energy source is needed to maintain the reaction temperature. Pyrolysis is a specific type of indirect gasification in which the gasification agent is either an inert gas or absent from the reaction altogether (Hauserman 1994).
2.8.2.5.1 Compositional Considerations.  Because of the numerous feedstock and processing methods used in obtaining syngas, the resultant gases will vary in composition. Syngas compositions vary not only with raw material (biomass, petroleum coke, coal, etc.), but also within each grouping, because of the composition of the actual feedstock. For example, successful uses of syngas-fired turbines have been demonstrated, and each facility operates a specific fuel composition dependent on the coal feed. Some syngas facilities have operated on fuels with hydrogen compositions of greater than 90 percent, although they have not attempted to implement advanced combustion technology for low emissions. IGCC projects of GE Energy include plants that operate on a wide range of compositions, as shown in Table 2.7. As shown in the table, many other constituents can exist in the syngas besides H2 and CO, including diluents such as CO2, N2, and H2O. Similarly, the syngas produced from biomass is strongly dependent on the type and composition, corresponding to another range of constituent concentrations. Because the composition of syngas varies so widely with production, the concept of determining a matrix engrossing all possible constituent ranges for the project was eliminated. Instead, fuel compositions were determined by likely future scenarios

2.8  Fuels
Table 2.7.  Minimum and maximum concentrations for syngas from GE power systems IGCC plants
Constituent H2 CO CO2 CH4 N2 + Ar H2O Minimum % 8.6 22.3 1.6 0.1 0.2 0 Maximum % 61.9 55.4 30.0 8.2% 49.3 39.8

75

Source:  Jones and Shilling, 2003

Table 2.8.  Dry, clean compositions (volumetric basis)
SOURCE Pure H2 Coal/Petcoke (O2 blown) Biomass (air blown) N2 Dilution H2 100 37 17 23 CO 0 46 17 31 CH4 0 1 5 1 CO2 0 14 13 10 N2 0 2 48 35 LHV (Btu/ft3) 265 247 142 165 Wobbe Index 1006 289 152 183

that involve the use of high-hydrogen fuels. One representative composition was selected for each of the following scenarios in addition to pure hydrogen: • Process and refinery gas; • Large-scale IGCC power plant (>50 MW); • Small-scale IGCC power plant (<50 MW); • Nitrogen dilution for NOx abatement. 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 percent 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 determines 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
Table 2.9.  Volume percent of species in typical fuels
Source High H2 Process and Refinery Gas Gasified Coal/Petcoke (O2 Blown) Gasified Biomass (air blown) & Gasification w/ N2 Dilution Landfill & Digester Gas LNG Shale Gas Associated Gases H2 90–100 25–55 35–40 15–25 CO 0–10 0–10 45–50 15–35 CH4 30–65 0–1 0–5 35–65 86–97 82–97 75–95 CO2 0–5 10–15 5–15 35–55 0–3 N2 C2 0–25 0–2 30–50 0–20 0–3 C3 0–25

2–11 0–14 5–25

0–2 0–1

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.

2.8.3  Water Injection In an effort to mitigate NOx formation, water may also be introduced. As discussed in Section 2.5.3, water could be injected into the working fluid. However, it can also be introduced into the combustion chamber. As shown in Chapter 10, incorporation of a water circuit into a gas system designed to operate on a backup fuel can lead to a fairly complex fuel injection system. In the case of a liquid fuel, a single circuit can also be used with water and liquid in form of an emulsion. In most cases, use of water poses additional considerations as discussed in Section 2.4.9.

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 environmental 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 pollutant 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 combustion 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): 051701–1: 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: 051507–1. 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: 101503–1–11. Belgiorno, V. et al. (2003). “Energy from Gasification of Solid Wastes.” Waste Management 23: 1–15. 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): 771–85. Bradley, D. (1992). “How Fast Can We Burn?” Symposium (International) on Combustion 24(1): 247–62. 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): 812–19. 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: 790–802. Collot, A. G. (2006). “Matching Gasification Technologies to Coal Properties.” International Journal of Coal Geology 65: 191–212. Dämkholer, G. (1947). “Influence of Turbulence on the Velocity of Flame in Gas Mixtures.” Z. Elektrochem 46: 601–26 (1950). English translation, NACA-TM-I 112. Demirbas, A. (2007). “Progress and Recent Trends in Biofuels.” Progress in Energy and Combustion Science 33(1): 1–17 . DeZubay, E. A. (1950). “Characteristics of Disk-controlled Flames.” Aero Digest, 54(6): 102–4. 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 179–212. 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: 011502–1–8.

78

Ground-Based Gas Turbine Combustion
EPRI (1993). A Feasibility and Assessment Study for FT4000 Humid Air Turbine (HAT), Report TR-102156. Fritz, J., Kröner, M., and Sattelmayer, T. (2004). “Flashback in a Swirl Burner with Cylindrical Premixing Zone.” Journal of Engineering for Gas Turbines and Power 126: 267–83. 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: 413–19. 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): 355–62. 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: 929–36. Knothe, G. (2010). “Biodiesel and Renewable Diesel: A Comparison.” Progress in Energy and Combustion Science 36(3): 364–73. 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. Kröner, 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: 693–700. 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): 729–32. Lewis, B., and Von Elbe, G. (1943). “Stability and Structure of Burner Flames.” Journal of Chemical Physics 11: 75–98.   (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: 471–80. 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): 011506–1 – 011506–10. Lieuwen, T. C., Yang, V., and Yetter, R., eds. (2010). Synthesis Gas Combustion: Fundamentals and Applications, CRC Press, FL.

References
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): 011502–011508. 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): 410–15. McDonell, V. (2008). “Lean Combustion in Gas Turbines,” in Lean Combustion – Technology and Control, Dunn-Rankin, D. ed., Academic Press, San Diego, pp. 121–160. 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 GT2009–59064, 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: 461–72. Nigam, P. S., and Singh, A. (2011). “Production of Liquid Biofuels from Renewable Sources.” Progress in Energy and Combustion Science, 37(1): 52–68. Notari, B. (1991). “C-1 Chemistry, A Critical Review.” Catalytic Science and Technology 1: 55–76. 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 2357–1, 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(1–2): 244–7 . 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  – GE’s 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. 473–83. 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: 141–69. 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: 1853–74. Ryan, N. E., Larsen, K. M., and Black, P. C. (2002). “Smaller, Closer, Dirtier, Diesel Backup Generators in California.” Environmental Defense Fund Report. Schäfer, 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: 520–30. 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: 98–120.

79

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-500–2006–110, February. Soares, C. (2008). Gas Turbines, A Handbook of Air, Land, and Sea Applications, Elsevier, London. Standby Systems, Inc. (2001–6). Propane Peak Shaving…an 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: 173–90. 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: 605–18. 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: 189–297 . Wohl, K. (1952) “Quenching, Flash-Back, Blow-Off Theory and Experiment.” 4th Symposium (Int.) on Combustion, pp. 69–89. Wright, F. H. (1959). “Bluff-body Stabilization: Blockage Effects.” Combustion and Flame 3: 319–37. 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: 630–6. 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.

3.1  Aero and Industrial Engines – Contrasting Requirements
Aero and ground-based engines can have many similarities. In fact, industrial engines derived from aero engines (aeroderivative engines) are used extensively in industrial service. However, technologies have been applied to industrial engines such that emissions from the lowest-emitting versions of the industrial engine are at least an order of magnitude lower than the original aero engine. These emissions reductions are achieved in a number of ways: • Natural gas fuel reduces conventional combustor NOx emissions by nearly 50 percent 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. The application of different emissions technologies is driven by a number of different factors, including local emissions impacts, geographical range of operation, fuel alternatives, weight and volume constraints, and transient operating requirements. 3.1.1  Emissions Impacts 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

Additionally, much smaller amounts of trace species, including oxides of nitro­ gen (NOx) formed at high temperature in the combustor, carbon monoxide and unburned hydrocarbons resulting from incomplete combustion of the fuel, and particulate matter (PM) produced by a range of mechanisms are emitted. • 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 particles that make up visible smoke) and volatile hydrocarbons and sulfur oxides that condense as the engine exhaust plume cools through mixing with the ambient air. 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: • “Primary” particulate matter, carbon monoxide, sulfur dioxide, and hydrocarbons classified as hazardous air pollutants are contained in the engine exhaust. • On a local scale, NO2 either in the engine exhaust or formed by oxidation of NO (typically with ozone) can contribute to elevated levels of NO2 near the runway. • On a regional scale, NOx and unburned hydrocarbons from a variety of sources, including aircraft and industrial engines, react to form ozone. • On a larger geographic scale, NOx and SOx react with other compounds (e.g., ammonia) to form “secondary” PM. 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, nitrogen dioxide, and sulfur dioxide. In the United States, the Environmental Protection Agency (EPA) determines the potential health impacts and sets NAAQS for pollutants 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 emissions. The study indicates that much of the NOx emitted at high altitude is transported 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 increasing 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. Aviation’s 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. 3.1.2  Flight Operations 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: • Provide uniform exit temperatures to ensure safe and durable turbine operation; • Have sufficient fuel system durability to avoid any chance of fuel leaks that could lead to fire; • Operate stably during fast accelerations, decelerations, and steady state conditions over a wide range of altitude and airspeed; • Light off and provide enough heat release to accelerate the engine during starts on the ground and at altitudes up to thirty thousand feet. 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 airworthiness certificates for engines and aircraft that have demonstrated that they meet all airworthiness requirements. From a combustor design standpoint, combustors designed to meet aircraft certification requirements for high-altitude air starts and fast transients have more stringent 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. 3.1.3  Geographical Range of Operation Local regulations may be justified for stationary ground-based engines, depending on the severity of local air quality problems. For example, it might be cost-effective to use exhaust gas treatment devices (e.g., selective catalytic reduction) to reduce stationary source emissions in a location with severe air quality issues. However, the cost and impact on engine efficiency (due to gas treatment system pressure loss) might outweigh the emissions benefits in other locations with better air quality. For an aircraft that flies into airports over a wide range of locations, it may make sense to average the cost-effectiveness for the airports served. Since aircraft fly internationally, it is more reasonable to consider cost-effectiveness on an international basis. 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, natural gas has advantages in that evaporation doesn’t 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. 3.1.5  Weight and Volume Experience with aeroderivative industrial engines (aircraft engines modified to meet industrial requirements) indicates that an aircraft engine combustor can meet basic

3.1  Aero and Industrial Engines – Contrasting Requirements

85

industrial operating engine requirements with relatively small fuel injector modifications 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 magnitude, 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 significantly 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): • Fuel staging is required because of the relatively narrow stability limits of a lean flame. Control systems have been developed to reliably handle the added complexity needed to meet aircraft requirements for altitude relight and fast transient operations. • Combustion dynamics (pressure fluctuations) are more likely in lean premixed combustors. Passive and active control strategies developed to manage combustion dynamics in industrial engines are being adapted to aircraft applications. • Wider operating limits and increased low-power operation are needed in aircraft applications. Combustors have been developed with “pilot” features specifically designed to extend low-power capability. • Slightly increased combustor volume is typically needed for complete combustion in lean flames. Short premixing combustors have been configured to fit into new aircraft engines with no impact on engine length and minimal impact on weight. • Challenges associated with prevaporizing and premixing jet fuels without autoignition have only recently been addressed for aircraft engines. In a land- or sea-based engine, thermal efficiency can be improved with intercooling, recuperative cycles and/or use of waste heat in steam cycles, so a high-efficiency engine can be designed with lower combustor operating temperatures. In a simple cycle aircraft engine, there has been a consistent push to higher engine pressure ratio and turbine inlet temperature. Both of these characteristics tend to increase NOx

86

Worldwide Aircraft Regulatory Framework

emissions with conventional technology, and make it more difficult to apply lean premising technology.

3.2  Regulatory Framework
3.2.1  United Nations International Civil Aviation Organization Aviation is an international industry, with individual aircraft likely to operate across national borders. Therefore, it became apparent early on that an international body would be the most effective way to handle many aviation issues. The International Civil Aviation Organization (ICAO), a United Nations (UN) organization, was established by the 1944 UN Convention on International Civil Aviation (also known simply as the Chicago Convention). One of the first steps for ICAO was to set recommended standards for engine and aircraft airworthiness. ICAO Annex 8 is a publication that describes international standards in this area. Through participation in ICAO, UN member states have largely adopted these standards for application by the national airworthiness authorities.
3.2.1.1  ICAO Committee on Aviation Environmental Protection Concerns with environmental issues increased during the 1960s and 1970s, as evidenced by the creation of the U.S. EPA in 1970 and the definition of initial ICAO Annex 16, Volume I aircraft noise standards in 1971. With its background in international airworthiness and noise standards, it was natural for ICAO to also take responsibility for international emissions standards. The ICAO Committee on Aircraft Engine Emissions was established, and initial aircraft engine emissions standards were introduced as ICAO Annex 16, Volume II in June 1981. In 1983, the ICAO Committees on Aircraft Engine Emissions (CAEE) and Aircraft Noise (CAN) combined to form the Committee on Aviation Environmental Protection (CAEP), a single committee to cover all aviation environmental issues. ICAO standards regulate aircraft engine emissions at low altitude. Limits have been set to control gaseous emissions of NOx, unburned hydrocarbons, and carbon monoxide, visible emissions of smoke, and fuel venting. Standards apply to aircraft engines with thrust greater than six thousand pounds. More stringent NOx standards have been adopted periodically, and are published in the current editions of Annex 16 Volume II. 3.2.1.2  ICAO Emissions Standards The current ICAO emissions standards are based on the ICAO landing-takeoff cycle (LTO Cycle): a ground test cycle intended to simulate aircraft operations below an altitude of three thousand feet above the runway, as shown schematically in Figure 3.1. The cycle uses four flight phases:

• Taxi-Idle: twenty-six minutes of operation at 7 percent of rated thrust; • Takeoff: 0.7 minutes at 100 percent rated thrust; • Climb: 2.2 minutes at 85 percent rated thrust; • Approach: four minutes at 30 percent rated thrust.

3.2  Regulatory Framework

87

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 conditions, 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 conducted as part of the airworthiness certification process, where a series of tests and analyses are conducted under the supervision of the national airworthiness authority (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 species, 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

Figure 3.2.  Example of an ICAO emissions data sheet.

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 chapter 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

to total aviation emissions; low-emissions technologies are more difficult to implement on a small scale; and past analyses indicate implementing emissions reduction technologies into small engines is not cost-effective. A summary of subsonic engine standards is shown in Table 3.1. Each standard includes maximum allowable levels for each species and specifies two applicability dates. The first is the date when the standard must be met by all new engine designs as part of the airworthiness certification process. The second date, known as a production cutoff, is the date when the standard becomes a requirement for all newly manufactured engines. Once an engine enters service, it is not required to meet any subsequent updates to the emissions standards. NOx has been the primary target of increased stringency over the years. CO and HC emission limits were set in the initial standards, and have not changed in subsequent revisions. The NOx standard has been reduced four times, such that the nominal limit for an engine pressure ratio of thirty has been reduced by 50 percent. As shown in Figure 3.3, the NOx limit (shown here for engines above 89 kN (20,000 lb)) thrust is a function of engine pressure ratio. CAEP has recognized that as new materials and cooling technologies have been introduced into new engines over the years, it has been possible to increase engine pressure to improve thermodynamic efficiency. Improved efficiency reduces fuel consumption and associated CO2 emissions. Temperatures within the combustor increase as engine pressure ratio goes up, resulting in more efficient combustion with associated reductions in CO and HC emissions. However, as CO2, CO, and HC emissions are reduced, NOx emissions increase because of higher rates of NOx formation with higher pressure and temperature in the combustor. Limits on smoke number were set to avoid visible jet engine exhaust plumes. The smoke particles in the plume obscure light, and the visibility of the plume depends on the concentration of soot particles in the plume and the path length of light through the plume. The smoke number is a measure of the concentration of particles in the plume, so for fixed smoke number, the visibility will be a function of the size of the plume, which will in turn be related to the size or thrust of the engine. With these relationships in mind, the limit on maximum smoke number was set as a function of engine thrust to ensure that the plume would not be visible over a typical range of viewing angles. A plot of smoke limit as a function of engine-rated thrust is shown in Figure 3.4. The LTO Cycle thrust levels and times in mode were defined to simulate operations of 1970s vintage jet aircraft, and are not necessarily representative of current engines and operating conditions. Today’s aircraft often take off at reduced thrust (depending on runway length, elevation, and ambient ­ temperature), and performance has improved such that they climb faster than their 1970s counterparts, but the LTO Cycle has been kept constant as a measure of relative performance. Methods have been developed to use the certification data to estimate emissions during actual operations (SAE International, “Procedure for the Calculation of Aircraft Emissions,” SAE AIR 5715, July 7, 2009), and have been integrated into analytical tools such as the FAA Aviation Environmental Design Tool (AEDT).

90 Table 3.1.  Summary of ICAO/CAEP emissions standards
Standard Applicability Certification CAEP/1* CAEP/2 CAEP/4 “ “ “ “ CAEP/6 “ “ “ “ CAEP/8 “ “ “ “ 1986 1996 2004 “ “ “ “ 2008 “ “ “ “ 2014 “ “ “ “ Production 1986 2000 2013 “ “ “ “ 2013 “ “ “ “ 2018? “ “ “ “ Engine characteristics EPR All All 30–62.5 <30 30–62.5 <30 >62.5 30–82.6 <30 30–82.6 <30 >82.6 30–104.7 <30 30–104.7 <30 >104.7 Foo, kN >26.7 >26.7 >89.0 >89.0 26.7–89.0 26.7–89.0 >26.7 >89.0 >89.0 26.7–89.0 26.7–89.0 >26.7 >89.0 >89.0 26.7–89.0 26.7–89.0 >26.7 Emission limits Smoke, SN 83.6*(Foo)–0.274 “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ CO, g/kN 118 “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ HC, g/kN 19.6 “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ “ NOx, g/kN 40+2*EPR 32+1.6*EPR 7+2*EPR 19+1.6*EPR42.71+1.4286*EPR0.4013*Foo+0.00642*EPR*Foo 37 .527+1.6 *EPR-0.2087*Foo32+1.6*EPR –1.04+2*EPR 16.72+1.4080*EPR 46.1600 +1.4286*EPR-0.5303*Foo+0.00642*EPR *Foo 38.5486+1.6823*EPR-0.2453*Foo.00308*EPR*Foo 32+1.6*EPR –9.88+2*EPR 7 .88+1.4080*EPR 41.9435+1. 505*EPR-0.5823*Foo+0.005562*EPR*Foo 40.052+1.5681*EPR-0.3615*Foo-0.0018*E PR*Foo 32+1.6*EPR
EPR = Engine Pressure Ratio Foo = Rated thrust at ISA sea level static conditions * Actually second meeting of Committee on Aircraft Engine Emissions (CAEE/2)

3.2  Regulatory Framework
160 140 ICAO LTO Cycle NOx limit, g/kN 120 100 80 60 40 20 0 15 CAEP/1 CAEP/2 CAEP/4 CAEP/6 CAEP/8 20 25 30 35 40 45 50 55

91

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 Engine rated thrust, kN

500

600

Figure 3.4.  ICAO/CAEP smoke limit.

3.2.1.3  Emerging Emissions Issues In the emissions area, CAEP has focused on low-altitude NOx for many years. However, the focus is expanding to give higher priority to CO2 emissions, particulate matter, and, to a lesser extent, high-altitude emissions. In response to global emphasis on climate change, at its eighth meeting in 2010, CAEP initiated an aggressive program to establish standards for aircraft CO2 emissions. The objective of this program was to develop metrics, certification methodologies, and emissions limits in time for the CAEP/9 meeting in 2013. CAEP is also working with the SAE E-31 Aircraft Exhaust Emissions Measurement Committee to develop a certification procedure by 2013 for a future

92

Worldwide Aircraft Regulatory Framework

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. 3.2.2  National and Local Emissions Policies
3.2.2.1  Adoption of CAEP Emissions Standards by National Authorities ICAO sets emissions standards for aircraft engines, but the individual member states must adopt the standards into national regulations before they can be enforced. The process is not straightforward in some countries and can delay local implementation. For example, historically, U.S. emissions standards for aircraft engines have first been set by the EPA under 40 C.F.R. Part 87 (Title 40 of the Code of Federal Regulations, Chapter I – Environmental Protection Agency, Part 87, Control of Air Pollution from Aircraft and Aircraft Engines). The FAA then executes the rules under 14 C.F.R. Part 34. This means that once the CAEP standard has been established and published by ICAO, the EPA normally prepares and issues a proposed rule and then, after a review period, issues the final rule. The FAA then needs to amend the engine emissions certification rule. As a result of this process, U.S. emissions requirements have not always harmonized with those of the rest of the world. For example:

• 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, FAA’s 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.
3.2.2.2  The U.S. National Environmental Policy Act (NEPA) ICAO emissions standards only apply to aircraft engines for civil aviation. Military engines are exempt, as are airframe manufacturers, airports, and air traffic control providers. However, under NEPA, U.S. federal agencies are required to consider the environmental impacts of their proposed actions and ways to reduce or offset the environmental impacts of those actions. NEPA requires environmental impact statements (EIS) to be prepared for military projects such as replacement of older

3.3  Future Outlook

93

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 combustor 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 generation facilities, road traffic, or ground support equipment.
3.2.2.3  European NOx Landing Charges Landing charges based on consideration of NOx and HC emissions, as published in the ICAO data bank, were first implemented in Sweden and Switzerland in 1997. Initially, two different systems were used that placed each aircraft/engine configuration into one of several charging classes based on its emissions levels. Landing fees were increased for aircraft with highest emissions. The charges generally promoted application of low-emissions combustors, but the use of two different systems sent a mixed message to manufacturers. In late 2000, the European Civil Aviation Conference (ECAC) formed a subgroup to conduct an Emission Related Landing charges Investigation Group (ERLIG). Based on the ERLIG’s work, ECAC Recommendation 27–4, NOx Emission Classification Scheme for Aircraft, was issued in 2003. The basic principle of this recommendation is that for all aircraft that meet the ICAO HC standard, charges should be proportional to the amount of NOx emitted per the ICAO Emissions Data Bank. This principle recognizes that environmental impact is approximately proportional to emissions, so an aircraft will be charged in proportion to the impact it causes. Airport emissions charges are now applied at airports in Switzerland, Sweden, and several other European countries. 3.2.2.4  European Union Emission Trading Scheme (ETS) Trading of CO2 emissions allowances under the European Union ETS was initiated in January 2005 as a means to provide incentive to reduce greenhouse gas emissions. Aviation was included in the ETS starting in 2012. Currently, the ETS focuses on CO2. However, considering the unique impacts of non-CO2 emissions emitted at high altitude, there have been proposals to require aircraft to buy additional credits to account for NOx emissions. For now, specific action on high-altitude NOx emissions is on hold, awaiting development of better scientific understanding of NOx impacts on climate.

3.3  Future Outlook
Aviation only produces a few percent of total emissions affecting air quality and climate, but its impact is expected to grow: • 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

• Absolute emissions are expected to increase because forecasted traffic growth will likely outpace improvements in emissions reduction technology. • Cruise emissions are not currently accounted for in emissions standards, and are likely to get more attention as scientific understanding of the impact of high-altitude emissions improves. ICAO is implementing emissions standards for CO2 and nonvolatile particulate matter 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 aviation 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: 7736–42. 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 2012–0936. 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: 1730–6. 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: 542–6. 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-2009–002, 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(10–11): 944–55. 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: 91–100, January.

4 Overview of Worldwide Ground-Based Regulatory Framework
Manfred Klein

Pollution prevention and energy conservation with system efficiency are key elements 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 appropriately 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 reliable 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 prevention planning and environmental assessments may require a more comprehensive 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.

4.1  Regional and Global Atmospheric Issues
Our regional and international concerns over atmospheric stress are in large part due to the way we produce and use energy – electrical, mechanical power, and various types of heat and cooling energy. The issues arise both from the health impacts of traditional air pollution and from the evidence suggesting a human influence on global climate change due to greenhouse gas emissions. For several decades, it has been well recognized that fuel combustion leading to air pollution from acid gases, particulates, and trace elements has had a major impact on human health and on the ecosystem. This has been a primary driver for national activities such as the acid rain programs and various smog reduction initiatives across Europe, North America, and other regions. Nitrogen oxides, SO2, and fine particulates are formed from fuel oil combustion in gas turbine engines, while natural gas-fueled gas turbines produce mainly NOx and some CO emissions. These types of emissions are often produced
95

96

Worldwide Ground-Based Regulatory Framework
Air emissions (Smog, Acid rain, Climate change, Toxics) GHGs • • • • Carbon dioxide CO2 Methane CH4 Nitrous oxide N2O SF6 et al. Ozone depletion • CFCs • • • • Air pollution Sulphur dioxide SO2 Nitrogen oxides NO2 Volatile organics VOC Carbon monoxide CO

• Fine particulate PM • Mercury & heavy metals • Ammonia

Figure 4.1.  Various types of 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 world’s 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 cannot 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 various 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 industrial 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 reproduction 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).

4.2  Air Pollution and Greenhouse Gas Emissions from Gas Turbine Systems
The following is a summary of various types of criteria air contaminant (CAC) emissions from gas turbines fueled by liquid fuels such as light distillates or kerosene, natural gas, or synthetic gas from coal gasification.
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 combustor 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 (~ 0.2–0.5%), or in natural gas (4–6 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 temperatures 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

in Sections 2.2 and 2.4, NOx is produced by the high-temperature (1800–2000K) oxidation of nitrogen contained in the very large quantities of air swallowed by the engine for its power production. Additional NOx is formed from liquid fuels such as distillate oil, which have higher local flame temperatures and some entrained nitrogen compounds. Combustion systems must therefore be designed to optimize firing temperature (i.e., maximize their own CO2 emissions) for the particular fuel in order to minimize these air pollutants and impurities. Over most of the operating range for combustors, CO emissions will rise as NOx emissions decline. This represents one of the difficult challenges in developing low-emission combustion systems for full-power range operation, reliability, and good thermal efficiency with low GHG emissions. Dry low NOx (DLN) combustion systems readily reduce NOx emissions from 3–4 kg/MWhr (150–300 ppm) down to about 0.5 kg/MWhr (20–30 ppm), and lower on large gas turbine engines. When DLN combustion and high plant efficiency are used for prevention, air pollution is reduced by 90–95 percent from existing steam systems using coal or oil (Figure 2.7). It should be noted that these types of modern efficient plants with dry low NOx combustion rarely cause a significant air pollution problem. Because of practical “balancing” issues in combustion and system design, it is very difficult to have strict, ultralow limits for air toxics, criteria air contaminants, and greenhouse gases at the same time. Very low NOx ppm levels may tend to increase air toxics and also encourage larger, more inefficient plants. A common question is whether CO emissions should be controlled down to the same concentration level as NOx in full- and part-load DLN combustion. Carbon monoxide from industrial stacks is not as serious an emission as NOx, and the requirement to have very low CO emissions during short-term off-design conditions can greatly compromise overall DLN combustor design and operating reliability (Klein, 1999). From the very same energy systems that produce the air pollutants listed earlier, useful energy is created from conversion of carbon into CO2. Carbon dioxide is the most common greenhouse gas, and although not usually considered a pollutant, it has a serious environmental impact on climate change. In assessing such matters as fuel combustion choices, it may be appropriate that air emission reductions of NOx, SO2, CO2, methane, mercury, ammonia, and particulate emissions be considered together, because they all occur at the same time in a given system and are highly interrelated. Air pollution cannot be produced in a system without making carbon dioxide. CO2 emission rates are determined by the fuel carbon content and the overall system efficiency. Fuels with high hydrogen content such as natural gas and synthetic gas, used in high-efficiency applications, represent the best way to reduce GHG emissions from thermal energy systems. Using the factors listed later in this section, one can easily estimate and compare the CO2 emission rates of various fossil-fueled plants, once the overall heat rate or efficiency (GJ/MWhr) has been determined. Combined heat and power systems have the best heat rates, in the 4–6 GJ/MWhr range. Typical CO2 factors (in kg/GJ) as shown later can be multiplied by net heat rates in GJ/MWhr (Figure 4.2).

4.2  Air Pollution and Greenhouse Gas Emissions from Gas Turbine Systems
1000 750 kg/MWhr 500 250 0 Coal Oil Gas GTCC GTCHP Bio IGCC Coal, Oil and Gas – steam boiler rankine cycles GTCC – gas turbine combined cycle GTCHP – gas turbine combined heat & power Bio – woodwaste biomass IGCC – coal gasification combined cycle CCS – carbon capture & storage Carbon dioxide

99

CCS

Figure 4.2.  Comparison of CO2 emissions from various energy-generating plants.

Because of its high hydrogen content, natural gas integrated systems can have a CO2 rate of 220–360 kg/MWhr. This represents a 60–80 percent net GHG reduction from current coal technology, with 90–95 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 cogeneration 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 processes, will play a vital role for energy security and provide economic and environmental 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

4.3  General Policy Considerations
Increasingly, energy and fuel choices may have to be judged by total emissions profile, water impacts, and environmental effects such as noise and visibility. Emission trading schemes, regulatory programs, publications, and books often describe potential solutions to either of these problems as individual concepts. There have long been international policies and regulations governing the installation of gas-turbine-based energy equipment, based mostly on direct NOx and CO combustion emissions. Permitting can be based on lowest NOx and CO emission rates for a particular unit, or it can alternatively be based on optimizing the whole plant within the fence line considering fuel choices, GHG profiles, and system design efficiencies. A case can be made for developing an integrated approach to these issues, through clear and comprehensive multi-pollutant, output-based emission policies for many combustion sources, including gas turbines. Environmental assessments that adopt emission regulations could choose to investigate linkages among these objectives for developing national strategies. Critical elements of the term clean energy are related to four main broad topics: human health, climate change, energy security/reliability, and other environmental issues (such as land use and water impacts). Emissions prevention is a key factor in making cost-effective long-term choices, often more effective than cleaning up existing emissions with back-end controls. Both technology and regulatory policy can enhance these opportunities, especially when there is a clear and balanced approach to the overall objectives, including elements of energy reliability and security (U.S. EPA, 2006).

4.4  Developing Emission Criteria and Standards
Most international energy and environment reports indicate that various types of gas-turbine-based facilities, fueled by natural gas or by synthetic gases from coal and petcoke, will provide very substantial benefits to large GHG reductions, improved energy security, and low air pollution (IEA, 2009). More efficient units with higher pressure ratios have difficulty meeting ultralow ppm-concentration-based standards for NOx and CO, especially under transient and cycling conditions. The intent here is to look at air emission rules and explore whether a different, more balanced approach to emissions standards would enhance engine reliability, system efficiency, and emission prevention effectiveness for a wider application of clean energy facilities. Air pollution measurements are done on a volumetric concentration basis, in parts per million by volume (ppmv or ppm) with continuous or periodic sampling equipment. Thus regulations have usually developed as ppm limits, and in some cases in a weight/volume fraction such as mg/m3, as in Table 4.1. This may also be because health concerns have often been associated with concentrated ambient or receptor emissions. Sometimes rules are based on a pollutant mass per unit heat input, such as lbs/MMBTU, grams/GJ, or ng/Joule.

4.4  Developing Emission Criteria and Standards
Table 4.1.  Standards with various emissions criteria
Examples of International Emission Standards, 2005 (for GT Units Larger than 10 MWe, gas fuel) United States United Kingdom Germany France Japan Canada EU LCPD World Bank Australia 2–42 ppm 60 mg/m3 75 mg/m3 50 mg/m3 15–70 ppm 140 g/GJout * 50–75 mg/m3 * 125 mg/m3 70 mg/m3

101

*  Facility Cogeneration Incentives

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 concentrations at the stack are usually corrected to 15 percent O2, but measured at a slightly different O2 level (13–16 percent) depending on the engine efficiency. Some source- and ambient-concentration-based rules give an incentive to dilute the emission 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 stimulus to conserve energy, as plant output and system efficiency is not directly considered. 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 recovery 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 standards 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 utilization 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 comprehensive 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 plant’s efficiency for power and heat, while minimizing total plant emissions. This opportunity 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 district energy as key energy and environmental solutions toward 80–90 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. 4.5.1  United States 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 concern 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 provision 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 EPA’s “top-down” approach for determining 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 requirements, 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 control as low as 2 ppmv. Most gas pipeline turbine engines have been regulated in the 15–42 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 pricing 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 regulation for gas turbines used for pipeline compressors, utility combined cycles, and

104

Worldwide Ground-Based Regulatory Framework
GT unit size, heat input <50 MMBTU/hr (electricity, 3.5 MWe) (mechanical, 3.5 MWe) 50–850 MMBTU/hr (3.5–100 MW) Over 850 MMBTU/hr (>110 MW) 42 100 25 15 2.3 5.5 0.55 0.43 ppm lb/MWhr

Units in arctic, offshore < 30 MW > 30 MW (New units, natural gas fuel) 150 96 8.7 4.7

(EPA OAR 2004-0490)

Figure 4.3.  Excerpts of new U.S. EPA rules for gas turbines (Sub-Part KKKK, 2006).

industrial cogeneration plants (Figure 4.3; U.S. EPA 2006, subpart KKKK). Key elements include: • 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 flexibility 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 policies. 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, several IGCC demo plants, and lower-vehicle fuel consumption, the more comprehensive 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 27–33 ppmv in simple cycle applications, and 37–42 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 difficulties. 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 effectiveness, such as steam/water injection flow rate measurement or some type of

106

Worldwide Ground-Based Regulatory Framework
Canada Output–based emission criteria power and heat (grams NOx per GJoutput) 60 50 NOx (ppm) 40 30 140 g/GJ 20 10 0 20 40 60 80 100 Power 240 g/GJ 3–20 MW > 20 MWe Heat 40 g/GJ

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 emissions to the 700 Mt level by 2010 (plus about 200 kT/yr of less air pollution). 4.5.3  European Commission Legislation 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 capture, 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
Table 4.2.  Gas turbine air pollution emission standards in European countries (2005)
Country Size range NOx (mg/Nm3) Gas fuel 150 150 350 300 70 150 100 150 100 150 100 80 60 618 35 58 105 60 NOx (mg/Nm3) Liquid fuel 200 200 350 300 176 70 200 150 200 150 CO (mg/Nm3) Gas fuel 100 100 100 100 No limits No limits 100 100 100 100 100 80 60 50 680 no limits no limits 100 100 CO (mg/Nm3) Liquid fuel 100 100 100 100

107

Austria Belgium Finland France Germany Italy Netherlands Spain Sweden United Kingdom

< 50 MWth > 50 MWth < 100 MWth > 100 MWth 150–300 MWth > 50 MWth 2–20 MWth 20–100 MWth < 100 MWe > 100 MWe < 8 MWth 8–15 MWth 15–50 MWth > 50 MWth All sizes < 500 MWth > 500 MWth < 50 MWth > 50 MWth

618

100 100 100 100 100 100 100 100 680

140 125

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 emission requirements for pipeline units ranged from about 100–200 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 (~15–20 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

The cogeneration efficiency allowance is a progressive incentive that allows this increased emission level to balance a lower-system CO2 rate. In 2004, a cogeneration directive was published by the European Union (European Union, 2004), and a new proposed energy efficiency directive was introduced in 2011 (COM 2011 0370). 4.5.4  European GHG Policies 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 environmental 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 allocation 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), supported by the European Turbine Network (ETN). Individual European countries are also developing their own GHG emission programs in conjunction with the ETS system, some of 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 reduction 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

accomplished through a combination of energy efficiency and renewable energy policies presented in the 2009 Climate Bill. Certain fuel-based energy systems are subject to both carbon and air pollution taxation, with special policy incentives for wind power, cogeneration, and district energy using biomass fuels and modest amounts of natural gas.
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 directives 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, France’s 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 policies in transportation, nuclear efficiency, and renewables, there are ongoing discussions 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 cogeneration 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.

4.5.5  Other International Regions
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). World Bank Group In 1998, the World Bank published emissions guidelines for various thermal energy systems, including gas turbine plants. This document covered both design and operation, with issues addressed including air pollution, GHGs and alternative energy systems, and other solid, liquid, and noise environmental issues. The Pollution Prevention and Abatement Handbook of the World Bank Group’s EHS guidelines included information for environmental assessments relevant to combustion processes designed to deliver electrical or mechanical power, steam, or heat with a total rated heat input capacity above 50 MW of thermal input. For combustion turbine units, the maximum NOx emissions levels were set at 125 mg/Nm3 for gas and 165 mg/Nm3 for diesel (300 mg/Nm3 for No. 6 fuel oil). The document notes that some measures, such as choice of fuel and use of measures to increase energy conversion efficiency, will reduce emissions of multiple air pollutants, including CO2, per unit of energy generation (World Bank, 1998). Optimizing efficiency of the generation process depends on a variety of recommended measures to prevent, minimize, and control air emissions, including:

• 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 combined 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 70–90 percent, compared with 32–45 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.

4.6  Environmental Assessment – Balancing Integrated Environmental and Energy Issues
Environmental assessments often consider best management practices for that sector, and they can take steps to involve all of the relevant environmental impacts

4.6  Environmental Assessment
Table 4.3.  Environmental air emissions
GT unit size 3–15 MWth 3–15 MWth 15–50 MWth > 50 MWth Application Electricity Mechanical drive All All Gas fuel (ppm) 42 100 25 25 Liquid fuel (ppm) 96 150 74 74–146

111

Source:  Excerpt from IFC World Bank, 2007

in a comprehensive and balanced manner. Synergies between various objectives involve concepts of pollution prevention, energy conservation, long-term planning alternatives, and system efficiency. Many aspects of energy project design and public policy involve the balancing of a variety of issues with sometimes conflicting parameters, as previously discussed in Section 2.4.8. Low air pollution, greenhouse gases, air toxics, water impacts, and noise are all important to local and regional areas, with policy trade-offs appropriately evaluated against energy reliability and security (Figure 4.5). This section is intended to summarize some of the choices and trade-offs involved in permitting environmental assessment and life cycle analysis for these facilities, with the following issues: • 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 regulatory strategy, and is there a need to integrate plant efficiency and system reliability 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 environmental 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 outlined 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 interconnected, 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
Energy supply choices

Health Ecosystem air pollution,

Energy security

water impact

Global atmos phere Climate Ozone Layer , toxics, gree nhouse gase s, CFCs, HFC s

Conservation & efficiency Emissions trading Policy Regulations Technology

Demand & consumption System reliability

Economic performance

Figure 4.5.  Balancing several important objectives.

What are cleaner energy choices? Low air pollution, GHG emissions, air toxics and water impacts • Aggressive energy conservation and efficiency • Small renewable energies, biomass fuels • High efficiency natural gas systems (GTCHP, GTCC) • Large hydro & nuclear facilities • Coal & petcoke gasification systems, w/ carbon capture • Waste energy recovery & material recycling

Figure 4.6.  A wide variety of clean energy choices.

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.

4.7  Life Cycle Analysis – Consideration of the Full Fuel Cycle
Environmental assessment (EA) evaluations of energy projects such as electric power, natural gas production, oilsands, gasification, and cogeneration systems should have the ability to compare various alternatives, on an integrated full fuel cycle basis, with emphasis on the operations phase. Although most air pollution is sourced at the point of combustion, the exhaust stack, there is also an upstream component attributable to the choice of energy technology and fuel. Air pollution and greenhouse gas emissions arise out of the production and delivery of their energy

4.7  Life Cycle Analysis – Consideration of the Full Fuel Cycle

113

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 pollution and GHG emissions. The following points relate to the natural gas industry, for example: 1. Before GHG emissions became important, fuel neutrality was generally accepted and emissions rules were developed for individual fuels at their source. With emissions caps and new international trading mechanisms, fuel choice and switching have now become more prominent solutions for GHG reduction. 2. Combustion source emissions from coal and oil are very high, so the additional contribution from upstream emissions do not make a large difference in an analysis. However, source emissions from natural gas and biomass facilities are much lower, so that the upstream GHG and health emissions can be significant to the analysis of natural gas impacts. 3. Most new types of natural gas systems providing cogeneration, district energy, and motive power from gas turbines, recip engines, fuel cells, and hybrid systems, as well as gasification and oilsands choices, will become parts of a sustainable solution. The growth of these sectors, with a more sustainable hydrogen-based fuel, can be balanced with consideration of the appropriate methane, NOx, and SO2 impacts from fuel production and delivery. 4. Methane is the fastest growing greenhouse gas in the atmosphere, and its association with natural gas, shale gas, and coalbed methane production is important. In future, there is a possibility that large reserves of offshore and Arctic gas hydrates could also become available. Their release to the atmosphere must be avoided and must be considered in the emissions quantification. 5. Hydrogen-based natural gas fuels can be supplemented with other gaseous fuels from solids gasification, waste products, or landfills. When all emissions and impacts are combined, rational use of valuable natural gas can be explored with opportunities for other H2 fuels. The pollution prevention improvements in DLN combustion and waste heat recovery 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 25–40 ppm level of NOx and those forced to further limit emissions to single digit levels of 2–9 ppm. The latter come with high cost and low marginal 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.

4.8  Valuation of Emissions for Gas Turbine Energy Systems
Many years ago, the technology choices were evaluated on cost per ton of NOx and/ or PM emissions; today, similar choices are assessed totally on the basis of $ per ton of only CO2. Economic feasibility often dismisses several of the tangible long-term benefits of the best clean thermal energy solutions, such as distributed CHP plants, gas turbine repowering, and coal gasification systems, versus central coal power or greenfield combined cycle utility generation. A challenge, therefore, is how to evaluate any conflicting factors, to determine the worth of combined solutions in terms of environmental performance over the medium and long-term time frames. An analysis of $ per ton of total cost could consider the probable range of benefits so that projects or design choices can be compared with clarity according to future financial risks. Complex large-energy project EAs need to have clear and simple-to-use criteria and values for full fuel cycle emissions, using basic information on methodology and emission factors for relevant pollutants. There has been a lot of past and recent work on this topic, although detailed assumptions and calculations differ somewhat. However, many results are within the +/- 20 percent range, which can easily be used effectively to arrive at an appropriate approximation for long-term planning. Simplicity, and the 80/20 rule, can be important in evaluating choices when detailed, specific information is unavailable. The use of conservative monetary estimates is far better than the disregard of relevant issues (i.e., deeming them of zero value). Project designs for higher $/MW cost systems with economic/financial analyses that are robust in the long term could take into account energy supply and multiple environmental factors, with benefits including: • avoided costs of replacing aging utility, industrial, and commercial boilers;

4.8  Valuation of Emissions for Gas Turbine Energy Systems

115

• 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.
1.  ReplaciNg Separate ENergY PrOdUctiON WitH COgeNeratiON.  A small-scale energy system such as a 1 MWe, 70 percent efficient gas-fired CHP plant has a heat-power ratio of about one to supply power, heat, and cooling to a building or small industry process for 7,000 hrs/yr (Figure 4.7). In upfront costs, the CHP may cost an additional $20/MWhr, or $140,000/yr, but this could reduce or prevent a range of air emissions, reasonably allocated to:

• Acid gases (NOx, SO2, PM) • Greenhouse gases • Air Toxics and CFC prevention • Cooling water impacts

$2000 per ton $20 per ton $ 1000 per kg $0.1 per m3

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 10 10 GJ fuel coal/oil/gas 1 Power boiler (36%) 1 MWe 5 5 GJ gas fuel 3.6 800 kg CO2 7 kg CAC 250 kg CO2 0.25 kg CAC 1 Process heat (80%)

4

5

10 GJ gas fuel 500 kg CO2 0.5 kg CAC 1 MWe 3.6 Gas turbine (36%) WHR

1 4

1

Combined heat and power

Figure 4.7.  Comparing separate heat and power production with CHP system.

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 sensitive 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 regardless of the actual monetary amount for each issue.

4.9  Summary
Today’s design of a modern gas turbine plant to replace aging coal and heavy oil power generation represents a dramatic 90–99 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
Emissions valuation : 300 MWe GTCC plant 2 TWhrs With SCR System tpy NOx Ammonia (5 ppm) PM2.5 N2O x310 CO2 100 50 $000/yr 200 250 DLN Without SCR tpy 400 0 $000/yr 800 0

117

50 10000 727000

250 200 14540 $15440 K

0 0 720000

0 0 14400 $15200 K

Air pollution @ $2000 & $5000/tonne

GHGs @$20/tonne

Figure 4.8.  Valuation comparison of DLN combustion and SCR controls.

clean hydrogen-based fuels, cogeneration, and solid fuel gasification with CO2 capture 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 environmental 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 sectors 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 America’s 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)

Newer output-based emissions standards could incorporate system efficiency to allow the plant designer to minimize fuel consumption and parasitic losses or to increase output to offset other emissions. Environmental assessments and plant permitting could be based on preventing or controlling a combination of emissions impacts, with a goal of achieving continuous improvement in system efficiency,

118

Worldwide Ground-Based Regulatory Framework

cost-effectiveness, energy security, process reliability, lower GHGs, and minimal water impacts.

Abbreviations
BACT BREF CO CFC CAC CCS CCME CHP DLN EIA EPA EU ETS ETN EHS GJ GHG GT GWP HRSG HFC IPCC LAER MACT MWhr IGCC LCPD IPPC NASA NOx NSPS NH3 ppm PM SCR T&D UHC UNECE UNFCC Best Available Control Technology BAT Reference documents Carbon Monoxide Chlorofluorocarbons Criteria Air Contaminant Carbon Capture and Storage Canadian Council of Ministers of the Environment Combined Heat and Power Dry Low NOx Energy Information Administration (United States) Environmental Protection Agency (United States) European Union Emission Trading Scheme European Turbine Network Environmental, Health, and Safety Gigajoule Greenhouse gases Gas turbine Global Warming Potential Heat Recovery Steam Generator Hydrochlorofluorocarbons Intergovernmental Panel on Climate Change Lowest Achievable Emission Rate Maximum Available Control Technology Megawatt hour Integrated Gasification Combined Cycle Large Combustion Plant Directive Integrated Pollution Prevention and Control National Aeronautics and Space Administration (United States) Oxides of Nitrogen New Source Performance Standards Ammonia Parts Per Million by Volume Particulates Selective Catalytic Reaction Transmission and Distribution Unburned Hydrocarbons United Nations Economic Commission for Europe United Nations Framework Convention on Climate Change

References
RefereNces

119

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. 21–23) 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 pollution 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). “America’s 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-2004–0490, pgs 38482–506, July.   40 CFR Part 60  – Subpart Da  – Standards of Performance for Electric Utility Steam Generating Units.

120

Worldwide Ground-Based Regulatory Framework
Western Australia. (2000). Guidance for the Assessment of Environmental Factors Western Australia; Guidance Statement for Emissions of Oxides of Nitrogen from Gas Turbines, May. The World Bank. (1998). Pollution Prevention and Abatement Handbook, WORLD BANK GROUP, Thermal Power: Guidelines for New Plants, July.

PART 2

FUNDAMENTALS AND MODELING: PRODUCTION AND CONTROL

5 Particulate Formation
Meredith B. Colket III

5.1 Introduction
5.1.1  Definition – Smoke/Soot/Carbonaceous Emissions Carbonaceous materials emitting from the exhaust of gas turbine engines are frequently referred to as soot emissions, nonvolatile particulates, or smoke. Frequently, these terms are used interchangeably. Such emissions typically consist of single particles ranging from 10–80 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 magnification for a combustor operating at 80 percent power is shown in Figure 5.1. These carbonaceous soot particles should be contrasted with volatile particulates (see Chapter 6), although volatiles may condense onto soot particles. In particular, 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 environment 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 military 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 2.5 ­ particulates 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).

2007; Anderson et al., 2011) on particulate emissions from various engines, as well as the examination of probe and sample-line effects and the role of volatile particulate emissions. The latter is discussed in detail in Chapter 6. 5.1.2  Environmental Considerations Ambient levels of airborne particulate matter (PM) is the subject of increasing attention because of recent studies showing correlations between short-term exposure to fine PM in the local environment and acute adverse health effects (Dockery et al., 1993; Bachmann et al., 1996; Wolff, 1996; Kumfer and Kennedy, 2009). These studies provide the basis for new national ambient air quality standards for PM with diameters less than 2.5 microns (PM2.5) set by the U.S. Environmental Protection Agency (1997). Regional areas out of compliance with these standards have developed local site implementation plans (SIP) for control of the ambient particulate levels. Such plans could include indirect control of emissions for aircraft by setting limitations on aircraft flight schedules or trajectories. Comparable European standards for PM2.5 were defined in the new Directive on Ambient Air Quality and Cleaner Air for Europe. This directive defines an upper limit value of 25 micrograms of PM2.5 per m3 in 2015 for the ambient atmosphere (see Priemus and Schutte-Postma, 2009). In addition to health effects, it is important to recognize that the Intergovernmental Panel on Climate Change (Penner et al., 1999) has implicated carbonaceous particulates as a contributor to environmental problems related to global warming, as such particulates absorb incoming radiation in the atmosphere and significantly increase the absorption characteristics of snow and ice in the Arctic. 5.1.3  Regulation Methods – Present and Future 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 visibility 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, substantially less information was available on health and other environmental effects. In addition, there was generally poor understanding of the control of particulate emissions 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 during 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 emissions 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
Reading of absolute reflectance made with a calibrated photodetector (ANSI PH 2.17) referenced to NIST Standard plaques

160 deg F from sample probe

0.023 lbm exhaust gas/ sq. in. filter at 0.5 cfm

Reflectance (%) 0

#4 Whatman filter paper

100

Typical filter spots

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 emissions (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 concentration can be obtained using Differential (Electrical) Mobility Analysis (DMA). Typically, a Scanning Mobility Particle Sizer (SMPS) electrostatic separator or classifier (see Figure 5.4) is coupled to a condensation nucleus counter (CNC). The electrostatic 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 photometry (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). 5.1.4  Interpretation of Smoke Number 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 manufacturer 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 parameter (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 uncertainties 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
10

Particulate Formation

Wood, A.D., 1975 8 Soot mass (mg/m3) Eckerle and Rosfjord, 1987 Norgren and Ingebo, 1975 Champagne, 1971 (filtered) 6 Champagne, 1971, unfiltered Hurley, 1993 4

2

0 0 5 10 15 Smoke No. 20 25 30

Figure 5.5.  Literature correlations relating soot mass and smoke number.

collection processes on filters. Hence, the assumption of single-particle emissions from present day aero engines is probably valid, to the first order. These more recent results are compared with literature results in Figure 5.6. The scatter of the new data set only slightly reduces the uncertainty observed in Figure 5.5. Also included in this figure are recent data sets from Stouffer (2001, University of Dayton Research Institute, personal communication) using experiments from a well-stirred reactor. A recommended correlation (Colket et al., 2003) is also plotted in Figure 5.6. This expression agrees well with the recent expression suggested by Wayson and colleagues (2009). The recommended expressions for the SN to mass correlation are: if m (mg/m3) < 2.5 then: if m (mg/m3) > 2.5 then: or equivalently: if SN < 18.7, then: if SN > 18.7, then: SN = −1.8743*m2 +12.117*m SN = 12.513*m0.4313 m = 3.232*(1−(1-SN/19.58)1/2) m = 0.002751*SN2.319.

5.1.5  Gas and Particulate Sampling  When sampling soot particles, investigators must consider a variety of loss mechanisms. 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
60 50 Soot mass (mg/m3) 40 30 20 10 0

129

Champagne, 1971 (filtered) Champagne, 1971, unfiltered Wood, A.D., 1975 Norgren and Ingebo, 1975 Eckerle and Rosfjord, 1987 Hurley, 1993 Stouffer, 2001 UTRC, 2000 UTRC Recommendation

0

10

20

30

40 Smoke No.

50

60

70

80

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) Thermophoretic Inertial, bends Diffusional Electrostatic

Total transport efficiency

Experimental measurements

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

5.2  Fundamentals of Particulate Formation and Oxidation
The formation and oxidation of soot is complex, with understanding of the controlling phenomena revealed slowly through scientific investigations over a period of more than fifty years. Excellent review articles documenting the slow advancement in the understanding of critical processes have been produced approximately every five to ten years (Palmer and Cullis, 1965; Wagner, 1979; Haynes and Wagner, 1981; Glassman, 1988; Kennedy, 1997; Wang, 2011). The best understanding of the origin of soot particles has come through studies in laboratory flames with simple flow fields. Such investigations permit samples to be collected at different stages of particle formation for time-resolved records of number concentration and size distribution of soot particles and composition of the accompanying gas (see Wang, 2011 and McKinnon and Howard, 1992 and references contained therein). Based on such studies, soot formation involves (see Figure 5.8): particle nucleation or inception, mass addition (i.e., surface growth) by reaction with gas-phase molecules; coagulation of particles through particle-particle sticking collisions; mass removal from particles by pyrolytic processes leading to dehydrogenation and structural rearrangement of the condensed material; and oxidation. The growth of polycyclic aromatic hydrocarbons (PAH) from the smallest, two to three aromatic ringed species to the larger species, with four or more aromatic rings, is inextricably linked to soot formation. PAH are the reactants in the formation of the initial soot particles (soot nucleation) and are also soot surface growth reactants. The composition of soot is polycyclic aromatic in character, and the distinction between the largest PAH molecules in sooting flames and the smallest soot particles is somewhat arbitrary, and will remain so until the structural and chemical characteristics of the soot nuclei are better understood. Violi and coworkers (2002, 2004) have attempted to model this growth process, and Wang (2011) has recently analyzed various proposals for the inception process, but regardless, this remains an unresolved yet active area of research. Despite evidence obtained about twenty years ago (D’ Anna et al., 1994; Dobbins and Subramaniasivam, 1994), the community has been slow to recognize that nascent soot particles are liquid-like rather than solid, carbonaceous materials. These liquid-like particles or nano-organic particulates (NOCs) have a density of about 1.2 grams/cc (versus 1.8 grams/cc for mature soots) and soon undergo dehydrogenation and carbonization at the elevated temperatures in flames. Based on more recent evidence and logical arguments, the research community has broadly adopted this picture through discussion at a 2007 conference on carbon particulates (Bockhorn et al., 2009). A simplified, sequential picture of the soot formation process is depicted here. It starts with fuel (oxidative) pyrolysis, the formation of the first aromatic ring structures, and is followed by the formation of PAH. Throughout the entire process, an intimate connection forms between the gas-phase species and the particulate formation/oxidation processes. A key reactant supporting PAH growth is C2H2. “Surface” growth of soot particles involves both

5.2  Fundamentals of Particulate Formation and Oxidation
Inception CxHy Jet A H C2H2 Spray vaporization Alkylatedaromatics PAH H2 C2H2 N aphthenes Surface growth & coalescence Oxidation

131

CO

Gas-phase kinetics

Ageing & coagulation

Emissions

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 aromatic 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 generally 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 temperatures (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 incipient particles may be characterized as liquid droplets (see discussion in Bockhorn et al., 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 particles, 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 surface 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 addition continually mix and alter the local conditions. Hence, the scenario depicted previously 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 (turbulent, diffusion) flames is difficult because of the complexities of the physical and chemical processes involved; it remains one of the greatest challenges to computational modeling of combustion. First, formation of soot is an incompletely understood 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 temporarily converted to soot. Finally, turbulent mixing and reacting flow complicate simulations 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 traditional modeling methods.

5.3 Inception
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

step is initiated by propargyl (C3H3) radical recombination or C3H3 + C3H4 (Colket and Seery, 1984; Wu and Kern, 1987) combination to form benzene or phenyl radical through a complex rearrangement pathway (Miller and Melius, 1992; Melius et al., 1993; Miller and Klippenstein 2001, 2003). Hence, processes leading to the formation of C3H3 and C3H4 can be the bottleneck to ring and hence soot formation; quantitative prediction of these species is likely a prerequisite to quantitative modeling for laboratory flames. It is worth noting that, while the rate coefficients for these steps are now established probably to within a factor of two (Miller and Klippenstein, 2003), uncertainties remain in the pathways that govern the formation and destruction of such critical species. Other steps, initiated by C2H2 addition to n-C4H5 or to n-C4H3, may contribute (Frenklach et al., 1985; Colket, 1986; Glassman, 1988), as well as reactions involving cyclopentadienyl moieties (Marinov et al., 1998). Questions about the importance of the “normal” variety of C4H3 or C4H5 radicals with the radical sites on the end (or terminal) carbons have been raised for some time as these isomers are strongly disfavored thermodynamically and recent measurements confirm these concerns (Hansen et al., 2006). Hence, it is probable that i-C4H5 and i-C4H3 play a more important role in ring formation (the latter species have radical sites on interior carbons, rather than terminal carbons) than existing models suggest. Modeling the formation of multi-ring aromatics is more challenging, with few quantitative demonstrations of such simulations. The reaction pathway (Figure 5.9) suggested originally by Frenklach and coworkers (1985) is: 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 5C 2 H (phenylacetylene) + H C 6 H 5C 2 H + H ⇔ C 6 H 4 C 2 H + H 2 C 6 H 4 C 2 H + C 2 H 2 ⇔ C 6 H 4 (CHCH)C2H ⇔ C 10 H 7 (naphthalenyl) Analogous reactions lead to the formation of phenanthrene, anthracene, pyrene, and other larger polycyclic aromatic hydrocarbons. Several other reaction pathways  – involving toluene/benzyl and indene/indenyl (Colket and Seery, 1994) or cyclopentadiene/cyclopentadienyl dimerization (Marinov et al., 1998, for example) – have been proposed. In addition, species such as C6H5CCH2 may play ­ a role via: 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 interior 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 intermediate species to estimate the formation of a large polycyclic aromatic structure.

134
Fragmentation

Particulate Formation
Fragmentation

• +C2H2(-H)

C≡C-H +H(-H2) •

C≡C-H

+C2H2 • C≡C-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 Where the rate constant, k, is calculated from collision theory.

5.4 Surface 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
Table 5.1.  HACA mechanism for surface growth as utilized by Appel and Colleagues (2000)
Reactions considered S1. S2. S3. S4. S5. S6. H + Csoot – H ⇔ Csoot· + H2 OH + Csoot – H ⇔ Csoot· + H2O H + Csoot· ⇒ Csoot – H C2H2 + Csoot· ⇒ Csoot· + H O2+ Csoot· ⇒ 2 CO + products OH + Csoot – H ⇒ CO + products log10(Af) n Efor 13 1.43 – 3.8 7.5

135

13.62 – 10.00 0.734 13.30 – 7.90 1.56 12.34 – reaction probability = 0.13

Table 5.2.  Modified version of the Frenklach and Wang soot growth mechanism as developed by Colket and Hall (1994)
Reactions considered 1. 2. 3. 4. 5. H + Csoot – H ⇔ Csoot· + H2 H + Csoot· ⇔ Csoot – H Csoot· ⇔ products + C2H2 C2H2 + Csoot· ⇔ C(s)CHC’H Csoot– CHC’H ⇔ Csoot· + H log10(Af) 14.40 14.34 14.48 12.30 10.70 Efor 12 – 62 4 – log10(Ar) 11.6 17.3 – 13.7 – Erev 7 109 – 38 –

Note:  Units are as in Table 5.1.

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 α. They proposed that:

α = tanh ( α logµ1 + b)
where μ1 is the first moment of the soot particle distribution and a and b are fitted parameters determined separately for each experiment. Zhang et al. (2009) instead utilized a value suggested by Xu et al. (1998), which depends on temperature: α = 0.004exp(10,800/T) but also pointed out that there are no universal values for α 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 computed by assuming steady-state levels of intermediate species: (k1[H] +k−2 )( k4 k5 [C 2 H 2 ] − k3 (k−4 + k5 )) χ dm = 2 mc dt (k−1[H 2 ] + k2 [H] + k3 )(k−4 + k5 )+ k4 k5 [C 2 H 2 ] where mc is the mass of a carbon atom and χ is the surface density of Csoot - H sites (~ 2.3·10 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, reversibility 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).

5.5 Soot 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 oxidizing 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 concentration levels are negligible) begin to transfer toward fuel-lean conditions. Specifically, this threshold occurs when equivalence ratios decrease below approximately 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 thermodynamically stable species such as CO2 and H2O is slow, as these processes are endothermic 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 equilibrium values of OH suggest that this process is too low to dominate over oxidation 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 temperatures 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 interpretations of this latter work are incomplete. Particle breakup simply may be enhanced in Echavarria’s 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 example 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 conditions 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 reasonably 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    k  x = 1+ T  kB po 2  
−1

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 stoichiometric 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 critical 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 combustion as the geometric equivalent of a set of spherical particles, of diameter, d. Some

138

Particulate Formation
Table 5.3.  Rate constants for the NSC oxidation rate by O2

kA
kB kC kT

20 exp (–15100/T) 4.46 · 10 exp(–7640/T) 1.51 · 10–5 exp(–48800/T) 21.3 exp (20600/T)
–3

g/cm2/sec/atm g/cm2/sec/atm g/cm2/sec 1/atm

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 usually given by: d[C(s)] / dt = − k x [X] SA MWc 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 Avogadro’s number, respectively, the OH oxidation is: ROH = 0.13NOH RgT 2 πWOH P 12 = 16.7 OH 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).

5.6 Coalescence and Agglomeration
Particle-particle collision frequencies are governed by the Smoluchowski equation (1916): dN/dt = − βN 2 The coagulation of soot particles is modeled usually as a free-molecule aerosol dynamics problem. Every collision results in a coalescence process in which the collision forms a new single particle with a larger diameter. Agglomerization results in a larger agglomerated particle with an extra (set of) particle(s) attached to each

5.6  Coalescence and Agglomeration
1.E-02 Po2(atm) Oxidation rate (gm/cm2/sec) 0.5 1.E-03 0.2 0.1 0.05 1.E-04 0.02 0.01 0.005 1.E-05 3 4 10 5
4/T(1/K)

139

6

7

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, β, in the previous equation is given by:

π β (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: Vm1m2 = 8kT  1 1  +  m1 m2   π 

A nominal of value of 1.5 is recommended for the Van der Vaals enhancement factor, 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: SA = 4.29
3

NM 2 ρ2

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 ρ (grams/cc soot) represents the primary particle mass density. Such monodisperse models offer a tremendous

140

Particulate Formation

computational advantage in that they eliminate additional equations necessary to track particle size distributions. The method of moments (Frenklach and Wang, 1990) is frequently utilized, but convergence issues arise with attempts to accurately treat multimodal size distributions. Quadrature method of moments (Blanquart and Pitsch, 2007) offers an advancement as it can simulate effects of multi-modal size distributions at a reasonable added cost for applications in turbulent combustion environments and hybridized methods (Mueller et al., 2009) help to resolve numerical issues. The well-known sectional method for particle size representation of spheres can provide detailed information on particle size distribution assuming that a sufficiently large number of sections are employed (typically > 40). However, a large number of sections will be a burden for turbulent combustion CFD codes and the moment methods may be the only practical alternative. Hall and colleagues first described the application of the sectional approach to soot modeling in 1997, and the method has been utilized more recently by a variety of authors (see, for example, Richter et al., 2005). The contributions from the inception processes are incorporated as a source term in the dynamical equation for the first sectional bin, whose lower mass boundary is set equal to the mass of the assumed inception species. The spherical particle sectional model nominally imposes no constraint on the final particle size, and, without modifications, does not account for aggregate formation. Coalescence (in which two or more particles combine to form a single larger particle) destroys particle surface area, whereas aggregation (in which two solid particles combine to form an aggregated, fractal structure), to the first order, does not. This is an important consideration because of the dependence of surface growth and oxidation on particle surface area. Adding equations for the number of primary spheroids within a section makes it possible to model the formation of soot aggregates (Zhang et al., 2009; Park and Rogak, 2004).

5.7  Related Phenomena
5.7.1  Formation Time Scales 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 temperatures 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. 5.7.2  Temperature/Pressure Effects 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 dependence of soot production varies with flame type. In premixed flames, Bonig and colleagues (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 exponent 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. 5.7.3  Soot Ageing Soot Ageing: Soot primary particles reach a maximum size because of active surface site deactivation (Woods and Haynes, 1994). This process is also referred to as “ageing” of surface sites. Singh and colleagues (2005) tested functional dependences 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 (α) that is a function of the average particle size and gas temperature, but not explicitly to individual 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 dependence 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 (40–50 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 process than the ageing process described earlier, and it is logical to assume it will also impact surface growth rates.

142

Particulate Formation

5.7.4  Impact of Radiation Loss 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 Planck’s 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 20–30 percent of the chemical energy to be lost to surroundings via radiation. Such losses can directly cause the reduction of the local flame temperature by hundreds of degrees Kelvin. In fact, in many atmospheric combustion devices, including home fireplaces, such radiative loss by soot particles is a key feature 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 critical as it is for atmospheric pressure flames. For flames with low sooting levels, the optically thin approximation (i.e., ignoring 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 thickness 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. 5.7.5  Fuel (Including Alternative Fuel) Effects 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

aromatic-laden fuels produce large amounts. Extreme examples would be methane at one end and heavy coal/shale-derived fuels at the opposite spectrum. Extensive NASA and AF-sponsored testing of fuels (including NASA’s ERBS, “Experimental Reference Broadened Specification”) in the late ’70s and early ’80s in various operational gas turbine engines showed that physical (rather than chemical) properties of the fuel were the dominant factors in most aspects of gas turbine engine performance. Only for smoke (soot) emissions and combustor liner heating (through radiation from soot) were chemical properties important, and they seemed best correlated against overall fuel H/C ratio or hydrogen content of the fuel (Mosier, 1984; Odgers and Kretschmer, 1984; Lefebvre, 1985) as shown in Figure 5.11. While the presence of aromatic and polycyclic aromatic hydrocarbons in the fuel should aggravate soot emissions, the experimental results were relatively independent of aromatic concentration or nature (e.g., single or fused rings). Naphthalene content was identified as a secondary factor that could not be neglected (Moses, 1984; Odgers and Kretschmer, 1984; Sampath and Gratton, 1984). Traditional combustors (pre-1985) with very fuel-rich primary zones seemed most sensitive to fuel composition effects on soot, while leaner-operating engines show less effect (Odgers and Kretschmer, 1984). None of these early studies, however, considered the impact on particle size, which has received increased attention recently. Multiple specifications for jet fuel collectively limit the emissions of soot beyond that which engine/combustor design can control. These include mass percent of hydrogen, aromatic content, naphthalene content, and smoke point. Actual minimum and maximum amounts may vary depending on whether the fuel is for military or commercial applications. Furthermore, specifications may be optional, that is, if one specification is met, another may not be required. Specifications for military jet fuel require hydrogen contents greater than 13.4 mass percent. Most liquid hydrocarbon fuels will have hydrogen mass ranging from about 13 percent to 15 percent. Methane and propane, with hydrogen masses of 25 percent and 18 percent, respectively, far exceed values in typical liquid fuels, and correspondingly soot emission levels with such fuels (such as in ground-based gas turbine engines) are relatively quite low. Upper limit aromatic levels in jet fuels are typically either 20 percent (commercial) or 25 percent (military) (by volume) and the concentration of naphthalenes must be 3 percent (volume) or less if Smoke Point limits are not achieved. Smoke Point (SAE Aerospace Recommended Practices 1179) is a direct, empirical measurement of soot produced by an atmospheric pressure flame burning a specific fuel. The Smoke Point of a given fuel is the height at which smoke is observed in a candle/ wick flame. Higher Smoke Point numbers indicate the fuel has a lower tendency to produce soot. Researchers have completed substantial work in the ’70s and ’80s and over the last five years in refining the (Yield) Threshold Sooting Index and relating it to Smoke Point in an attempt to develop a more quantitative method (Yang et al., 2007; McEnally and Pfefferle, 2009; Mensch et al., 2010). Such new methods are not yet industrial standards, but they are useful research tools. Applications of premixed, lean-burn technologies, required to control NOx emissions, result in very low or zero soot emissions. Any soot emissions that do occur are likely due to use of a diffusion flame pilot or incomplete premixing.

144
2.00

Particulate Formation

Smoke No. /(Smoke No.)14%H

J79 1.75 TF30 1.50 F101 TF33 1.25 TF41 1.00 F100

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 8–25 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 micrographs 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

5.7.6  CO2, H2O, N2 Dilution Effects Direct dilution by adding inerts can noticeably reduce soot production (Gomez et al., 1987). The inert may be added with the fuel or the air, but this effect is strongest 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 flameless 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).

5.7.7  PAH Absorption
Polycyclic aromatic hydrocarbons (PAH) are believed to play a critical role in particle inception and particle surface growth, as described in Chapters 5.3 and 5.4. As temperature decreases through the turbine and into the exhaust plume, unburned hydrocarbon species, including PAH, condense onto soot particles. As the temperatures in such regions are low, additional chemical conversion of such species are slow. Instead, the soot particles carry the condensed species into the atmosphere. Other species, such as H2SO4, which combines with water upon cooling, may also condense onto the soot particles, although past work has indicated that some of the particles are hydroscopic and some hydrophobic. Additional discussion of this and related topics is provided in Chapter 6.

5.8  Particulate Formation in Combustion Systems
5.8.1  Impact of Combustor Design on Soot Emissions The biggest early impact of combustor hardware changes on soot/smoke emissions occurred in the ’70s. In early engine designs, pressure atomized nozzles with no adjacent air addition led to localized, very fuel-rich regions in the primary zone of the

146

Particulate Formation

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 equivalence 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 conditions 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 rapidly 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 necessary 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 particle 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 conditions for an aeroderivative would severely degrade its durability and hence soot emissions 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 produced 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 exceeding 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, D1331–89. 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/ TM–2011–217059, 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-96–013 (NTIS PB97–115406), Office of Air Quality Planning and Standards, Research Triangle Park, NC.

148

Particulate Formation
Balthasar, M., and Frenklach, M. (2005). Proc. Combust. Inst. 30: 1467–75. Bendetto, D., Pasini, S., Falcitelli, M., Marca, C. La., and Tognotti, L. (2000). Comb. Sci. & Tech. 153: 279–94. Bengtsson, K. U. M., Benz, P., Schären, R., and Frouzakis, C. E. (1998). Proc. Comb. Inst. 27: 1393–9. Benish, T. G., Lafeur, A. L., Taghizadeh, K., and Howard, J. B. (1996). Proc. Comb. Inst. 26: 2319–26. 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): 405–11. 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 25–8. Blazowski, W. S., and Jackson, T. A. (1978). “Evaluation of Future Jet Fuel Combustion Characteristics,” AFAPL-TR-77–93, July. Bockhorn, H., D’ Anna, 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: 1581–7 . Brandt, O., and Roth, P. (1989). “A Tunable IR-Diode Laser Technique for Measuring Reaction Rates of High Temperature Aeroslos.” Combustion and Flame 77: 69–78. 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, GT2003–38627 , June. Bruno, C., and Vallini, L. (1999). “Flameless Combustion and its Application to Aero-engines.” Fourteenth International Symposium on Air Breathing Engines, September 5–10, 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, GT2010–23524, 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: 851–64. 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 F49620–94-C-0059, UTRC97–5.906.0001–1, 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, GT2004–54001, 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. R03–6.100.0007–1, AF Technology Investment Agreement F33615– 00–2–2001, 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: 883–91. D’ Anna, A., D’ Alessio, A., and Minutolo, P. (1994). “Spectroscopic and Chemical Characterization 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
Dobbins, R. A. (1996). Comb. Sci and Tech 121: 103. Dobbins, R. A., Fletcher, R. A., and Lu, W. (1995). Combust. Flame 100: 301–9. Dobbins, R. A., and Subramaniasivam, H. (1994). “Soot Precursor Particles in Flames,” in Soot Formation in Combustion, Bockhorn, H., ed. Springer-Verlag, Berlin, p. 290. Dockery, D. W., Pope, III, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay, M. E., Ferris, Jr., B. G., and Speizer, F. E. (1993). New England Journal of Medicine 329: 1753. Echavarria, C. A., Jaramillo, I. C., Sarofim, A. F., Lighty, J. S. (2011). Proc. Comb. Inst. 33: 659–66. Eckerle, W. A. and Rosfjord, T. J. (1987). “Soot Loading in a Generic Gas Turbine Combustor.” AIAA-87–0297 . Edwards, T. (Chair) (2006). “Jet Fuel Surrogate Research: A White Paper for Presentation and Discussion at the Surrogate Fuels Workshop – Part IV.” February 13. Fairweather, M., Jones, W. P., Ledin, H. S., and Lindstedt, R. P. (1985). Proc. Comb. Inst. 24: 1067–74. Fairweather, M., Jones, W. P., and Lindstedt, R. P. (1992). Combust. Flame 89: 45. Fenimore, C. P., and Jones, C. W. (1967). J. Phys. Chem. 71: 593. Flower, W. L. and Bowman, C. T. (1986). “Soot Production in Axisymmetric Laminar Diffusion Flames at Pressures from One to Ten Atmospheres.” Proc. Comb. Inst. 26: 1115–24. Friedlander, S. K., and Wang, C. S. (1966). Journal of Colloid Interface Science 22: 126. Frenklach, M., Clary, D., Gardiner, Jr., W. C., and Stein, S. E. (1985). “Detailed Kinetic Modeling of Soot Formation in Shock-Tube Pyrolysis of Acetylene,” Proc. Comb. Inst. 20: 887–901. Frenklach, M., and Wang, H. (1990). “Detailed Modeling of Soot Particle Nucleation and Growth,” Proc. Comb. Inst. 23: 1559–66.   (1994). “Detailed Mechanism and Modeling of Soot Particle Formation,” in Soot Formation in Combustion, Mechanisms and Models, Bockhorn, H. ed., Springer Verlag, Berlin, p. 165. Fuchs, N. A. (1964). Mechanics of Aerosols, Pergamon, pp. 291–4. Gelbard, F. (1982). MAEROS Users Manual, NUREG/CR-1391, (SANDD80–0822). Gelbard, F., and Seinfeld, J. H. (1980). “Simulation of Multicomponent Aerosol Dynamics,” J. Coll. Int. Sci. 78: 485. Glarborg, P., Kee, R. J., Grcar, J. F., and Miller, J. A. (1986). “PSR: A Fortran Program for Modeling Well-Stirred Reactors.” Sandia Report No. SAND86–8209, Sandia National Laboratories. Glassman, I. (1988). “Soot Formation in Combustion Processes.” Proc. Combust. Inst. 22: 295–311. Glaude, P. A., Curran, H. J., Pitz, W. J., and Westbrook, C. K. (2000). “Kinetic Study of the Combustion of Organophosphorous Compounds,” Proceedings of the Combustion Institute 28: 1749–56. Gomez, A., and Glassman, I. (1988), “Quantitative Comparison of Fuel Soot Formation Rates in Laminar Diffusion Flames,” Proceedings of the Combustion Institute 21: 1087 . Gomez, A., Littman, M., and Glassman, I. (1987). “Comparative Study of Soot Formation on the Centerline of Axisymmetric Laminar Diffusion Flames: Fuel and Temperature Effects,” Combustion and Flame 70: 225. Gosman, A. D., and Ioannides, E. (1983). “Aspects of Computer Simulation of Liquid- Fueled Combustors,” Journal of Energy 7: 482–90. Gran, R., Melaaen, M. C., and Magnussen, B. F. (1994). “Numerical Simulation of Local Extinction Effects in Turbulent Combustor Flows of Methane Air.” Proceedings of the Combustion Institute 25: 1283–91. Hall, R. J. (1994). “Radiative Dissipation in Planar Gas Soot Mixtures,” J. Quant. Spectrosc. Radiat. Transfer 51: 635–44. Hall, R. J., Smooke, M., and Colket, M. B. (1997). “Predictions of Soot Dynamics in Opposed Jet Diffusion Flames,” in Physical and Chemical Aspects of Combustion: A Tribute to Irvin Glassman, Sawyer, R. F. and Dryer, F. L. eds., Combustion Science and Technology Book Series, Gordon and Breach, Langhorne, PA, pp. 189–230.

149

150

Particulate Formation
Handbook of Aviation Fuel Properties (1983). Prepared by the Coordinating Research Council. Hansen, N., Klippenstein, S. J., Taatjes, C. A., Miller, J. A., Wang, J., Cool, T. A., Yang, B. et al. (2006). “Identification of C5Hx Isomers in fuel-Rich Flames by Photoionization Mass Spectrometry and Electronic Structure Calculation,” J. Phys. Chem. A, 110(10): 3670–8. Harris, S. J., and Maricq, M. M. (2001). “Signature Size Distributions for Diesel and Gasoline Engine Exhaust Particulate Matter.” Journal of Aerosol Science 32(6): 749–64. Harris, S. J., and Weiner, A. M. (1983). “Surface Growth of Soot Particles in Premixed Ethylene Air Flames,” Combustion Science and Technology 31: 155–67 . Haynes B. S., and Wagner H. Gg. (1981). “Soot Formation.” Progress in Energy and Combustion Science 7: 229–73. Hinds, W. C. (1982). Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles, New York: John Wiley. Hoshizaki, H., Wood, A. D., Seidenstien, S., Brandalise, B. B., and Myer, J. W. (1975). “Development of an Analytical Correlation between Gas Turbine Engine Smoke Production and Jet Plume Visibility.” Lockheed Palo Alto Research Laboratory, Air Force Aero Propulsion Laboratory Report AFAPL-TR-76–29, July 31. Howard, J. B., and Kausch, W. J. (1980). “Soot Control by Fuel Additives,” Progress in Energy and Combustion Science 6: 263–76. Hura, H. S., and Glassman, I. (1988). “Soot Formation of Diffusion Flames of Fuel/Oxygen Mixtures,” Proc. Comb. Inst. 22: 371–8. Hurley, C. D. (1993). “Smoke Measurements inside a Gas Turbine Combustor.” AIAA-93–2070. ICAO (1995). Engine Exhaust Emission Data Bank, first edition, Doc 9646-AN/943, Montreal, Quebec, Canada. Kee, R. J., Rupley, F. M., and Miller, J. A. (1989). “CHEMKIN II: A Fortran Chemical Kinetics Package for the Analysis of Gas-Phase Chemical Kinetics.” Sandia Report No. SAND89– 8009, Sandia National Laboratories. Kennedy, I. M. (1997). “Models of Soot Formation and Oxidation.” Prog. Energy Combust. Sci. 23: 95–132. Köylü, Ű. Ő., McEnally, C. S., Rosner, D. E., and Pfefferle, L. D. (1997). “Simultaneous Measurements of Soot Volume Fraction and Particle Size/Microstructure in Flames Using a Thermophoretic Sampling Technique,” Combustion and Flame 110: 494–507 . Kumfer, B., and Kennedy, I. (2009). “The Role of Soot in the Health Effects of Inhaled Airborne Particles,” in Combustion Generated Fine Carbonaceous Particles, Bockhorn, H., D’ Anna, A., Sarofim, A. F., and Wang, H. eds., KIT Scientific Publishing, Karlsuher, pp 1–15. Lai, F. S., Friedlander, S. K., Pich, J., and Hidy, G. M. (1972). “The Self-Preserving Particle Size Distribution for Brownian Coagulation in the Free-Molecule Regime,” J. Colloid Interface Science 39: 395. Lee, W., and Na, Y. D. (2000). “Soot Study in Laminar Diffusion Flames at Elevated Pressure Using Two-Color Pyrometry and Abel Inversion.” JSME Int. J. B 43: 550–5. Lefebvre,A. H. (1985).“Influence of Fuel Properties on Gas Turbine Combustion Performance,” AFWAL-TR-84–2104, January. Lindstedt, P. (1994). “Simplified Soot Nucleation and Surface Growth Steps for Non-Premixed Flames,” in Soot Formation in Combustion: Mechanisms and Models, Bockhorn, H. ed., Springer-Verlag, New York. Liscinsky, D., Colket, M. B., Hautman, D. J., and True, B. (2001). “Effect of Fuel Additives on Particulate Formation in Gas Turbine Combustors,” AIAA 2001–3745, presented at the 37th AIAA Joint Propulsion Conference and Exhibit, Salt Lake City, Utah, July 8–11. Liscinsky, D., and Hollick, H. (2008). “Effect of Particle Sampling Technique and Transport on Particle Penetration at the High Temperature and Pressure Conditions Found in Gas Turbine Combustors and Engines.” Summary Report of Year 1 Activities, NASA Contract No. NNC07CB03C, United Technologies Research Center, East Hartford, CT, February 29.

References
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: 719–29. 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-GT0062), New Orleans, LA, June. Marchal, C., Delfau, J.-L., Vovelle, C., Morèac, G., Mounaïm-Rousselle, C., and Mauss, F. (2009). “Modelling of Aromatics and Soot Formation from Large Fuel Molecules,” Proceedings of the Combustion Institute 32: 753–9. 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: 605–13. 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: 621–8. Mensch, A., Santoro, R. J., Litzinger, T. A., and Lee, S.-Y. (2010). “Sooting Characteristics Of Surrogates For Jet Fuels,” Combust. Flame 157: 1097–105. McCrain, L. L. and Roberts, W. L. (2005). “Measurements in the Soot Volume Field in Laminar Diffusion Flames at Elevated Pressures.” Combust. Flame, 140: 60–9. McEnally, C., and Pfefferle, L. (2009). “Sooting Tendencies of Nonvolatile Aromatic Hydrocarbons,” Proc. Comb. Inst. 32: 673–9. 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: 1497–505. 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: 202–9. 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: 7254–66.   (2003). “The Recombination of Propargyl Radicals and Other Reactions on a C6H6 Potential,” J. Phys. Chem. A 107: 7783–99. Miller, J. A., and Melius, C. F. (1992). Combust. Flame 91: 21–39. Moses, C. (1984). “U.S. Army Alternative Gas-Turbine Fuels Research: MERADCOM.” Combustion Problems in Turbine Engines, AGARD-CP-353, January, pp. 7–1 to 7–10. Mosier, S. A. (1984). “Fuel Effects on Gas Turbine Combustion Systems.” Combustion Problems in Turbine Engines, AGARD-CP-353, January, pp. 5–1 to 5–15. Mueller, M. E., Blanquart, G., and Pitsch, H. (1975). “Hybrid Method of Moments for Modeling Soot Formation and Growth,” Combust. Flame 156: 1143–55. 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 1000-2000°C,” in Proceedings of Fifth Carbon Conference 1: 154, Pergamon Press, Inc., Oxford.

151

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. 8–1 to 8–10. O’Rourke, P. J. (1981). Collective Drop Effects on Vaporizing Liquid Sprays, PhD dissertation, Princeton University. O’Rourke, 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): 1155–73. 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: 1397–405. 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: 1485–91. 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. 6–1 to 6–12. 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: 1457–66. 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: 593–606. 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: 613–28. 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: 117–39. 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: 2013–20. Smoluchowski, M. (1916). “Drei Vorträge über Diffusion, Brownsche Molekularbewegung und Koagulation von Kolloidteilchen,” Physik. Zeit. 17: 557–71, 585–99. 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
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 4–5. 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: 2343–9. Violi, A., Sarofim, A. F., and Voth, G. A. (2004). “Kinetic Monte Carlo-Molecular Dynamics Approach To Model Soot Inception,” Combust Sci. Technol. 176: 991–1005. Von Gersum, S., and Roth, P. (1990). “High temperature oxidation of soot particles by. O atoms and OH radicals,” J. Aerosol Science 21: S31–S34.   (1992). “Soot oxidation in high temperature N2O/Ar and NO/Ar mixtures,” Proc. Combust. Inst. 24: 999–1006. 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: 41–67 . 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: 91–100. 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: 898–905. 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-96–008, 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. 275–89. Wu, C. H., and Kern, R. D. (1987). “Shock-Tube Study of Allene Pyrolysis,” J. Phys. Chem 91: 6291–6. 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: 43–57 . Xu, F., Lin, K-C., and Faeth, G. M. (1998). “Soot Formation in Laminar Premixed Methane/ Oxygen Flames at Atmospheric Pressure,” Comb. Flame 115: 195–209. Xu, F., Sunderland, P. B., and Faeth, G. M. (1997). “Soot Formation in Laminar Premixed Ethylene/Air Flames at Atmospheric Pressure,” Combust. Flame 108: 471–93. 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: 191–205. 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: 761–8.

153

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 mandated a smoke number (SN) measurement, which served to control the visible opacity 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 products 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 vehicle, 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 determine 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 CFM55–2C1 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).

For regulatory purposes, exhaust emissions measurements are required. For most sources, an emissions exit point can be defined, like the top of a smokestack or the end of an exhaust tailpipe. For an aircraft engine, the engine exit plane has been a reasonable place to define the exit point of emissions, and measurements within a half of a diameter of the engine exit plane have been the requirement for locating certification measurements for several decades (Figure  6.1). This ensures that the engine exit temperature defines the balance between particles and precursor gases, and volatile contributions to particles are necessarily in the gas phase at the engine exit How these aerosol particle precursors later evolve depends on the mixing of the exhaust with the ambient atmosphere that, in turn, depends on ambient temperature, pressure, and relative humidity, as well as any background pollutant gases. These ambient conditions can vary quite dramatically because airplanes can traverse the atmosphere from ground level, possibly in a polluted urban environment, up to as high as the lower stratosphere, depending on the airplane and its flight path.

156

Gaseous Aerosol Precursors

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 pressure species forming new particles and/or condensing on existing emitted particles, under certain circumstances, particles or their coatings may reevaporate, especially as the ­ aerosol precursor 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. 6.1.1  Precursor Emission Species 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 contains 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 aerosol precursors are a smaller subset of the SOx and organic species that have sufficiently 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
Table 6.1.  Sources of species that contribute to volatile PM mass in near-field aircraft plumes
Engine sources Fuel sulfur compounds: sulfides, disulfides, and benzothiophenes Fuel HCs: Liquid Aliphatic and Aromatic HCs Engine combustor exit species SO2 [gaseous] Engine turbine and early plume species SO2 [gaseous] ~1 percent SO3 (from partial oxidation of SO2) [gaseous] Possible further oxidation of key species Near-field plume species SO2 [gaseous] ~1 percent H2SO4 (from reaction of SO3 with H2O)

157

Lubrication System Venting

CO2 [gaseous] small amounts of CO [gaseous], Aliphatics, aromatics, and oxygenated organics [all gaseous in combustor] n/a

CO2 [gaseous], small amounts of CO [gaseous], Aliphatics, aromatics, and oxygenated organics [most gaseous], Lubrication oil

Lubrication oil

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 condensable 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 possibility 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 200–1200, 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 species 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 hydrogen 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 collectively 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 concentration 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, formaldehyde, 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 measurement under-counts the HC emissions and that it measures with a reduced sensitivity 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 levels 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
5 Total HC (FID) PTR-MS: methyl naphthalene naphthalene C5-benzene C4-benzene C3-benzene m/z 107 styrene hexanal phenol toluene m/z 83 high ralkenes benzene m/z 73 m/z 59 m/z 57 acetaldehyde m/z 43 m/z 41 TILDAS: C2H4 HCHO

159

Carbon emission ratio, ERC (ppb C/ppm CO2)

4

3

2

1

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 infrared (TILDAS) and mass spectrometric (PTR-MS) measurements. The most relevant condensable species are larger than even the naphthalenes presented here and include larger alkanes, PAHs, and oxygenates such as organic acids (credit NASA/TM-2006–214382 APEX Report).

low-vapor pressure. This combination, low concentrations and low-vapor pressures, makes their measurement difficult – sampling can be challenging because the measurement sensitivity must be high and the already small concentrations may be further reduced because of possible condensational losses in the sampling system. For all of these reasons, including measurement challenges and potential health and environmental impacts, recent emissions research has focused on the speciation of HC emissions (Figure 6.2), especially for better quantification of products of incomplete combustion and for identification of those emissions known to be important health hazards (FAA-EPA, 2009). 6.1.3  Gaseous Emissions, Gas-to-Particle Conversion in A/C Exhaust: SOx and Organics The processes by which particles form, grow, and interact have received increasing attention in the last several decades as their impact on human health and environmental 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 understanding 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 processing 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 (Kärcher 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 sulfuric acid can form new particles, as well as form coatings on emitted soot particles (Miake-Lye et al., 1994, Kärcher 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 (Kärcher 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 reactive 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 (Kärcher 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 (Kärcher et al., 1995; Brown et al., 1996; Kärcher 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 reinforced in these modeling studies, given the range of organic species concentrations available and the expected sulfuric acid levels with current fuel sulfur levels. 6.1.4  Organic Speciation and HAPs Based on the early recognition of the role of sulfuric acid in nucleating and contributing 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 measurements of aircraft PM were made to quantify PM emissions from the existing commercial fleet using an aerosol mass spectrometer (Anderson et al., 2005; Onasch et al., 2009; Timko et al., 2010). These compositional measurements clearly demonstrated that organic species not only make significant contributions to volatile PM, but typically dominate the volatile PM mass at low-power, near-idle engine operating 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 somewhat 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 individual 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 pollutants (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 emissions 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 concentrations 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
Compound 1,2,3-trim ethylbenzene 1,2,4-trim ethylbenzene 1,3,5-tnm ethylbenzene l,3-butadiened 1-decene 1-heptene 1-hexene 1-methyl naphthalene 1-nonene 1-octene 1-pentene 2-m ethyl-1-butene 2-m ethyl-1-pentene 2-m ethyl-2-butene 2-m ethyl-naphthalenee 2-m ethylpentane 3-m ethyl-1-butene 4-m ethyl-1-pentene acetaldehyded acetone acetylene acroleind benzaldehydee benzened butyraldehyde cl4-alkane cl5-alkane cl6-alkane cl8-alkane c4-benzene + c3-aroald c5-benzene + c4-aroald cis-2-butene cis-2-pentene crotonaldehyde dim ethylnapthalenes ethane ethylbenzene d ethylenef formaldehyde d,f Sum of all compounds CAS Mass Registry No.a fraction 526–73–8 95–63–6 108–67–8 106–99–0 872–05–9 25339–56–4 592–41–6 90–12–0 124–11–8 25377–83–7 109–67–1 563–46–2 763–29–1 513–35–9 91–57–6 107–83–5 563–45–1 691–37–2 75–07–0 67–64–1 74–86–2 107–02–8 100–52–7 71–43–2 123–72–8 No CAS No CAS No CAS No CAS No CAS No CAS 590–18–1 627–20–3 4170–30–3 28804–88–8 74–84–0 100–41–4 74–85–1 50–00–0 0.00106 0.00350 0.00054 0.01687 0.00185 0.00438 0.00736 0.00247 0.00246 0.00276 0.00776 0.00140 0.00034 0.00185 0.00206 0.00408 0.00112 0.00069 0.04272 0.00369 0.03939 0.02449 0.00470 0.01681 0.00119 0.00186 0.00177 0.00146 0.00002 0.00656 0.00324 0.00210 0.00276 0.01033 0.00090 0.00521 0.00174 0.15461 0.12310 Compound glyoxal isobutene/1-butene isopropylbenzene d isovaleraldehyde methacrolein methanold methylglyoxal m-ethyltoluene m-tolualdehyde m-xylene and p-xylened naphthalened n-decane n-dodecane n-heptadecane n-heptane n-hexadecane n-nonane n-octane n-pentadecane n-pentane n-propylbenzene n-tetradecane n-tridecane n-undecane o-ethyltoluene o-tolualdehyde o-xylened p-ethyltoluene p-tolualdehyde phenold propane propionaldehyded propylene styrened toluened trans-2-hexene trans-2-pentene valeraldehyde unidentifiedb CAS Mass Registry No.a fraction 107–22–2 106–98–9 98–82–8 590–86–3 78–85–3 67–56–1 78–98–8 620–14–4 620–23–5 108–38–3/ 106–42–3 91–20–3 124–18–5 112–40–3 629–78–7 142–82–5 544–76–3 111–84–2 111–65–9 629–62–9 109–66–0 103–65–1 629–59–4 629–50–5 1120–21–4 611–14–3 529–20–4 95–47–6 622–96–8 104–87–0 108–95–2 74–98–6 123–38–6 115–07–1 100–42–5 108–88–3 4050–45–7 646–04–8 110–62–3 NA 0.01816 0.01754 0.00003 0.00032 0.00429 0.01805 0.01503 0.00154 0.00278 0.00282 0.00541 0.00320 0.00462 0.00009 0.00064 0.00049 0.00062 0.00062 0.00173 0.00198 0.00053 0.00416 0.00535 0.00444 0.00065 0.00230 0.00166 0.00064 0.00048 0.00726 0.00078 0.00727 0.04534 0.00309 0.00642 0.00030 0.00359 0.00245 0.29213 1.00000

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

impact on the emissions of products of incomplete combustion (Corporan et al., 2007; Anderson et al., 2011). Most notably, reductions in fuel aromatic content cause significant reductions in soot production. In concert with reductions in soot, total organic emissions are also reduced, although individual species may actually rise (Anderson et al., 2011, appendix C), depending on the specific fuel considered. As a result of these effects, the amount of condensed organic PM measured in the plume is also reduced for these reduced aromatic fuels. The reduction in soot reduces the availability of soot surface area for condensation of low-vapor pressure gases, and the amount of organic in coatings on soot particles is reduced in concert with the reductions in surface area. The available low-vapor pressure organics are also reduced, so there is less mass available to add to particles of all types. In addition, these alternative fuels typically have very low sulfur levels, which limits the sulfate mass in the nucleation mode, which may also further limit the availability of surface area for organic condensation. 6.1.5  Environmental Issues and Reasons for Interest/Possible Future Regulatory Pressures The environmental consequences of PM are well established as a very significant contributor and are subject of much ongoing work to better understand and quantify the impacts. The issue is complicated by the fact that PM is not characterized by a single number or metric. Unlike a gas, where the overall emissions are well characterized by a single quantity, such as total emissions of CO2, PM’s properties and potential impact can be different for different types of PM emissions. For example, the same mass of PM, the metric first adopted in many countries to monitor and control PM emissions, will have dramatically different numbers of particles if the particles are 25 nm in diameter rather than 2.5 microns in diameter (at the upper limit for PM2.5): one million times more when the particles are 100 times smaller. In addition, the impacts of particles on human health may be very different if they are composed of toxic or carcinogenic species, and the impact on cloud properties may depend strongly on the surface properties and surface structure of the PM emissions. So, all particles are not the same, and research on their impacts and steps toward more effectively regulating PM emissions will depend on continuing to better understand the properties of the emitted PM. PM has been shown to play significant roles in environmental issues as diverse as human health, especially respiratory problems; local air quality and visibility; cloud formation and properties; and global radiative balance (Penner et al., 1999; Lee et al., 2010). Health effects have been correlated with exposure to PM mass and size (Oberdörster et al., 2005), but exploration of the role of structure and composition on mediating these health effects is only just beginning. Local air quality assessment has moved from using an upper size of 10 micrometers to quantify PM (PM10) to a 2.5 micrometer upper cutoff (PM2.5), but even that metric is a crude representation of how exposed populations ingest particles and does not account for variations in dependences on size or composition. Regulatory processes, especially in Europe, are

6.2  Sulfur Chemistry: SO2, SO3, H2SO4

165

moving to include a number-based metric in addition to mass, but particle composition and structure PM health effects are still far from understood well enough to be included in assessments or regulation. Local air quality concerns are driven largely by human health concerns, but also include visibility impacts and physical damage to property and structures due to exposure to pollution. These types of concerns are also driven by PM properties that go beyond simply the total mass of PM emitted. The impacts of soiling due to soot are much different from the impacts of acid particles damaging surfaces and changing the acidity of bodies of water, so composition of PM is important for LAQ as well. Size and morphology also determine lifetimes in the atmosphere and the range over which the PM is transported and deposited. For regional and global climate concerns, PM can play a role in radiative processes that affect the overall energy balance of the planet. Most important, PM can serve as nucleation sites for cloud formation, and much interest is focused on aviation PM’s potential role in affecting high-level cloud cover in the upper troposphere. Aviation is unique among major human emission sources in that the emissions are deposited in the upper atmosphere where clouds form and persist. Whether a linear cloud in the form of a contrail forms immediately behind an airplane or the PM emissions are left behind in a clear sky, the particles deposited in the atmosphere by an aircraft in flight have the potential to interact with other particles and the water vapor present at the flight level to affect how clouds form and persist and to affect the resulting cloud properties. Since clouds have such a significant role in determining the global radiative balance, much research is focused on better understanding the impact aviation has on global cloud cover. Current estimates indicate that aviation impacts on clouds could be quite significant (Penner et al., 1999; Lee et al., 2010), especially if projected growth in air traffic is realized in coming years, but aviation cloud effects are still very uncertain and definitive estimates of these impacts are still being pursued.

6.2  Sulfur Chemistry: SO2, SO3, H2SO4
6.2.1  Formation Mechanisms and Time Scales, Turbine Chemistry and Determination of S(VI) Fraction: Uncertainties and Bounds, Temperature/Pressure Effects 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 temperatures, 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 nozzle, the exhaust gases cool and expand and the thermodynamic state favors complete conversion to the higher oxidation state S(VI). However, the driving oxidative reactions with OH (and to a lesser extent atomic oxygen) are not sufficiently fast with

166
2000 0.0001%

Gaseous Aerosol Precursors

0.0005% 0.001%

0.00

3%

Temperature (K)

1500

5% 0.00 % 0 0.0 7 0.01%

0.1% 1000 1% 5%

500 0 10 20 Pressure (atm) 30 40

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) production 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 expansion 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 (Kärcher 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 percent (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

6.2.2  Fuel (Including Alternative Fuel) Effects 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 previously, 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 combustor 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 sustainability or carbon footprint perspective.

6.3  Organic Precursors Formation
6.3.1  Formation and Oxidation Kinetics of UHCs/HAPs The gaseous products of incomplete combustion, which include those low-vapor pressure organic species that contribute to volatile PM, are species that are sidetracked in the primary path of fuel being oxidized completely to carbon dioxide and water. Modern aircraft gas turbine combustors are highly efficient chemical reactors, especially at cruise powers, and most of the fuel ends up as carbon dioxide and water, so very small amounts of the fuel end up as these products of incomplete combustion. The resulting organics emitted can be (1) the result of intermediates in the oxidative pathway being kinetically quenched before completion as the combustion gases flow through the combustor or (2) products of a pyrolytic pathway wherein new organic species are formed in fuel-rich regions and are not consumed in the subsequent oxygen-rich zones of the combustor. The latter, pyrolytic products are related to the nonvolatile soot PM released from the engine in the particle phase, since they both arise from a pyrolytic pathway. The distinction between the oxidative and pyrolytic pathways is important since these two classes of species have different dependences on engine power. The oxidative pathway suffers the most inefficiencies at the lower-power settings  – low idle  – and the resulting organic emissions from incomplete oxidation are highest at low power. Pyrolytic processing is maximized at high power, with the maximum temperatures and pressures occurring in the combustor, and both soot and pyrolytically derived organics such as PAHs are more evident under such conditions. While both types of organics may contribute to organic in

168

Gaseous Aerosol Precursors

the particle phase, the pathway and detailed organic speciation may depend on the power setting and the chemical pathways dominant for those engine conditions. 6.3.2  Formation Mechanisms The chemical mechanisms discussed in Chapters 1, 5, and 7 for combustion chemistry and particle formation are the kinetic mechanisms that also form the organic species of interest for volatile particle contributions. Research in oxidative chemistry has focused on determining the overall oxidation correctly, with the total of the products of incomplete combustion primarily used until now as a metric of the combustion efficiency. Beyond qualitative agreement on the general types of species involved, thus far little emphasis has been placed on calculating the detailed speciation of the organic emissions correctly. And, indeed, relatively few measurements have been available to date for comparison with kinetic predictions until fairly recently. So the oxidative chemical mechanisms, both full and reduced, available until now for organic emissions are the same as those pursued for understanding combustion chemistry generally. Future work will look at the detailed speciation of organic emissions. This may put constraints on the implementation of mechanism reduction, if the details of which species representing incomplete combustion are important. Reduction that focuses on which reactions are locally important (Oluwole et al., 2007) may prove more helpful in this regard than globally simplified reaction schemes that might neglect species present in small concentrations but that might be important for PM or HAPs reasons. For pyrolytic mechanisms, the situation is similar, in that detailed kinetics are being developed for soot production (see Chapter 5). The low-vapor pressure emitted species of interest for volatile PM are intermediates in the soot production pathway, and are left as gaseous species when the combustion gases leave the reaction zones, not otherwise consumed by soot production or later oxidation. Similar to oxidative chemistry, soot formation research has focused on the final nonvolatile soot production and less emphasis has been placed on which intermediates might remain when the fluid leaves the reaction zones. So, similar to oxidative chemistry, future work on soot modeling might perform more detailed bookkeeping on what intermediates can be left behind and are not consumed by either soot production or later oxidation. 6.3.3  Fuel Effects, Including Alternative Fuels 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 properties 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. 6.3.4  Atmospheric Chemistry and Climate Questions 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 pertain 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 campaigns, 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 population exposed to them. While strong evidence suggests that aged atmospheric particles have organic PM contributions that all converge to similar, highly oxidized 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  Summary and Open Questions
6.4.1  Summary Gases emitted in the exhaust of aircraft gas turbine engines include a number of species 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 nonvolatile 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 production 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 coatings 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 different 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 emitted, 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 composition on the eventual impacts of the emitted PM. 6.4.2  Open Questions 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 sulfur 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 microphysical processes and engine and fuel technologies can affect the composition of particles formed in the aircraft plume and beyond.

6.4  Summary and Open Questions

171

A key source of condensable organic species has been identified as species vented from the engine lubrications system. Since this source has not been considered an emissions source previously, significant variation exists in emission levels between engine technologies and can constitute a major source of organic PM mass in some cases. If this contribution to volatile PM is a concern, the appreciation of lube oil as a precursor to volatile PM in aircraft plumes must be connected to aircraft lubrication system design. In parallel with increased understanding of gaseous PM precursor emissions, the potential climate and human health effects of the resulting PM must also be better understood. Recent advances in understanding volatile PM have been driven by the knowledge that climate and health may be affected by the PM emissions from aircraft. With those effects as a motivation, PM measurements and microphysical modeling now provide greater details about the PM characteristics than those previously available. Now the question turns back to the climate and health communities with regard to whether questions of composition, size, and morphology influence the expected impacts of PM on climate and health. If the mass is unaffected, but the number and composition of the particles are different because of different fuels or different engine technology, does that increase or decrease the predicted impacts? The advances in measurement and modeling allow a refinement in the impacts analysis and spur more detailed studies on the impact of particles beyond simply a mass-based metric. For climate, the most important questions revolve around cloud cover and radiative impacts of particles. The role of sulfuric acid and organic species in affecting the surface properties of the emitted particles may affect the cloud nucleation properties of aviation particles, and thus their role in contrails and subsequent clouds. As engine and fuel technologies evolve, it is important that impacts assessments properly account for the possible changes in PM properties that may be occurring. For local air quality and human health, it is likely that the various species associated with respirable particles will mediate the potential impact they may have on exposed populations. Given that the PAH emissions, and thus some of the key organic gaseous precursors for volatile PM, are affected by the fuel composition seen in some alternative fuels, the exposure near airports to such carcinogenic species may be controllable by changes to fuel composition. Better understanding of the health implications of PM composition and of the ability to tailor jet fuel to control organic gaseous precursors is needed. The related issue of lubrication oil contributions to volatile PM also needs further exploration. If exposure to lubrication oil, and the associated tricresylphosphate (TCP) additive, is a concern in and around airports, specific control of lube oil as an emission might warrant further research. Since lube oil has not been treated as an emission to date, limited data or impacts assessments have been obtained and there is significant variation across current engine technologies. Despite a large number of open questions, recent measurements and modeling have provided a fairly detailed picture of how gaseous precursor emissions interact with each other and with emitted soot particles to determine the particle properties

172

Gaseous Aerosol Precursors

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: 4380–92. 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/ TM–2011–217059. 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-2005–213783. 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: 1689–92. 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): 22939–53. 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: 2615–26. 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): 70–4. 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. Kärcher, B., Peter, T., and Ottmann, R. (1995). “Contrail Formation – Homogeneous Nucleation of H2SO4/ H2O Droplets.” Geophysical Research Letters 22(12): 1501–4. Kärcher, 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): 29379–86.

References
Kärcher, 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 during the Aircraft Particle Emissions eXperiment (APEX) 1 to 3.” Environmental Science Technology 45: 3415–21. 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): 949–58. 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): 10314–15. 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: 4678–734. 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-2000–00182. 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 18–20, DLR-Mitteilung 94–06, Deutsches Zentrum für 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: 4135–47 . 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: 4625–41. Oberdörster, G., Oberdörster, E., and Oberdörster, J. (2005). “Nanotoxicology: An Emerging Discipline Evolving from Studies of Ultrafine Particles.” Environmental Health Perspectives 113(7): 823–39. 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: 127–46. 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): 1121–37 . 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: 3603–12.

173

174

Gaseous Aerosol Precursors
Schumann, U., Arnold, F., Busen, R., Curtius, J., Kärcher, 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 1–7 .” 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(10–11): 944–55. 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): 111–17 . 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): 1109–12. 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: 417–30. 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-2006–214382. 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: 897–905. 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: 3261–72. 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 02–03, ACRP Report 7 , ISBN 978–0–309–11745–6, 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 Aviation’s Contribution to Current and Future Fine Particulate Matter in the United States.” Atmospheric Environment 45: 3424–33. 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: 912–18. 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): 779–95.

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 combustion 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 combustion 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 hydrocarbon 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 environmental 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 pollutants (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 technologies (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 underlying 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 presented 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 hydrocarbon 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.

7.2  Hydrocarbon Oxidation and CO Formation
Natural gas and petroleum-based liquid hydrocarbon fuels (e.g., fuel oil and jet fuels) are widely used in gas turbine applications. In recent years, alternative fuels such as syngas and biofuels have also attracted considerable interest because of the concern regarding combustion-generated CO2 emissions and the need for national energy security. Natural gas, widely used in stationary power generation, is predominantly composed of methane (CH4) with lesser amounts of ethane (C2H6) and propane (C3H8). Petroleum-based liquid hydrocarbon fuels, such as kerosene-type jet fuels, consist of hundreds of chemical components ranging from carbon number C7 to C17. These components can generally be categorized into four chemical classes: normal paraffins, iso-paraffins, cyclo-paraffins, and aromatics (Edward and Maurice, 2001). Figure 7.1 shows the chemical class composition of a commercial jet fuel, Jet A1, as function of carbon number.

7.2  Hydrocarbon Oxidation and CO Formation
12 10 8 6 4 2 0 i+c-Paraffins n-Paraffins Aromatics

177

Volume (%)

C7

C8

C9 C10 C11 C12 C13 C14 C15 C16 C17 Carbon number

Figure 7.1.  Chemical class composition distribution for petroleum-derived commercial aviation fuel Jet A-1. (Reproduced from Edwards, 2007; originally published by ASME.)

A thorough understanding of the combustion chemistry of the multicomponent gas turbine fuels is important for the reduction and control of the pollutant emissions. Combustion is fundamentally a highly exothermic chemical process governed by a series of chain reactions involving radical species in which fuel molecules are consumed. This process eventually leads to the formation of major products such as CO2 and H2O along with minor pollutant species. With the advances in computing capabilities, researchers have increasingly used detailed chemical kinetic models to study the combustion and pollutant formation behavior of fuels. Accurate models for chemical kinetics and turbulent combustion are required to account for turbulence-chemistry interactions over a wide range of chemical time scales in order to fully resolve the numerous chemical species present in the system. As practical petroleum-based liquid gas turbine fuels consist of hundreds of chemical species, a set of representative chemical components, known as “surrogate fuels,” is often used to develop chemical kinetic models for liquid gas turbine fuels (Dagaut, 2002; Violi et al., 2002; Colket et al., 2007; Gokulakrishnan et al., 2007; Dooley et al., 2010). The number of species and reactions in a detailed kinetic model will increase dramatically as the number of carbon atoms in the fuel rises (Lu and Law, 2009). As a result, a detailed surrogate kinetic model of a multicomponent liquid hydrocarbon fuel can include a collection of thousands of elementary reactions involving hundreds of species. However, a reaction mechanism of this size is too computationally expensive to be of practical use in detailed flow-field modeling for combustor design. Hence, one of the daunting tasks for computational fluid dynamics (CFD) simulations of practical gas turbines is the prediction of pollutant emissions using reduced chemical kinetic models. Based on the combustion conditions, hydrocarbon fuel oxidation can be classified into low-, intermediate-, and high-temperature oxidation regimes. Figure 7.2 shows a schematic of the main reaction pathways for the chain-branching route of

178

NOx and CO Formation
Fuel RH OH H-atom abstraction RH HO2 + Alkene R + H2O2 R Branching β-scission Alkene + small alkyl radicals (C1, C2, C3) OH + OH H + Alkene Intermediate temperature oxidation Branching O2 OH + O High temperature oxidation

Products + OH Propagation O2 RO2 QOOH O2 O2QOOH

OH + ROOH Branching RO + OH Low temperature 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 generation 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 formation at typical gas turbine operating conditions are influenced by high-temperature fuel oxidation pathways, combustion kinetics of hydrocarbons for this regime is presented in this chapter. The low-temperature kinetic pathways are relatively more complex than the high-temperature oxidation route, since the formation of alkylperoxy (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

conducted concerning the high-temperature oxidation of hydrocarbons (Warnatz, 1984; Westbrook and Dryer, 1984). The framework of a chemical kinetic model for high-temperature oxidation follows a hierarchical structure where long-chained hydrocarbon molecules (C4 species and larger) thermally and chemically decompose into smaller hydrocarbon molecules (i.e., C1 and C2 species) through initiation, propagation, and termination reactions. The chemical kinetics of the H2-O2 system forms the foundation for any hydrocarbon oxidation as it is essential for the generation of the combustion radical pool. Chaos and colleagues (2010) have provided a detailed review of the current status of the chemical kinetic modeling of H2-O2 in the context of syngas combustion. The H2-O2 reaction subset includes the primary branching reaction, R1, critical for the high-temperature oxidation of any hydrocarbon, and the recombination reaction R2. H + O2 <=> OH + O  H + O2 + M <=> HO2 + M  R1 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 (symbolically shown as R), RH + X <=> R + XH  R3

where X represents species such as OH, H, O, HO2, O2, or CH3. The alkyl radicals, formed in reaction R3, will undergo chain reactions with other combustion radicals and beta-scission reactions to form smaller hydrocarbon molecules (e.g., CH3, C2H4, C2H5, etc.) (Westbrook and Dryer, 1984). These smaller hydrocarbon molecules then react with oxygenated combustion species (i.e., O2, O, OH, and HO2) to form formaldehyde (CH2O), one of the main intermediate species in hydrocarbon fuel oxidation. Formaldehyde undergoes thermal and chemical decomposition to form the formyl radical (HCO). The thermal dissociation of CH2O proceeds via the following channels: CH 2 O + M<=>HCO + H + M  CH 2 O + M<=>CO + H 2 + M  R4 R5

where M represents a third-body collisional species. Formaldehyde also reacts with H, O, OH, and HO2 to form HCO: CH 2 O + H<=>HCO + H 2  CH 2 O + O<=> HCO + OH  CH 2 O + OH <=> HCO + H 2 O  CH 2 O + HO2 <=> HCO + H 2 O2  R6 R7 R8 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. HCO + O2 <=> CO + HO2  R10

HCO + M <=> H + CO + M  R11 Yetter and colleagues (1991a, 1991b) experimentally investigated moist CO oxidation in a high-pressure plug-flow reactor to develop a comprehensive chemical kinetic model for CO oxidation. This has served as the basis for the CO oxidation reaction subset in the chemical kinetic mechanisms for hydrocarbon oxidation. Carbon monoxide then reacts with OH, HO2, O, and O2 to form CO2 via reactions R12 to R15, respectively. CO + OH <=> CO2 + H  R12 CO + HO2 <=> CO2 + OH  R13 CO + O (+M)<=> CO (+M) R14 2  CO + O2 <=> CO2 + O  R15 The exothermic reaction R12 is the main pathway for the oxidation of CO while producing an H atom available to react with O2 to promote the branching reaction R1. Reactions R13 to R15 are relatively slow compared to reaction R12 in the high-temperature regime; hence, they play a smaller role in CO oxidation. Chemical kinetic mechanisms are often validated against laboratory-scale experimental data in order to minimize the uncertainty in the reaction rate parameters. For example, experimental measurements from shock tubes, laminar flames, and flow reactors (i.e., plug-flow reactors (PFR) and perfectly stirred reactors (PSR)) are valuable sources of validation data. Very often, flow reactor experiments are performed with dilute fuel/oxidant mixtures in order to minimize the uncertainty caused by the effect of heat release on species measurements. For this reason, accurate measurement of CO from practical combustion systems with high heat release remains a difficult task. Schoenung and Hanson (1981) and Nguyen and colleagues (1995) have demonstrated the discrepancy between in situ techniques (e.g., tunable diode laser spectroscopy) and extractive sampling techniques (e.g., non-dispersive infrared (NDIR) analyzers) for measuring CO. This is because of the propensity of the CO + OH/HO2 reactions to oxidize CO into CO2 in the sample probe, affecting the accuracy of the measurement (Gokulakrishnan et al., 2012). A significant experimental effort with well-designed probes is needed to “freeze” the chemistry at the point of extraction for accurate measurements (Colket et al., 1982). Although it appears the reaction pathways for the production and consumption of CO are relatively straightforward, the prediction of CO in practical gas turbine combustors is rather complicated, especially in non-premixed combustion systems. In aeroderivative gas turbine engines, spray atomization, fuel evaporation, and turbulent mixing of dilution air with multiple recirculation zones, coupled with radiative heat transfer, adds significant complexity to the prediction of CO formation (Sturgess et al., 2005).

7.2  Hydrocarbon Oxidation and CO Formation
PB= 300 Premix burner fuel mass flow Total fuel mass flow NOx emission
NOx, CO emissions (ppm)

181

CO emission

200 PB = 0% PB = 92% 100 PB = 100%

Flame extinction 0 0 1

Operating range 4 5

3 2 Air/Fuel ratio 1600 1000 2200 1200 4000 2500 3000 2000

800 °C 1500 °F

Figure 7.3.  NOx and CO emissions as a function of normalized air-to-fuel ratio, λ, (i.e., actual air-to-fuel ratio (AFR) to stoichiometric AFR) for diffusion flames (PB = 0%), partially premixed (PB = 92%), -and fully premixed (PB = 100%) combustion in gas turbines (reproduced from Maghon et al., 1988 with the permission of Illinois Institute of Technology). The shaded area is the optimal region for minimizing both CO and NOx production.

Maghon and colleagues (1988) experimentally investigated CO and NOx emission characteristics of a natural-gas-fired combustion system with varying degrees of fuel-air premixing. Figure 7.3 (adopted from Maghon et al., 1988) depicts the trade-off between CO and NOx formation as a function of air-to-fuel ratio (λ). Lean, premixed combustion systems reduce NOx emissions by operating close to lean blowout limits in natural-gas-fired stationary gas turbines (Gokulakrishnan et al., 2008). However, lean, premixed systems that approach blowout conditions are susceptible to combustion instabilities and can produce high CO and UHC emissions (Gokulakrishnan et al., 2008). The need to accurately model fuel-air mixing and the coupling of turbulence and chemistry is paramount when predicting CO levels from combustion devices. Increasingly, large chemical kinetic models have been coupled with various turbulent combustion models for the simulation of turbulent flames (Hilbert et al., 2004). However, the high degree of stiffness caused by a wide range of chemical time scales (varying over several orders of magnitude) in hydrocarbon oxidation makes it computationally expensive to fully resolve all of the species in a detailed kinetic model for the simulation of practical combustion systems. Figure 7.4 shows the chemical kinetic time scales of selected species, including some of the important

182
1.0E+03

NOx and CO Formation
2500 Temperature Chemical time scale (sec) 1.0E+00 2000 CO H2O2 CH3 O CH CH2O HO2 H OH HCO Temperature (K)

1500

1.0E-03

1000 1.0E-06

500

1.0E-09 2.0

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, premixed laminar flame zone (highlighted region) for an atmospheric pressure CH4/air flame (equivalence ratio = 0.9) using the GRI3.0 mechanism (Smith et al.).

combustion radicals, within the flame zone of a one-dimensional CH4/air flame as a function of flow residence time from the burner surface. The adiabatic laminar premixed flame simulation was performed in Cantera (Goodwin) using the GRI-3.0 chemical kinetic mechanism (Smith et al.). The chemical time scale, τi, can be com­ puted as (Gou et al., 2010):

τi =

Ci ωi

(7.1) 

where Ci is the concentration of species i and ωi is the rate of consumption of species i. The chemical time scales provide a comparative measure to identify the fastand slow-forming species, and hence, indicate the numerical stiffness of the system. It can be noted that CH2O and CO exhibit a wide spectrum of time scales within the flame zone ranging from 10–7 to 10–2 s and 10–4 to 10 s, respectively. However, HCO and OH radicals show a narrow spectrum of time scales between 10–8 to 10–7 s and 10–6 to 10–5 s, respectively. This variation in the time scale spectrum indicates the consumption of HCO (the main precursor for the formation of CO) is relatively faster than the consumption of CO (the precursor for the formation of CO2). The formation of CO mainly occurs within the flame zone because of the presence of fast-forming species such as HCO via reactions R10 and R11, while the slower conversion process of CO to form CO2 starts in the flame zone and extends to the post-flame zone, mainly via reaction R12. Therefore, it is essential to have sufficient residence time to allow for CO burnout and conversion to CO2 in the gas turbine combustor design. Important design considerations for minimizing CO quenching include the placement of dilution air introduction points and combustor length prior to expansion into the turbine.

7.3  Formation of Nitrogen Oxides

183

7.3  Formation of Nitrogen Oxides
More than forty years of fundamental and applied research has generated a large collection of literature on the combustion chemistry of NOx and N2O formation, including several comprehensive review articles (Miller and Bowman, 1989; Correa, 1992; Hayhurst and Lawrence, 1992; Kramlich and Kinak, 1994; Dean and Bozzelli, 1999; Glarborg et al., 2003). Continued progress in experimental and computational capabilities has greatly enhanced the understanding of the chemical mechanisms involved in NOx formation and destruction in practical combustion systems. Four different chemical pathways produce NO from molecular nitrogen present in the combustion air: (a) thermal NO, (b) prompt NO, (c) the N2O route, and (d) the NNH route. In addition, NOx can be generated from chemically bound fuel-nitrogen, commonly present in many solid and liquid fuels (Glarborg et al., 2003). Therefore, nitrogen chemistry plays a crucial role in understanding the different pathways for NOx formation and destruction, as well as in developing pollution control technologies (Correa, 1998). Hanson and Salimian (1984) reported an early review of the reaction rate parameters of elementary reactions involved in nitrogen chemistry, while Miller and Bowman (1989) provided a comprehensive review and discussion on chemical kinetic modeling of nitrogen oxides. Dean and Bozzelli (1999) reported subsequent advances in this area. A number of laboratory-scale experimental data sets from fluidized bed reactors (Hayhurst and Lawrence, 1996; Gokulakrishnan and Lawrence, 1999; Lawrence et al., 1999), plug-flow reactors (Johnsson et al., 1992; Kristensen et al., 1996; Allen et al., 1997; Mueller et al., 1999), perfectly-stirred reactors (Steele et al., 1995; Rutar et al., 1998), and flame structure measurements (Drake et al., 1990; Naik and Laurendeau, 2004; Harrington et al., 1996; Klassen et al., 1995) have been conducted to investigate the various chemical processes involved in NOx formation and destruction under different combustion conditions. To demonstrate NOx formation from molecular nitrogen introduced with the oxidizer stream, this chapter presents laminar premixed hydrocarbon flame simulations using the experimental conditions of Drake and colleagues (1990). Figure 7.5 compares one-dimensional laminar flame simulation results with the experimental data of Drake and colleagues (1990) for NO obtained in a burner stabilized ethane premixed flame at 6 atm and an equivalence ratio of 0.9. The figure also shows the contribution of various NO formation pathways to the total NO. The simulation was performed in Cantera (Goodwin) using the chemical kinetic mechanism of Gokulakrishnan and colleagues (2012) with the flame temperature profile reported by Drake and colleagues (1990). Since the peak flame temperature is less than 1800 K, the contribution of thermal NO to the total is relatively small. The sum of contributions of the prompt-NO and N2O routes account for approximately 90 percent of the total NO production at 1 cm from the burner surface. It is also noteworthy that these results show as the residence time increases (indicated by the distance from the burner surface), the contribution from the thermal-NO route continues to increase. Therefore, in gas turbine combustor design, an optimal residence time

184
8

NOx and CO Formation
2000 Temperature Total NO 6 Exp. Data-NO
NO (ppmv)

1500
Temperature (K)

4 N2O Route Prompt-NO 2 Thermal-NO 0 0.0 NNH Route

1000

500

0.2

0.4

0.6

0.8

0 1.0

Distance (cm)

Figure 7.5.  Contributions of different NO formation pathways from a simulation of the burner-stabilized premixed flame configuration of Drake and colleagues (1990)with 6 mole% C2H6/23 mole% O2 in N2 at an equivalence ratio of 0.9 and at 6 atm pressure. Symbols denote the experimental measurements of Drake and colleagues (1990) for NO, and the lines denote the modeling results.

needs to be maintained to minimize post-flame NO production via the thermal-NO pathway, while still allowing enough time for CO burnout. The equivalence ratio of the fuel-air mixture also has a significant impact on NO formation. Figure 7 .6 shows computational results for the total NO production and the contribution of different NO pathways to the total NO as a function of equivalence ratio for the same experimental conditions shown in Figure 7 .5. The use of a fixed flame temperature profile will assist to highlight the role of equivalence ratio on NO formation by isolating the effect of temperature. Otherwise, the thermal NO will increase dramatically with flame temperature and mask the contributions of the other chemical pathways. It can be noted in Figure 7 .6 that the contribution from the prompt-NO pathway increases with the equivalence ratio, though it will eventually peak in fuel-rich flames. Klassen and colleagues (1995) experimentally demonstrated that the equivalence ratio at which the peak occurs varies depending on the pressure: the higher the pressure, the leaner the equivalence ratio at which the peak occurs. For example, in a CH4/O2/N2 mixture at 3 atm, the maximum NO formation was observed around an equivalence ratio of 1.3, while the peak shifts to an equivalence ratio of 1.0 at 14.6 atm (Klassen et al., 1995). A discussion on various chemical pathways for NO formation and their importance at gas turbine relevant conditions is presented in the following sections. 7.3.1  Thermal NO The relative importance of an individual chemical route to the overall NOx emissions 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
100 Percent contribution to total NO Prompt N2O Route NNH Route Thermal Total NO 15

185

80

12 Total NO (ppmv)

60

9

40

6

20

3

0

0.7

0.8

0.9

1.0

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 temperature 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 concentration of oxygen in the post-flame zone using Equation 7.2 (Bowman, 1975): d[NO]
16 −0.5 dt = 6 × 10 Teq exp

(

−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 combustion 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 investigated 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, premixed combustion systems operating at high inlet pressures and temperatures. 7.3.2  N2O Pathway 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  N 2 O + O<=> NO + NO  N 2 O + H <=> NO + NH  R19 R20 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  N 2 O + H <=> N 2 + OH  R22 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: N 2 + CH <=> HCN + N  R24

However, quantum chemists have disagreed about the plausibility of reaction R24, as it does not conserve electron spin. Miller and colleagues (2005) have discussed this issue in detail. They concluded, based on the work of Moskaleva and Lin (2000), that NCN (via reaction R25) is a possible intermediate species that can conserve electron spin. N 2 + CH <=> NCN + H  R25

Subsequently, NCN will undergo fast oxidation with O and OH radical species to form NO via reactions R26 and R27. NCN + OH <=> NO + HCN  R26 NCN + O<=> NO + CN  R27 In addition, NCN can further react with H to form N and HCN species that can lead to NO formation. NCN + H <=> HCN + N  R28 N R29 2 + C <=> CN + N  The nitrogen atoms produced via reactions R28 and R29 can react with O2 and OH to enhance thermal-NO formation via reactions R17 and R18, while the cyano compounds react with various oxygen-containing species to form NO. Prompt NO does not generally form in any significant quantity in the post-flame zone because the concentration of hydrocarbon radicals is quite small away from the flame front. As shown in Figure 7.6, the formation of prompt NO in the flame zone increases under fuel-rich conditions because of the availability of larger quantities of hydrocarbon radicals. 7.3.4  NNH Route 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 particularly 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 molecules (reactions R32 and R33) and prevent reaction R31 from contributing to NO formation (Klippenstein et al., 2011). In a recent publication, Klippenstein and colleagues (2011) discussed in detail the role of the NNH species in NO formation as well as NO destruction in the thermal-deNOx process. 7.3.5  Fuel-bound Nitrogen 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 percent (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 nitrogen 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 nitrogen is converted to HCN and/or NH3 intermediates, which then react with combustion 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. HCN + O<=> NCO + H  HCN + OH <=> CN + H 2 O  CN + OH <=> NCO + H  NCO + O<=> NO + CO  R34 R35 R36 R37 NNH + O<=> N 2 + OH  NNH + O<=> N 2 O + H  R32 R33

7.4  Influence of Vitiated Air on Fuel Oxidation
Table 7.1.  Fuel-Bound nitrogen content of solid fuels
Fuel Biomass Peat Coal Household waste Sewage sludge
Source:  Dagaut et al., 2008

189

Nitrogen content (wt%) 0.1–3.5 0.5–2.7 0.5–2.5 0.5–1.0 2.5–6.5



NCO + OH <=> NO + CO + H 

R38

During oxidation, NH3 breaks down to form NH2, NH, and N via reaction route R39. NH j (j= 3 to 1) + O/H/OH <=> NH i (i = 2 to 0) + OH/H 2 /H 2 O  R39

NH and N radicals undergo further reaction to form NO via reactions R40 and R41. NH + O<=> NO + H  N + O2 <=> NO + O  R40 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 responsible for the destruction of NO in the thermal-deNOx and reburning processes, respectively. 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.

7.4  Influence of Vitiated Air on Fuel Oxidation
Many practical applications combine exhaust gas with fresh air to create an oxidizer stream known as vitiated air. This is utilized in combustors through exhaust gas recirculation (Correa, 1998) or through product gas entrainment (Fleck et al., 2000) to increase thermal efficiency and/or reduce emissions. A detailed discussion on the effect of exhaust recirculation on NOx emissions can be found in the chapter entitled “Emissions from Oxyfueled or High-Exhaust Gas Recirculation Turbines.” A brief description of the effect of vitiated air on fuel oxidation is proved in this section. Vitiated air generally contains combustion products such as CO2, H2O, CO, NOx, and unburned hydrocarbons, along with O2 and N2. A detailed investigation of the effect of vitiated air on the ignition (Fuller et al., 2009) and flame propagation

190
Thermal NO
pa th

NOx and CO Formation
N2 +H +O+M Prompt NO
O N2 th pa

NN H

+O

NNH +O +O NO

N2O +H +OH +O

+CH

N +H +OH NH

+OH +O2

NCN

+OH +O +H NCO (HNCO) (NH3) Fuel N (HCN) +O2 CN

Figure 7.7.  Major NOx formation reaction pathways and their interconnections.

(Fuller et al., 2012) properties of hydrocarbon fuels showed that NO has a significant 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 oxidizer 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 reduction 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 produce CH3O via reaction R44. NO + HO2 <=> NO2 + OH  R42

7.5  NOx Abatement Strategies

191



NO2 + CH 3 <=> NO + CH 3O  NO + CH 3O2 <=> NO2 + CH 3O 

R43 R44

Although these reaction pathways have been mainly validated for methane oxidation (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 oxidation. Reactions R42 to R44 are fuel-independent pathways to promote oxidation by NOx irrespective of the type of hydrocarbon molecule. In addition, NOx interactions 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 hydrocarbon 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 oxidation 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 termination 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

7.5  NOx Abatement Strategies
Nitrogen oxide emissions can be controlled either during the combustion process (i.e., in situ combustion control) or after the combustion process is complete (i.e., post-combustion control). Lean, premixed combustion has become an increasingly popular approach to keep the NOx level below 10 ppm from stationary gas turbines that operate on natural gas. This is achieved by significantly reducing NOx formed

192
1.0 Relative reduction in ignition delay time

NOx and CO Formation

20 mole% O2 0.8 12 mole% O2

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 stoichiometric 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 et al., 2005) and lean prevaporized, premixed (LPP) combustion (Gokulakrishnan et al., 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 temperatures (>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 operating 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. 7.5.1  Thermal-deNOx Process The thermal-deNOx process is a non-catalytic, gas phase NOx reduction technique originally invented and patented by Lyon (1975). This technique uses ammonia in the post-combustion zone to induce reactions that can reduce NOx production. Numerous works (Lyon and Benn, 1978; Kjaegaard et al., 1996; Miller and Glarborg, 1999; Schmidt, 2001) have been devoted to the understanding of the chemistry involved and the influence of operating variables on the thermal-deNOx process efficiency. In the presence of excess O2, the addition of NH3 was found to convert the NO into N2 over a narrow window of temperatures in the range of 1100 to 1400 K (Kjaegaard et al., 1996). In this temperature window, NH3 reacts primarily with OH to form the NH2 radical via reaction R46, which then reacts with NO to form N2 and H2O via reaction R49. In addition, NH3 reacts with the radical species O and H to form NH2 via reactions R47 and R48. NH 3 + OH <=> NH 2 + H 2 O  R46 NH R47 3 + O<=> NH 2 + OH  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, reactive 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). NH 2 + NO <=> NNH + OH  NNH <=> N 2 + H  R50 R51

The H atom formed in reaction R51 promotes the chain-branching reaction R1 to produce OH and O. However, excess availability of NNH can scavenge O atoms to form NO via reaction R31. Also, a termination reaction between NNH and O2 (R52) can reduce the generation of the reactive radicals such as O, H, and OH, which are necessary to sustain the initiation reactions of NH3 via reactions R46 through R48.

194

NOx and CO Formation



NNH + O2 <=> HO2 + N 2 

R52

Therefore, in the thermal-deNOx process, it is important to maintain a balance between the recombination reaction R49 and the chain-branching reaction sequence (i.e., R50 and R51) to sustain the NO reduction by NH3 while minimizing the NO formation via reaction R31. The branching fraction (α) of the NH2+NO reactions (R49 and R50) is given by Equation 7.3 and must stay within an optimal range to self-sustain the deNOx process.

α=


k50 k49 + k50 

(7.3)

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 additional NO from NH3. However, this window can be shifted toward higher temperatures 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 burnout 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 hydrocarbon 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

Pioneering work of Myerson and colleagues (1957) on the ignition of propane in the presence of NO2 led to the finding that hydrocarbon radicals can consume NO, and hence, the concept of reburning chemistry. Wendt and colleagues (1973) and Myerson (1974) performed early experiments on reburning to investigate the interaction of mixtures of hydrocarbon fuels with NO. Subsequently, numerous works were reported in the literature (as summarized in Smoot et al., 1998) to provide a better understanding of the complexities of the chemical kinetic processes using a variety of hydrocarbon fuels. The plug flow reactor experiments of Glarborg and colleagues (1998) and jet-stirred reactor experiments of Dagaut and colleagues (1998, 2000) showed that the NO reduction potential depends highly on the type of fuel used in the reburning process. Methane was found less effective as the reburning fuel compared to other hydrocarbons fuels such as C2H6, C3H8, and nC4H10, which produce significant amounts of acetylene (C2H2). Acetylene is one of the main sources of the HCCO radical (via reaction R53) in any hydrocarbon oxidation process (though less important in methane oxidation). C 2 H 2 + O <=> HCCO + H  R53

In the absence of significant concentrations of oxygenated radicals (which can occur under fuel-rich conditions), HCCO reacts with NO via reactions R54 and R55 (Miller et al., 1998; Vereecken et al., 2001): HCCO + NO <=> HCNO + CO  R54 HCCO + NO <=> HCN + CO R55 2  Most of the HCN formed in reaction R55 is converted to N2 while further oxidation of the HCNO from R54 can lead to the production of NO (Miller et al., 2003). This will reduce the overall effectiveness of the reburning process. Therefore, it is important to have the proper branching fraction between R54 and R55 to correctly predict the reburning chemistry. For methane, the most important route for reducing NO is promoted by reaction R56: CH 3 + NO <=> HCN + H 2 O  R56

In addition, the interaction between NO and other hydrocarbon radicals, such as CH, can contribute to the reduction of NO via reaction R57 (Glarborg et al., 1998; Dagaut et al., 2000). CH + NO <=> HCN + O  R57 Further experimental data is needed to reach a consensus on a branching fraction that works for all hydrocarbon fuels. A detailed discussion on this subject can be found in other references (Glarborg et al., 1998; Miller et al., 1998; Dagaut et al., 2000; Vereecken et al., 2001; Frassoldati et al., 2003).

7.6  Effect of Pressure on CO and NOx Formation
Most gas turbine combustors operate at elevated pressures to maximize the output of the device. For example, many stationary utility gas turbines operate at pressures

196

NOx and CO Formation

around 15 atm, while aeropropulsion gas turbines operate at pressures around 40 atm (Correa, 1992). In general, increasing pressure tends to accelerate the overall ignition and oxidation of hydrocarbon fuels. Increasing pressure will affect the combustion process in terms of thermodynamics, transport, and chemical kinetics. Increasing pressure also impacts the thermal diffusion and mass diffusion properties of the reactants by increasing the density of the mixture. The effect of pressure on flame propagation is largely controlled by the reaction kinetics, because transport properties such as the density-weighted diffusive terms are generally pressure independent (Law, 2006). Similarly, pressure will play a role in the formation of pollutants through its effect on the reaction kinetics. Increasing pressure from 1 to 10 atm will raise the adiabatic flame temperature by roughly 50 K for a stoichiometric CH4/air mixture because of a slight increase in the CO to CO2 conversion rate. At equilibrium, the CO concentration should be proportional to P-0.5 (Bhargava et al., 2000). This pressure dependence was experimentally demonstrated by Bhargava and colleagues (2000) and Rink and Lefebvre (1989), where CO emissions at the exit of the combustor for a given equivalence ratio decreased as the combustor operating pressure increased. Several factors, including the effectiveness of the fuel/air mixing, the equivalence ratio, and the residence time, play important roles when accounting for the effect of pressure on CO and NOx levels in gas turbine combustion. As a result, the role of pressure on the formation of CO and NOx is not straightforward in practical gas turbine combustors because of complex interactions between combustion chemistry and flow-field variables such as turbulent mixing. Therefore, numerous groups have developed physics-based correlations over the years for predicting NOx and CO emissions from gas turbine combustors. Among the most notable of these correlations were those developed by Lefebvre (1984, 1985) and Mellor and co-workers (Mellor, 1976; Connors et al., 1995; Newburry and Mellor, 1996). Rizk and Mongia (Rizk and Mongia, 1993, 1995; Mongia, 2010a, 2010b) extended many of these formulations, developing semi-empirical expressions that have been extensively tested against the measured output of industrial aero gas turbines. A number of studies have investigated the role of pressure on NOx production using simplified laboratory-scale experimental systems. A series of counterflow laminar flame experiments (Naik and Laurendeau, 2004) of methane/air diffusion and partially premixed flame showed that the peak NO concentrations decreased with increasing pressure up to 15 atm. But NO concentrations in partially premixed flames showed less pressure dependence than those measured in diffusion flames (Naik and Laurendeau, 2004). A review of turbulent, non-premixed flame data and combustor emission measurements has shown that NOx emissions for conventional combustors have a dependence that scales with P0.5 to P0.8 (Correa, 1992). The pressure dependence in these diffusion and partially premixed flames is mainly due to the dominant role the thermal-NO pathway plays in the formation of NO. The formation of NO via the thermal pathway, at a given equivalence ratio, is limited by the concentration of O atoms. At equilibrium conditions, O-atom concentration will scale with the P0.5 (Bhargava et al., 2000). However, other variables such as super-equilibrium

7.6  Effect of Pressure on CO and NOx Formation
16 2000

197

NO (ppmv, wet at 15% O2)

12 NO at 14.6 atm 8 NO at 6.1 atm 4 NO at 1 atm

1750 Temperature (K)

1500

1250

0 0.0

0.2 0.4 0.6 Distance from burner surface (cm)

1000 0.8

Figure 7.9.  Model comparison of Gokulakrishnan and colleagues (2012) with the experimental 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 conditions (with equivalence ratios less than 0.7) found that NOx emissions have very little dependence on pressure. Bhargava and colleagues (2000) performed an experimental 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 influence of pressure on NOx formation. Figure 7.9 shows the effect of pressure on an NO profile obtained in burner stabilized premixed flame experiments of Klassen and colleagues (1995), which were further 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
100

NOx and CO Formation

Prompt-NO
Percent contribution to total NO 80

60

Thermal-NO

40

N2O Route
20

NNH Route

0

1

3

6 9 Pressure (atm)

12

15

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 experimental 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 pathway 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 laminar flame simulation results shown in Figure 7.11 were performed using the chemical kinetic mechanism of Gokulakrishnan and colleagues (2012). The figure also

7.6  Effect of Pressure on CO and NOx Formation
20 φ = 0.8 (P0.42) 15 φ = 0.7 (P0.35) φ = 0.6 (P0.39) 5

199

NO (ppmv, wet at 15% O2)

10

0

0

5

10 Pressure (atm)

15

20

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.

provides approximate pressure-dependant factors for each equivalence ratio calculated based on the experimental data. The pressure dependence of the contribution of different NO formation pathways at equivalence ratios of 0.7 and 0.8 is very similar to that shown in Figure 7.10 for an equivalence ratio of 0.6. However, the contribution of prompt NO to the total NO increases with equivalence ratio. In addition, the increase in flame temperature and the residence time (due to variations in the mass flow rates used in the experiments at different pressures [Klassen et al., 1995]) with increasing pressure significantly enhances the NO formation via thermal NO. These results highlight that the pressure dependence of NO formation is determined by various factors, including local equivalence ratio, heat loss from the flame zone, and residence time. Figure 7.12 shows the NOx emissions obtained in a lean, premixed swirl burner experiment for fuel oil #2, which had 0.04 wt percent fuel-bound nitrogen, at varying equivalence ratios as a function of pressure (Gokulakrishnan et al., 2008). The results are also compared with NOx emissions for methane obtained in the same experimental facility at an equivalence ratio of 0.7. The fuel oil in these experiments was vaporized prior to mixing with air. Experimental data for the fuel oil in Figure 7.12 show that NOx increases by a factor of three at 6 atm when the equivalence ratio was increased from 0.5 to 0.7. As shown in Figure 7.12, the pressure dependence power factor on NOx formation for fuel oil decreased from 0.54 to 0.17 when the equivalence ratio was increased from 0.5 to 0.7. As the equivalence ratio is increased, the contribution of NO from fuel-bound nitrogen and the prompt-NO route will increase. In addition, the thermal NO will also increase because of increasing flame temperature at a given pressure. It can be noted that pressure power factor for methane and fuel

200
20

NOx and CO Formation

Fuel oil #2: φ = 0.70 P0.17

NOx (ppmv, dry at 15% O2)

16

12

Fuel oil #2: φ = 0.60 P0.29 Fuel oil #2: φ = 0.50 P0.54

8

4 Methane: φ = 0.70 P0.44 0 1 4 Pressure (atm) 7 10

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 colleagues (1996) measured and modeled NOx production in a CH4/air flame as a function 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 temperature) on the relative importance of different NO formation pathways on total NO production.

7.7╇ NO2 Formation
The total concentration of NOx exhausting from a gas turbine includes both NO and NO2. A concentration of NO2 above 10 ppm in the exhaust plume of a typical power-producing gas turbine will create a visible plume (i.e., a “brown” or “yellow” plume) (Feitelberg and Correa, 1999). The U.S. EPA characterizes NO2 Â as a reddish-brown, highly reactive gas linked with a number of adverse effects on the respiratory system. For this reason, many municipalities regulate the opacity of exhaust plumes from electrical power production plants, citing plants that produce a visible, brown plume.

7.8  Summary

201

In typical gas turbine applications, little NO2 is sustained in the combustion chamber because of the high temperatures and tendency for NO2 to be converted to NO at these conditions via: NO2 + H <=> NO + OH  R58

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 (800–1000 K), and is aided by reactions with unburned hydrocarbons 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 temperature 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 conversion 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 conversion 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
1.0 n-C4H10 NO2 / NOx 0.5 i-C4H10 0

NOx and CO Formation

CH4 C2H6 C3H8 10 C2H4 102

C3H6 H2

CO

1

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 conditions 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 combustion 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 balance 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 pathways 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: 449–70.

References
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: 397–404. 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: 31–71. 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: 33–45. 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: 1097–109. Brasseur, G. P. et al. (1998). “European Scientific Assessment of the Atmospheric Effects of Aircraft Emissions.” Atmospheric Environment 32: 2329–418. 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: 899–907 . 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: 56–62. 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. 29–70. 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 2008–972. Colket, M. B., Guile, C. R., Zabielski, M. F., and Seery, D. J. (1982). “Internal Aerodynamics of Gas Sampling Probes.” Combustion and Flame 44: 3–14. 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): 925–32. Correa, S. M. (1992). “A Review of NOx Formation under Gas-Turbine Combustion Conditions.” Combustion Science and Technology 87: 329–62.   (1998). “Power Generation and Aeropropulsion Gas Turbines: From Combustion Science to Combustion Technology.” Proceedings of the Combustion Institute 27: 1793–807. Dagaut, P. (2002). “On the Kinetics of Hydrocarbons Oxidation from Natural Gas to Kerosene and Diesel Fuel.” Physical Chemistry and Chemical Physics, 4: 2079–94. Dagaut, P., Glarborg, P., and Alzueta, M. U. (2008). “The Oxidation of Hydrogen Cyanide and Related Chemistry.” Progress in Energy and Combustion Science 34: 1–46. 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: 329–63. 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: 2459–65. 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. 125–343. Dooley, S. et al. (2010). “A Jet Fuel Surrogate Formulated by Real Fuel Properties.” Combustion and Flame 157: 2333–9.

203

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: 387–95. Edwards, T. (2007). “Advancements in Gas Turbine Fuels From 1943 to 2005.” Journal of Engineering for Gas Turbines and Power 129: 13–20. Edwards, T., and Maurice, L. Q. (2001). “Surrogate Mixtures to Represent Complex Aviation and Rocket Fuels.” Journal of Propulsion and Power 17(2): 461–6. 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. GT2011–45090. 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: 188–207 . 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: 373–80. 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: 89–112. Frassoldati, A., Faravelli, T., and Ranzi, E. (2003). “Kinetic Modeling of the Interactions between NO and Hydrocarbons at High Temperature.” Combustion and Flame 135: 97–122. 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 2012–0167 . 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 2009–4295, 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 2011–96. 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: 1–27 . 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: 89–113. 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: 051501–1. 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: 640–52. 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. GT2012–69902. 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: 655–63. Gokulakrishnan, P., McLellan, P. J., Lawrence, A. D., and Grandmaison, E. W. (2005). “Kinetic Analysis of NO-Sensitized Methane Oxidation.” Chemical Engineering Science 60: 3683–92.

References
Goodwin, D. Cantera. [Online]. http://cantera.github.com/docs/sphinx/html/index.html. Gou, X., Sun, W., Chen, Z., and Ju, Y. (2010). “A Dynamic Multi-Timescale Method for Combustion Modeling with Detailed and Reduced Chemical Kinetic Mechanisms.” Combustion and Flame 157: 1111–21. Grewe, V., Fichter, M., Sausen, C., and Dameris, R. (2002). “Impact of Aircraft NOx Emissions: Part 1: Interactively Coupled Climate-Chemistry Simulations and Sensitivities to Climate-Chemistry Feedback, Lightning and Model Resolution.” Meteorologische Zeitschrift 11: 177–86. Hanson, R., and Salimian, S. (1984). “Survey of Rate Constants in the N/H/O System,” chapter 6 in Combustion Chemistry, Gardiner, W. ed., Springer-Verlag, pp. 361–421. Harrington, J. E. et al. (1996). “Evidence for a New NO Production Mechanism in Flames.” Proceedings of the Combustion Institute 26: 2133–8. Hayhurst, A. N., and Lawrence, A. D. (1992). “Emissions of Nitrous Oxide from Combustion Sources.” Progress in Energy and Combustion Science 18: 529–52.   (1996). “The Amounts of NOx and N2O Formed in a Fluidized Bed Combustor during the Burning of Coal Volatiles and also of Char.” Combustion and Flame 105: 341–57 . Hayhurst, A. N., and Vince, I. M. (1980). “Nitric Oxide Formation from N2 in Flames – The Importance of Prompt NO.” Progress in Energy and Combustion Science 6: 35–51. Hilbert. R., Tap, F., El-Rabii, H., and Thévenin, D. (2004). “Impact of Detailed Chemistry and Transport Models on Turbulent Combustion Simulations.” Progress in Energy and Combustion Science 30(1): 61–117 . Hori, M., Matsunaga, N., Malte, P. C., and Marinov, N. M. (1992). “The Effect of Low-Concentration Fuels on the Conversion of Nitric Oxide to Nitrogen Dioxide.” Proceedings of the Combustion Institute 24: 909–16. Hori, M., Matsunaga, N., Marinov, N. M., Pitz, W., and Westbrook, C. K. (1998). “An Experimental and Kinetic Calculation of the Promotion Effect of Hydrocarbons on the NO-NO2 Conversion in a Flow Reactor.” Proceedings of the Combustion Institute 27: 389–96. Johnsson, J. E., Glarborg, P., and Dam-Johansen, K. (1992). “Thermal Dissociation of Nitrous Oxide at Medium Temperatures.” Proceedings of the Combustion Institute 24: 917–23. Kjaegaard, K., Glarborg, P., and Dam-Johansen, K. (1996). “Pressure Effects on the Thermal DE-NOx Process.” Proceedings of the Combustion Institute 26: 2067–74. Klassen, M. S., Thomsen, D. D., Reisel, J. R., and Laurendeau, N. M. (1995). “Laser-Induced Fluorescence Measurements of Nitric Oxide Formation in High-Pressure Premixed Methane Flames.” Combustion Science and Technology 110–111: 229–47 . Klippenstein, S. J., Harding, L. B., Glarborg, P., and Miller, J. A. (2011). “The Role of NNH in NO Formation and Control.” Combustion and Flame 158: 774–89. Kramlich, J. C., and Kinak, W. P. (1994). “Nitrous Oxide Behavior in the Atmosphere and in Combustion and Industrial Systems.” Progress in Energy and Combustion Science 20: 149–202. Kristensen, P. G., Glarborg, P., and Dam-Johansen, K. (1996). “Nitrogen Chemistry during Burnout in Fuel-Staged Combustion.” Combustion and Flame 107: 211–22. Law, C. K. (2006). “Propagation, Structure and Limit Phenomena of Laminar Flames at Elevated Pressures.” Combustion Science and Technology 178: 335–60. Lawrence, A. D., Bu, J., and Gokulakrishnan, P. (1999). “The Interactions between SO2, NOx, HCl and Ca in a Bench Scale Fluidized Bed Combustor.” Journal of the Institute of Energy 72: 34–40. Lefebvre, A. H. (1984). “Fuel Effects on Gas Turbine Combustion Liner Temperature, Pattern Factor, and Pollutant Emissions.” Journal of Aircraft 21(11): 887–98.   (1985). “Fuel Effects on Gas Turbine Combustion  – Ignition, Stability and Combustion Efficiency.” Journal of Engineering for Gas Turbines and Power 107: 24–37 . 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: 542–6.

205

206

NOx and CO Formation
Lu, T., and Law, C. K. (2009). “Toward Accommodating Realistic Fuel Chemistry in Large-scale Computations.” Progress in Energy and Combustion Science 35(2): 192–215. 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: 601–10. 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: 218–28. Mellor, A. M. (1976). “Gas Turbine Engine Pollution.” Progress in Energy and Combustion Science 1: 111–33. Miller, J. A., and Bowman, C. T. (1989). “Mechanism and Modeling of Nitrogen Chemistry in Combustion.” Progress in Energy and Combustion Science 15: 287–338. 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: 235–43. 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: 757–65. Miller, J. A., Klippenstein, J., and Glarborg, P. (2003). “A Kinetic Issue in Reburning: The Fate of HCNO.” Combustion and Flame 135: 357–62. 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: 43–88. 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 2010–1530.   (2010b). “On Initiating 3rd Generation of Correlations for Gaseous Emissions of Aero-Propulsion Engines,” in 48th AIAA Aerospace Sciences Meeting, AIAA Paper 2010–1529. 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: 2749–55. 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: 512–20. 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: 2393–401. 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: 705–24. Myerson, A. L. (1974). “The Reduction of Nitric Oxide in Simulated Combustion Effluents by Hydrocarbon-Oxygen Mixtures.” Proceedings of the Combustion Institute 15: 1085–92. 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: 154–63. 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: 1809–53. 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): 527–33, May–June.

References
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: 395–406. 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. 187–213. 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: 113–21. 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: 612–19.   (1995). “A Semi-analytical Emissions Model for Diffusion Flame, Rich/lean and Premixed Lean Combustors.” Journal of Engineering for Gas Turbines and Power 117: 290–301. 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: 129–44. 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: 227–37 . Skalska, K., Miller, J. S., and Ledakowicz, S. (2010). “Trends in NOx Abatement: A Review.” Science of the Total Environment 408: 3976–89. 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: 501–18. 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: 385–408. 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: 440–9. 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): 193–217. 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: 307–18. 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: 400–6. Violi, A. et al. (2002). “Experimental Formulation and Kinetic Model for JP-8 Surrogate Mixtures.” Combustion Science and Technology 174: 399–417 . 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. 1–120. Warnatz, J. (1984). “Rate Coefficients in the C/H/O System,” chapter 5 in Combustion Chemistry, Gardiner, W. ed., Springer-Verlag, pp. 197–360. 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: 897–904. Westbrook, C. K., and Dryer, F. L. (1984). “Chemical Kinetic Modeling of Hydrocarbon Combustion.” Progress in Energy and Combustion Science 10: 1–57 .

207

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: 97–128.   (1991b). “Flow Reactor Studies of Carbon Monoxide/Hydrogen/Oxygen Kinetics.” Combustion Science and Technology 79: 129–40. Zeldovich, Y. B. (1946). “The Oxidation of Nitrogen in Combustion and Explosions.” Acta Physicochimica U.R.S.S. 21: 577–628.

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 example, both are prominent candidates for carbon capture and storage (CCS) (Griffin et al., 2008; Budzianowski, 2010), where emissions requirements are driven by pipeline or geologic reservoir constraints rather than by atmospheric pollution considerations. 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 perturbation 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 (Lückerath 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 combustor, 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 differentiater 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 conductivity, 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 equivalence ratios closer to stoichiometric compared to operation with air. These changes in fuel and oxygen concentration, in turn, influence the mixture’s kinetic characteristics as discussed later. Difference in gas transport properties have important influences upon quantities such as laminar flame speed, sensitivity to aerodynamic stretch, entrainment processes in mixing layers, and heat transfer in boundary layers. Figure 8.1 compares values of thermal conductivity, λ, binary mass diffusivity (with O2), D, and dynamic viscosity, µ, for N2, CO2, and H2O obtained from the transport database of CHEMKIN (Kee et al., 2007). While the thermal conductivity and viscosity 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 water’s 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), dilution 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 processes (Naik et al., 2003).

8.1  Introduction
x 10−4 N2 6 D (m2/s) CO2 H2O

211

8

4

2

0

500 x 10−5 N2

1000 T (K)

1500

2000

8

6 µ (Pa−s)

CO2 H2O

4

2

0

500

1000 T (K)

1500

2000

0.3 N2 CO2 λ (W/(mK)) 0.2 H2O

0.1

0

500

1000 T (K)

1500

2000

Figure 8.1.  Binary diffusion coefficient (with O2) at P = 1atm, D, dynamic viscosity, μ, and thermal conductivity, λ, 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 chemically 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)  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 H + O2 (+ M)  HO2 (+ M)  H + OH(+ M)  H 2 O(+ M) R3 R4

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 concentration 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 combustion 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 different 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 combustion. The final part of each section provides insight into CO and NOx formation processes and emissions characteristics under EGR and oxyfuel conditions.

8.2  CCS Emissions Requirements
CO2 capture and storage (CCS) is usually considered as a chain consisting of three separate elements: capture, transport, and storage (Metz, 2005). The purpose of the

8.2  CCS Emissions Requirements

213

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 sequestered 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 accommodating CCS is to reduce the concentration of compounds other than CO2 in the exhaust gases to acceptable levels for transport and storage while minimizing excessive energy and economical penalties in the capture process. Control of pollutant emissions in the turbine exhausts may have a significant impact on the costs associated with flue gas cleanup. The following paragraphs describe the emission constraints imposed by CCS on EGR and oxyfuel combustion systems. 8.2.1  Emissions Requirements for EGR CO2 capture from exhaust gases containing large concentrations of incondensable gases (e.g., N2), such as those produced by gas turbines engines with EGR, is usually accomplished with liquid solvents. A wide range of technologies currently exists, even if most of them have yet to be demonstrated at power plant scale operation (Aaron and Tsouris, 2005). Ammine-based CO2 absorption systems are currently considered the most suitable for combustion-based power plants (Rao and Rubin, 2002) and are the focus of the following discussion. This technology produces a relatively clean stream that generally satisfies all the gas purity specifications imposed by later steps in the CCS process; the effects of impurities in the flue gases on the solvents are then the primary interaction between the combustion and CCS. Emissions requirements imposed by ammine-based capture systems stem from the sensitivity of these solvents to particulates and O2 concentrations in the flue gases, as well as acid gases content, such as SOx and NOx. A common approach is to use aqueous solutions of monoethanolammine (MEA). SOx, NOx, and particulates are known to react with MEA to form heat stable salts that reduce the CO2 absorption capacity. Thus, SO2 and NO2 concentrations must be maintained below 10 ppm in order to avoid excessive loss of costly solvent, while requirements on particulates content are less understood (Rao and Rubin, 2002). The problem is especially acute for SO2 because its concentration in flue gases is controlled by the presence of sulfur in the fuel and cannot be controlled in the combustion process (Lefebvre, 2010). NO2 is less of a problem because the majority of NOx emissions from gas turbines are generally nitric oxide (NO), whereas NO2 is typically below 10 percent of the total NOx. Similarly, the presence of O2 in the flue gases causes degradation of the ammines, with the byproducts leading to corrosion problems. Combustion optimization to reduce the presence of acid gases, O2, and particulates in flue gases can potentially reduce operating cost, but is not usually considered (Wall, 2007). Most research efforts are directed toward the development of solvents tolerant to impurities, for

214

Emissions from Oxyfueled/EGR Turbines

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). 8.2.2  Emissions Requirements for Oxyfuel Combustion CO2 capture from oxyfuel exhaust gases generally consists of a water condensation 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 processing 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 impurities (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 corrosion) 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
Table 8.1.  Specifications for two CO2 transport pipelines for EOR
Component CO2 CO H2O H2S N2 O2 CH4 Hydrocarbon Temperature Pressure Canyon Reef >95% – No free water < 0.489 m-3 in the vapor phase <1500 ppm 4% <10 ppm (weight) – <5% <49°C – Weyburn pipeline 96% 0.1% <20 ppm 0.9% <300 ppm <50 ppm 0.7% – – 15.2 MPa

215

Source:  Pipitone and Bolland, 2009

Table 8.2.  Issues arising from the presence of impurities in captured CO2 stream
Issue Transport (pipeline) and injection facilities Corrosion Hydrate formation Two-phase flow Leakages of toxic components Fouling Increased compression work Blockage of pores Impurities involved H2O, O2, CO, Acid forming compounds: SOx, NOx, H2S, HCN, HF, HCl H2O, CO2, H2O, H2S, CH4 Ar, O2, H2, H2S H2S, COS, CO, SO2, NOx, heavy metals Particulates All inerts: O2, Ar, N2 Particulates and O2, H2S, SO2 thorough precipitation SO2, H2S, NOx, HCl, HCN, HF H2S, COS, CO, SO2, NOx HCl, HCN, HF, H2S, NO, SO2, O2 O2, N2, Ar, H2, CO

Storage and interaction with geological formations Dissolution of carbonates minerals (Conditions are strongly Toxic compounds in case of leakage site dependent) Chemical effects on storage rocks Minimum miscibility pressure (MMP) with oil (for EOR applications)
Source:  Adapted from Anhenden et al., 2008.

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 preferable 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 considerations, 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
Table 8.3.  Gas purity requirements for different transport/storage scenarios: Scenario 1: Pipeline transport, Aquifer storage Scenario 2: Pipeline transport, On-shore storage/EOR Scenario 3: Ship transport, Off-shore storage/EOR
Scenario 1 (Modest CO2 quality) CO2 H2O SO2 O2 NOx T P >96% <500 ppm <200 mg/Nm3 Total inerts < 4%vol* – 50°C 110 bar Scenario 2 (High CO2 quality) >96% <50 ppm <50 mg/Nm3 <100 ppm – 50°C 110 bar Scenario 3 (Very high CO2 quality) >96% <5 ppm <5 mg/Nm3 <100 ppm <5 ppm −50°C 7 bar

*  No individual restriction on O2 content

as input to design the combustor. As an example of the possible variability of purity requirements, it is interesting to compare the three different transport/­ storage scenarios listed in Table 8.3. The data have been adapted from Anhenden et al. (2008), where oxyfuel combustion of coal was investigated, with similar considerations applicable to oxyfuel gas turbine cycles.

8.3  Exhaust Gas Recirculation
8.3.1  Combustor Considerations 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 chamber 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 (Wünning and Wünning, 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 combustion 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; Lückerath et al., 2008; Schütz et al., 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
Table 8.4.  Main N2-reactions involved in the different NOx formation pathways under lean conditions in GRI-Mech 3.0
Zeldovich (thermal) pathway: N 2 + O  N + NO   (R7) Nitrous oxide (N2O) pathway: N 2 + O + M  N 2O + M   (R9)
Source:  Smith et al.

217

Prompt pathway: N 2 + CH  HCN + N   (R8) NNH pathway: N 2 + H + M  NNH + M   (R10)

While internal EGR has been proposed mostly as a substitute of premixed swirl-stabilized combustors in current DLN gas turbines, application of external EGR has been investigated almost exclusively for power cycles concepts implementing carbon dioxide capture (Griffin et al., 2008). Combustor tests have focused mostly on “cold EGR,” with water condensed out of the exhaust gases before recirculation, and on “once through” experiments, in which the main species present in the exhaust gases (N2, CO2, O2) are artificially added to air without recycling the burned gases. High-pressure, large-scale tests performed by several gas turbine OEMs typically show that EGR ratios, defined as the ratio of EGR to that of the fuel and air, must be limited to 30–35 percent to avoid blowoff (ElKady et al., 2009; Evulet et al., 2009; Burdet et al., 2010; Guethe et al., 2011). 8.3.2  Emissions Trends and Kinetics Important emission impacts of EGR can be understood by analyzing results from simplified model calculations using detailed kinetics. The following simulations examine a fuel composition typical of natural gas (95 percent CH4, 3 percent C2H6, 2 percent N2 by volume) and two oxidizer composition: a baseline standard air (21 percent O2, 79 percent N2 by volume) and air with EGR (16 percent O2, 77 percent N2, 3 percent CO2, 4 percent H2O by volume), which corresponds to dilution with 30–40 percent exhaust gases. The simulations use the GRI-Mech 3.0 (Smith et al.) chemical mechanism, which is optimized for natural gas and air combustion with NOx formation and reburn chemistry, though it has not been validated for high vitiation levels or high pressures (see Chapter 7). An initial reactant temperature, Tin, of 650 K was chosen to simulate gas turbine combustor conditions. Most of the discussion in this section focuses on NOx emissions and more briefly on CO emissions, because the formation mechanism of the latter is less complex. In Table 8.4, the initiation reactions involved in NO formation are reported for ease of reference; a detailed discussion of NOx formation mechanisms is provided in Chapter 7. The combustion is modeled as a simple premixed flame, which is simulated with the PREMIX code of CHEMKIN (Kee et al., 2007). First, NO emissions are presented at a fixed combustor residence time, τres = 25 ms, a value similar to the conditions of many industrial gas turbine combustors. The residence time was calculated

218
(a) 100

Emissions from Oxyfueled/EGR Turbines

Baseline, air With EGR NO (ppm, dry) 25 atm 1 atm 10

1 1750 (b) 100

1800

1850

1900

1950

2000

Adiabatic flame temperature (K)

Baseline, air NO (ppm, dry at 15%O2) With EGR 25 atm

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 1750–2000 K at 1 and 25 atm pressure. The equivalence ratios needed to obtain these flame temperatures are φ ~ 0.5–0.64 when air is used as the oxidizer, and φ ~ 0.67–0.86 for EGR; significantly higher levels of excess oxygen are available in the undiluted air case. Figure 8.2a shows that EGR generally 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
50 Baseline, air 40 With EGR P = 25 atm 30

219

NO (ppm, dry)

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 φ ~ 0.58 for the baseline-air case and φ ~ 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 reduction in atomic oxygen O levels (Guethe et al., 2009; Li et al., 2009). This, in turn, lowers 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 promotes prompt NOx formation through reactions like R8. This is more evident at

220
(a) 80 70 NO contribution (%) 60 50 40 30 20 10 0 1750

Emissions from Oxyfueled/EGR Turbines

Zeldovich N2O NNH Prompt

1800

1850

1900

1950

2000

Adiabatic flame temperature (K) (b) 80 70 NO contribution (%) 60 50 40 30 20 10 0 1750 1800 1850 1900 1950 2000 Zeldovich N2O NNH Prompt

Adiabatic flame temperature (K)

Figure 8.4.  Contribution of primary NO production mechanisms to total NO yield for 1 atm, τres = 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 combustors 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 presence 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
(a) 80 70 NO contribution (%) 60 50 40 30 20 10 0 1750 (b) 80 70 NO contribution (%) 60 50 40 30 20 10 0 1750 1800 1850 1900 1950 2000 Zeldovich N2O NNH Prompt Zeldovich N2O NNH Prompt 1800 1850 1900 1950 2000

221

Adiabatic flame temperature (K)

Adiabatic flame temperature (K)

Figure 8.5.  Contribution of primary NO production mechanisms to total NO yield for 25 atm, τres = 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 replacing 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 reactions 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
(a) 100

Emissions from Oxyfueled/EGR Turbines

NO (ppm, wet)

25 atm 1 atm 10

Reacting CO2, H2O Non-reacting CO2, H2O 1 1750 1800 1850 1900 1950 2000

Adiabatic flame temperature (K)

(b)

5 4

x 10–3 OH
Reacting CO2, H2O Non-reacting CO2, H2O

Mole fraction

3 O 2 1 0 H 0 1 2 3 4 5 Residence time (ms)

Figure 8.6.  Total NOx concentration at 1 and 25 atm with τres = 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 formation 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
10 9 NO contribution (ppm, wet) 8 7 6 5 4 3 2 1 0 1750 1800 1850 1900 N2O 1950 2000 Adiabatic flame temperature (K) NNH Reacting CO2, H2O Non-reacting CO2, H2O Zeldovich Prompt

223

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 according 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 equilibrium 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 distribution, characterized by its average, μ, and standard deviation, σ (Lyons, 1980; Li et al., 2009), as given by f (φ ) =  ( φ − µ )2  exp  −  2σ 2  2 πσ    1

Perfectly premixed NO values are then weighted by the Gaussian distribution function to predict the total NOx emission according to the expression NO = ∫ NO(φ) f (φ)dφ 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 σ/μ = 7.5 percent. The results are plotted against adiabatic flame temperature at

224

Emissions from Oxyfueled/EGR Turbines

100

σ/µ =7.5%
NO (ppm, dry)

10

σ/µ =0
1 1750

Baseline, air With EGR 1900 1950 2000

1800

1850

Adiabatic flame temperature (K)

Figure 8.8.  Effects of different degree of fuel-air unmixedness on NOx emissions at 25 atm, τres = 25 ms.

the mean equivalence ratio. In general, unmixedness increases NO emissions, since NO emissions increase nonlinearly with flame temperature. The results also show that with EGR, NO formation is less sensitive to unmixedness. This is evident from the smaller (absolute and relative) difference between the premixed and partially premixed emissions for each case. The lower sensitivity to unmixedness is primarily a result of the reduced sensitivity of NO to temperature variations for EGR-diluted systems. This agrees with experimental results that indicate EGR is relatively more beneficial for NOx reduction in gas turbines with poor premixing or running in a partially premixed mode. For example, NOx levels were found insensitive to different fuel staging strategies in a GT24/26 Alstom gas turbine combustor with external EGR (Burdet et al., 2010). A similar conclusion was suggested, based on reductions in NOx emissions obtained with external EGR in a GE DLN system (ElKady et al., 2009; Li et al., 2009).

8.4  Oxyfuel Combustion
8.4.1  Combustor Considerations Employing oxyfuel in post-combustion separation of CCS avoids the large flow rates of N2 contained in the exhaust gases of air combustion. For this reason, oxyfuel combustion has been applied to coal-fired furnaces (Buhre et al., 2005; Toftegaard et al., 2010). In these applications, the energy penalty associated with oxygen production by the ASU is less than that imposed by separating CO2 by absorption, mainly because of the large flow rates of flue gases generated by air-fired coal plants and the large content of impurities they usually contain. For gas turbines, a variety of power cycles have been proposed that utilize oxyfuel combustion of natural gas. One class of cycles, usually denoted semi-closed oxyfuel combustion combined cycles (SCOF-CC), utilizes CO2 as a working fluid and as a diluent to control the

8.4  Oxyfuel Combustion

225

combustion temperature. Basic designs of these cycles require only a slight modification 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 characteristic 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 integrated gasification combined cycles (IGCC) (Casleton et al., 2008). For this application, nearly zero NOx emissions provides a compelling motivation relative to air combustion options. This is enhanced by the fact that oxygen-blown gasifier technology 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 attendant 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 combustor (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). 8.4.2  Emissions Trends and Kinetics 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 stoichiometry are strongly coupled. In contrast, flame temperature and fuel-oxidizer ratio are decoupled in diluted oxyfuel combustors, with the reactant mixture near stoichiometric to avoid wasting fuel and oxygen. Flame temperature is controlled independently 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, especially for long residence time combustors. Figure 8.9 demonstrates the dependence of equilibrium CO and O2 emissions for oxyfuel combustion on flame temperature

226
(a) 106 105 104 CO (ppm, dry) 103 102 101 100 10–1

Emissions from Oxyfueled/EGR Turbines

Rich CH4/air

φ =1.05 φ =1.0

φ =0.95
Lean CH4/air P=1 atm P=15 atm

10–2 1350 1450 1600 1750 1900 2050 2200 2350 2500 Adiabatic flame temperature (K) (b) 106 105 104 O2 (ppm, dry) 103 102 101 100 10–1 Lean CH4/air

φ =0.95 φ =1.0

φ =1.05
Rich CH4/air P=1 atm P=15 atm

10–2 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 generally 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 mixtures. 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
1000 P = 1 atm P = 15 atm Lean CH4/air

227

φ = 0.90
COmax/COeq 100

φ = 1.0
10

φ = 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 equilibrium 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 mixtures, 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 primary flame zone), as shown in Figure 8.10. There is a large CO overshoot in the oxyfuel flame relative to equilibrium for φ = 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 sensitivity is decreased, as indicated by the lower amount of the CO overshoot for those

228
(a) 105

Emissions from Oxyfueled/EGR Turbines

CO (ppm, dry)

104

φ = 0.90 φ = 1.00 φ = 1.05 φ = 1.10 τres = 40 ms Equilibirum

103 1500 (b) 105

1700 1900 2100 2300 Adiabatic flame temperature (K)

2500

104 O2 (ppm, dry)

103

102

φ = 0.90 φ = 1.00 φ = 1.05 φ = 1.10 τres = 40 ms Equilibirum

101 1500

1700 1900 2100 2300 Adiabatic flame temperature (K)

2500

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 equilibrium 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

10–2

Mole fraction

10–3

10–4

10–5 0.90

O2 CO τres = 40 ms Equilibrium 0.95 1.00 Equivalence ratio 1.05 1.10

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).

mainly oxidized by OH, as indicated in reaction R1, while OH is produced in the burnout region by the reaction of O2 with H mostly via reaction€R6. These competing trends suggest that an important trade-off exists in the optimal design of an oxyfuel combustor between flame temperature (the result of fuel, oxidizer, and diluent staging) and the combustor residence time. Furthermore, these calculations show that for long combustor residence times (approaching equilibrium), the combination of CO and O2 emissions are sensitive to the exact stoichiometry and temperature at which combustion occurs, while for limited residence time this sensitivity is reduced. These trends have been experimentally verified, as shown in Figure 8.12, which presents measurements and calculations of O2 and CO emissions from a CH4/CO2/O2 atmospheric pressure, premixed, oxyfuel combustor (Amato et al., 2011a). Returning to the discussion of Figure 8.11, the dramatically reduced sensitivity of the emissions to stoichiometry suggests that mixing quality is not critical, at least at atmospheric pressure. At higher pressure, however, both CO and O2 recombination reactions are accelerated so that CO and O2 emissions are likely to approach equilibrium within the 40 ms residence time (they also have lower concentrations than at atmospheric pressure as shown in Figure 8.9). As such, the sensitivity of the finite residence system to stoichiometry and, thus, mixedness, should increase with pressure.

8.5╇ Concluding Remarks
This chapter has reviewed emissions issues for combustion systems that rely on extensive levels of EGR and operate with oxygen rather than air. Additional considerations apply for these systems relative to air-fueled systems, partly because of the alteration of mixture and chemical kinetic rates, and also because of differences

230

Emissions from Oxyfueled/EGR Turbines

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 pollutant, 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): 321–48. Amato, A., Hudak, B., D’Souza, P., D’Carlo, 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): 3399–405. Amato, A., Hudak, B., D’Carlo, 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): 39–59. 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(5–6): 656–68. 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): 343–54. 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): 1653–63. Budzianowski, W. M. (2010). “Mass-recirculating Systems in CO2 Capture Technologies: A Review.” Recent Patents on Engineering 4(1): 15–43. 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): 283–307 . 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
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): 1013–52. Cavaliere, A., and de Joannon, M. (2004). “Mild Combustion.” Progress in Energy and Combustion Science 30(4): 329–66. 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): 2693–700. 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): 740–7 . 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): 1261–78. De Visser, E., Hendriks, E., Barrio, M., Mølnvik, M. J., De Koeijer, G., Liljemark, S., and Le Gallo, Y. (2008). “Dynamis CO2 Quality Recommendations.” International Journal of Greenhouse Gas Control 2(4): 478–84. Ditaranto, M., and Hals, J. (2006). “Combustion Instabilities in Sudden Expansion Oxyfuel Flames.” Combustion and Flame 146(3): 493–512. 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): 279–95. 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 GE’s Dry Low NOx Combustors Utilizing Exhaust Gas Recirculation for Post Combustion Carbon Capture.” Energy Procedia 1(1): 3809–16. 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): 291–6. 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): 3401–8. Griffin, T., Bücker, 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 García, M., and Burdet, A. (2009). “Flue Gas Recirculation in Gas Turbine: Investigation of Combustion Reactivity and NOx Emission.” ASME Turbo Expo, Orlando, FL, USA.

231

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(1–6): 49–64. 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): 1255–67 . 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): 2619–26. 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): 247–54. 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): 3383–90. Kvamsdal, H., Jordal, K., and Bolland, O. (2007). “A Quantitative Comparison of Gas Turbine Cycles with CO2 Capture.” Energy 32(1): 10–24. Lammel, O., Schütz, H., Schmitz, G., Lückerath, R., Stöhr, 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): 333–49. 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): 725–34. 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): 1593–605. 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): 202–13. Li, S., and Williams, F. (1999). “NOx Formation in Two-stage Methane-air Flames.” Combustion and Flame 118(3): 399–414. Liu, F., Guo, H., Smallwood, G. J., and Gülder, Ö. 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(1–2): 778–87 . 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): 495–7 .

References
Lückerath, 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): 660–5. 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): 1223–30. Mathieu, P., and Nihart, R. (1999). “Zero-emission MATIANT Cycle.” Journal of Engineering for Gas Turbines and Power 121(1): 116–20. 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): 425–31. 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): 1141–60. 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): 551–65. Pipitone, G., and Bolland, O. (2009). “Power Generation with CO2 Capture: Technology for CO2 Purification.” International Journal of Greenhouse Gas Control 3(5): 528–34. 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): 4467–75. 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(1–2): 225–30. 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): 776–83. Sadanandan, R., Lückerath, 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): 970–80. 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)): 765–72. Sanz, W., Jericha, H., Bauer, B., and Göttlich, 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): 535–42.

233

234

Emissions from Oxyfueled/EGR Turbines
Schütz, H., Lückerath, 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): 831–44. 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): 581–625. 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): 1537–44. Vandenhengel, W., and Miyagishima, W. (1993). “CO2 Capture and Use for EOR in Western Canada 2. CO2 Extraction Facilities.” Energy Conversion and Management 34(9–11): 1151–6. Wall, T. (2007). “Combustion Processes for Carbon Capture.” Proceedings of the Combustion Institute 31(1): 31–47 . White, V., Torrente-Murciano, L., Sturgeon, D., and Chadwick, D. (2009). “Purification of Oxyfuel-derived CO2.” Energy Procedia, 1(1): 399–406. 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): 64–88. Wünning, J., and Wünning, J. (1997). “Flameless Oxidation to Reduce Thermal NO-formation.” Progress in Energy and Combustion Science 23(1): 81–94. Yantovski, E. (1996). “Stack Downward: Zero Emission Fuel-fired Power Plants Concept.” Energy Conversion and Management 37(6): 867–77 . 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): 1061–71. 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): 352–60.

PART 3

CASE STUDIES AND SPEcIFIc TEcHNOLOGIES: POLLUTANT TRENDS AND KEY DRIVERS

9 Partially Premixed and Premixed Aero Engine Combustors
Christoph Hassa

9.1 Introduction
Premixed combustors for aero engines have been under development for nearly forty years, yet, at the time of writing, the first airplane with premixed combustion 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 partially premixed combustion. An obvious difference between both applications lies in the fuels, which are predominantly gaseous for industrial gas turbines and exclusively 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, prevaporized (LPP) combustion, which for a large part were concurrently achieved with development efforts on LPP combustors. Understanding the limitations and difficulties 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): Prioritized Combustor Design Considerations: • Safety – Critical for Flight (Prime Reliable) • Operability – Takeoff to 45,000 ft. (Lean Blowout/Altitude Relight, Heavy Rain, and Hail) • Efficiency – 99.9% • Durability – 6,000 Cycles / 36,000 hr • Emissions – Best Available Emissions at 7%, 30%, 85%, and 100% Power and During Cruise • Downstream Turbomachinery Thermal and Life Integrity The presenter goes on to state: “Emissions are normally the fourth or fifth combustor design priority,” a statement that will find the agreement of everybody boarding 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 combustor into the engine, and safety ultimately needs to be established by flight testing. 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.

9.2 Some Results of Research Relevant to Premixed, Prevaporized Combustion in Aero Engines
9.2.1  NOx Reduction Potential, Influence of Pressure and Residence Time 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 theory, a three order of magnitude reduction of NOx should be possible for homogeneously 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 Φ = 0.4. CO equilibrium was reached at all conditions after 1.5 ms within experimental 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
10 Nitrogen oxides (NOx) emission index (gNO2/KgC3H8) Equivalence ratio 0.6

239

Combustion efficiency (%) 99.70

99.50

0.5

1 99.25

0.4 97.00 0.5 1.0 1.5 2.0 2.5 3.0

Residence time (msec)

Figure 9.1.  NOx emission and combustion efficiency with residence time, T = 800 K, P = 5.5 bar (Anderson, 1975).

shows that Φ = 0.4 and 0.5 curves display an asymptotic behavior with time, which is not apparent for Φ = 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 temperature stays below 1900 K, the thermal NOx formation is very slow. However, for Φ = 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 different pressures because the prompt mechanism does not depend on it. For higher temperatures, the thermal mechanism takes over and displays the square root dependence 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 stoichiometric condition makes the control of NOx emission with RQL combustors very difficult 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
Forward thermal NO formation rate (ppm/ms) 104

Premixed Aero Engine Combustors

30 atm, 900 K 102 10 atm, 600 K 100 10–2 10–4 10–6 0.4 1 atm, 300 K

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 production 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 combustor 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 premixing; an additional limitation of the residence time is needed. Because of this coupling of cooling and emission at high combustor outlet temperatures, a reduction 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 combustors now in development still need to respect the limitations given by the available 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 reaction (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
S=0 10 0.1 0.2 0.28 NOx emission index (gm/kg fuel)

241

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  Mixture Homogeneity and Prevaporization
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 equivalence 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 Φ = 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

global equivalence ratios around stoichiometry, wide mixture distributions actually reduce the NO emission, such that premixing only makes sense below Φ = 0.8. However, if the global equivalence ratio is leaner than that, homogeneity plays a decisive role and can have an order of magnitude influence. For lean combustors, the rapidity of mixing is, therefore, perhaps the most important parameter, which can be influenced by the combustor design. Often this behavior is also used during combustor development. The NOx versus AFR curves of different configurations are compared and their steepness is taken as a measure of the homogeneity of the respective mixture.
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 diffusion 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 prevaporization 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 conditions and quantifying it right before the flame has not been achieved yet. However, experimental investigations exist that quantify the degree of prevaporization entering 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 percent. At Φ = 0.7 and 700 K preheat, no influence of prevaporization was measured, whereas at Φ = 0.6, the NOx emission doubled for the lower prevaporization. These results were reproduced with a superposition of reactor calculations from pure diffusion burning for the liquid part and pure premixed combustion for the prevaporized fuel. The calculated NOx variation of 6 percent for Φ = 0.7 was within the experimental 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
10 8 EINO (g/kg) 6 4 2 0

243

0

5

10 15 cslip (m/s)

20

Figure 9.4.  Emission index of NOx for a 50 µm heptane droplet versus relative droplet velocity 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 µm 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, prevaporization doesn’t 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 µm. 9.2.3  Lean Stability The available literature reports lean blowoff for premixed combustion around 1650 K. The operating range of aero engine combustors and its consequences on combustor 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 Φ = 0.25 for 600 K preheat. The study was confined to the recirculation rates, possible with bluff body flame holders. As the recirculation 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 stabilization or large-scale, jet-driven stabilization, as used in the so-called flameless combustion, 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 temperature, which is again set by AFR. Here the OH radical plays a double role as precursor 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. 9.2.4  Combustion Efficiency 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 combustor, 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 convectively cooled and the mixture is fully premixed before combustion, satisfying the residence 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
2.0 1.8 1.6 1.4 Ratio 1.2 1.0 0.8 0.6 0.4 1000 1050 1100 1150 NOx/NOx1050 – ratio CO / CO 1050 – ratio Loads 94%, Pressure 4bar

245

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 demonstrated 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 1100°C. 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 In premixing passages of LPP burners, autoignition can occur at sufficiently high temperatures and residence times. If a flame stabilizes in the premixing passage, destruction of the burner is likely to occur. For that reason, the ignition delay time is the primary design criterion for premixers. It is defined by the chemical kinetics of the fuel even for liquid fuels, since evaporation starts right at the beginning of the fuel injection and the chemical reactions start at the same time. To measure it in a reproducible manner is a difficult task, because the final ignition is the integral outcome of reactions whose velocity depends on the local conditions, which the ignition kernel has passed on its way. Since complete premixing cannot be done infinitely fast,

246

Premixed Aero Engine Combustors

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 combustion 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 temperatures 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.

τ=


0.508  3377  exp   P 0.9  T  

(9.1)

Here τ is the autoignition time in ms, P the pressure in bar, and T the air temperature in 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 turbines. 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 temperatures, 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 passages. 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: 1. Flashback through the boundary layer (Plee and Mellor, 1978): For an upstream movement of the flame through the boundary layer, the mixture composition has to be within the flammability limits of the fuel. For a fuel injection in the center of the channel, this will not be the case until the dispersion transports the fuel to the wall. Although the gas velocity has its minimum in the boundary layer, at some point the quenching distance will be reached that leads to extinction. Accordingly, the flow in the passage should be accelerated first to prevent separation of the boundary layer but also to keep it small. New results concerning adverse pressure gradients and confined flames are given in Eichler and Sattelmayer (2011).

9.3  Development of Lean, Prevaporized Combustors with Premixing Ducts
Table 9.1.  Subsonic turbofan performance data, ISA SLS rated take-off thrust 459 kN (Wedlock et al, 1999)
Condition Thrust Combustor inlet conditions Temperature Pressure Air flow K bar kg/s Take-off Climb out Approach Idle Cruise 100% 85% 30% 7% 917 878 709 551 846 47.6 41.09 18.17 7.01 17.58 157.2 140.3 73.1 33.3 60.3 Combustor Combustor AFR exit temperature K 36.88 40.83 68.57 115.11 39.91 1792 1683 1232 886 1672 Fuel flow kg/s

247

4.262 3.436 1.066 0.2893 1.511

2. Combustion-oscillation-induced flashback (Guin, 1999): The amplitude of a pressure oscillation can be so high that during the phase with positive pressure gradient, the velocity is decelerated to a level below the flame velocity. Since this condition is first reached near the wall, it thereby induces a flashback through the boundary layer. 3. Combustion-induced vortex breakdown (Fritz et al., 2004): This mechanism is only relevant to swirl flames, which stabilize near the burner outlet. The combustion-induced dilatation of the flow generates a reduction of the pressure gradient at the burner outlet. If the distribution of axial velocity already has a near minimum at the centerline, breakdown of the vortex can occur and a recirculation bubble moves upstream. If the mixture in this zone is within the flammability limits, flashback can occur. Lower swirl strengths will then increase the security against flashback. A further possibility to suppress this type of flashback is the installation of an inner body blocking the recirculation at the axis. However, combustion-induced flashback has also been observed with central bodies (Heeger et al., 2010) where flame-induced separation of the boundary layer on the inner body seems to have occurred within a straight annular channel. 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 flashback. During such tests, the burner flow function is continually reduced at the relevant 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.

9.3 Development of Lean, Premixed, Prevaporized Combustors with Premixing Ducts
After the information given in the last paragraph, a section with this title might seem futile. How can premixing be done in such a short time? There are two reasons why a review of some of the development projects and their supporting research is useful

248

Premixed Aero Engine Combustors

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. 9.3.1  Premixed Combustor Development in NASA-funded Programs
9.3.1.1  Technology for Subsonic Flight in the NASA Experimental Clean Combustor Program (ECCP) and the Energy Efficient Engine Program (E3) These programs were part of aviation’s answer to environmental concerns about air pollution arising in the early ’70s. The EPA proposed regulations (1972), which entered the specifications of those programs. Under the ECCP program, Combustor projects began that covered the full range of aero engines (Jones, 1978). Only the programs for large subsonic turbofans will be considered here. The baseline engines were the GE CF-6–50 and the Pratt & Whitney JTD-9. Fuel staging was part of all considered combustor architectures. One of four GE concepts was a radial/axial staged combustor with a premixing main stage, whose cross section is depicted in Figure 9.6 (Bahr and Gleason, 1975). The inner pilot stage has a conventional design, whereas the outer main stage consists of an annular premixing channel terminated by bluff body flame holders of semicircular cross section. Main fuel injection consisted of sixty fuel bars with a pair of circumferentially directed holes of 1.03 mm diameter. Average residence time in the premixer was 0.4 ms, Φ at full load 0.5. Sixty percent of the air went through the main, 23 percent through the pilot stage. The pressure in the annular combustor was limited to 9.53 bar. During the emission tests, which were done for simulated takeoff, climb, and cruise, the part of the overall fuel going through the pilot was varied. Figure 9.7 shows the results of these variations extrapolated to 30 bar. Qualitatively, it resembles the curves measured in the experiments with premixers for a varying degree of prevaporization, which suggests that the pilot burned indeed as a diffusion flame. At 24 percent pilot fuel fraction, an efficiency of 99.6 percent was reached. For a fuel fraction above 0.3, the emissions are at the level of today’s conventional CF6–50 engines. Of course, the comparison of a thirty-five-year-old combustor with the present-day RQL version is somewhat unfair; however, its purpose is the comparison of a simple version of an LPP combustor with the benchmark RQL technology at the operating point where LPP should do particularly well. If the trends in Figure 9.7 are extrapolated to zero percent pilot fuel, a main stage EI of around five results. That number allows the conclusion that it is possible to reduce NOx by one order of magnitude with a rather simple premixing stage, without elaborate mixing schemes and with a rather crude atomization device. But if the homogeneity is not right, and with the limitations of bluff body stabilization, too high temperatures must be created to burn out the very lean bits of the mixture

9.3  Development of Lean, Prevaporized Combustors with Premixing Ducts
15 mounting struts 60 main stage Fuel spraybars 8.61 3.81 21.74 14.58

249

0.103 dia.

78.08 dia. 61.31 dia. 60 flameholders 3 m 2c 3 1 8, l.=3 Vo 83.70 dia.

30 75.03 dia. 72.29 dia. Vol.=14,979 cm3 63.55 dia. 30 Pilot stage Fuel injectors 2.16



Linear dimensions in centimeters 56.22 dia. 30.35

Figure 9.6.  Radial/axial staged GE combustor of ECCP Ph I, all dimensions in cm (Bahr and Gleason, 1975).

50

NOx emission index (g/kg fuel) Extrapolated to hot day takeoff

40 30 20 10

II-6,7 II-10 II-12,15 II-14 III-2 f = 0.024+ 0.001

0 0.0

0.0

0.2 0.3 0.4 Pilot fuel flow Total fuel flow

0.5

0.6

Figure 9.7.  Variation of NOx emissions with pilot fuel fraction of radial/axial staged GE combustor 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 unsatisfactory. 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 therefore no cruise emissions can be found in the ICAO data bank, it can be assumed that

250
30 aerating nozzles

Premixed Aero Engine Combustors
60 inner main swirlers

Main combustion zone Dilution air Pilot combustion zone

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 combustion. No data exist to tell where stabilization actually happened. Since pilot exhaust temperatures resulted in very short autoignition times, it can be suspected that ignition 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 concept 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

Although pressure scaling of NOx for combustors with an unknown degree of diffusion and premix burning is arguable, the result is in line with the mixing characteristics, which can be inferred from the architecture. For the successful lean stability extension of McVey and Kennedy (1979), the pilot fuel was injected into the wake of the stabilizing bluff body. In the radially staged premixer, the pilot stream has little chance of reaching the stabilization zone of the bluff body and increasing temperature there, as it had to cross the high-velocity jets emerging between the flameholder bodies. Hence, main combustion was only enhanced downstream where the flameholder wakes had collapsed and remaining residence time was small. The Vorbix concept is inherently more stable because the main fuel mixes in the hot pilot instead of the pilot mixing into the main exhaust as in the premixing combustor. Any combustion of main fuel taking place at non-premixed condition effectively acts as a pilot of the main stage, only that its part cannot be controlled by a separate fuel stream. As in RQL combustors, the transition through stoichiometry has to be controlled by mixing, and here the lower pressure ratio of the JTD 9 helped. The potential problem was amended in later phases of the program by shortening the distance between atomizer and swirler and, in later programs, by integrating the atomizer in the main swirlers and adding a premixing channel to the swirlers. In ECCP Phase II, approach in staged mode could be demonstrated with 99 percent efficiency (Roberts et al., 1976). At the end of the ECCP program, combustion technologies to satisfy the EPA rules were at hand, however without the control technology for staged injectors to go with them. Obviously, the creation of an airworthy fuel control system for two separate stages is a development program in itself. Here the introduction of digital control systems permitting much higher freedom in staging was an important benefit to premixing combustors. As stated in Section 9.2.1, cooling is another enabling technology required for successful premixed combustion. Reading about the air distribution of the radially staged premixer, one might have noted that only 17 percent was foreseen for liner cooling. Lean combustion has its advantages with respect to cooling, however in the analysis of the ECCP results, doubts about liner survival arose and it was almost certain that liner lifetime goals would be missed (Sokolowski and Rohde, 1981). New cooling methods were developed in the following E3 program. Impingement-film cooling and double-walled liners were used as well as segmentation of the liners for reduction of stresses. But the liner surface also had to be reduced to reach the goals. The lesson learned was that only a short combustor is a good lean combustor, with the residence time requirements satisfied by a higher equivalence ratio of the main combustion zone. The values for the Vorbix combustors in the final tests were around 0.7; the price to pay was that the reduction of the EPA parameter, similar to today’s ICAO parameter for NOx, was around 50 percent compared to the production engine. After ten years of lean combustor development, workable concepts were found and many operability issues solved, but most of the theoretical promise of a three order of magnitude reduction in NOx got lost on the way.

252

Premixed Aero Engine Combustors

9.3.1.2  Technology for Supersonic Flight in the Stratospheric Cruise Emission Reduction (SCERP), the Advanced Low Emission and the High Speed Civil Transport Program (HSCT) For supersonic transport, a tenfold reduction of NOx was deemed necessary according to the findings of atmospheric science of the day. The program goal for the NOx EI of 3g/kg is comparable to the current standard of 25 ppm at 15 percent O2 for industrial gas turbines and much more demanding than those of the subsonic programs. From the previous section, it is clear that the NOx goal could not be achieved with a diffusion pilot stage. In principle, variable geometry offers the potential of circumventing the staging of lean combustion. However, for the combustors of these programs, it was used to extend the stability range of staged premixing combustors, allowing for turndown of the pilot at full power. All but one of the concepts preserved the pressure loss of the combustors by keeping the effective area of the combustor constant. Mechanical valves and turning vanes were used, sometimes with crank arms and joints moving in the secondary channel of the combustor (Fiorentino et al., 1980; Ekstedt and Fear, 1987). One “fluidic” concept influenced the diffuser flow with bleeds to divert the flow from pilot to main but was discarded because the achievable air modulation was judged too low (Fiorentino et al., 1980). At high power, the variable area vanes of the variable tube parallel staged combustor (Ekstedt and Fear, 1987) shifted the flow to the premixer in the outboard position, giving the outlet temperature profile a peak on the outward side. This directs the attention to another problem of staged combustors, where special attention must be given to the mixing of pilot and main flow if a degradation of the outlet temperature profile compared to unstaged combustors is to be avoided. The premixer of the parallel staged combustor ended in a perforated plate flameholder with internal air cooling and 70 percent blockage, set by the program limit of 5.5 percent for the overall combustor pressure loss. While cruise NOx was achieved with the series staged combustor tested in parallel, the efficiency goal of 99 percent could not be reached at the same time even with the variable airflow. Yet tests never reached maximum conditions at 30 bar because of autoignition. Wakes of the variable area vanes in one case and wakes of the fuel-injector tips in the other were held responsible for the damage incurred at 7 bar. Additional purge air shifted the onset of burning in the premixing duct to 19 bar. An additional conclusion from these results was that premixing ducts with a rapid area constriction at their end are inherently unsafe, as is the case with bluff body flame holders. If either autoignition or flashback occurs, the combustion-induced pressure loss leads to a redirection of the air, causing the upstream velocity to go down, further exacerbating the problem. However, it is very difficult, or next to impossible, to avoid flashback at all times. If combustion oscillations occur somewhere in the operating envelope with main operation, maybe even only during a transient, flashback or autoignition is bound to happen as soon as the amplitude of the oscillation reaches the level of the liner pressure loss. The velocity in the duct falls below the flame velocity, or the low velocity causes the residence time in the duct to exceed the autoignition limit. Hence, premixing has to be achieved in a way that will lead to extinction of the flame in the premixing

9.3  Development of Lean, Prevaporized Combustors with Premixing Ducts

253

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 integrated 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 requirement 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 velocity 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, staging to achieve the operating range became a more difficult task as variable geometry 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 configurations, 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 formulated 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
5000

Premixed Aero Engine Combustors

4000

Tflame (°R)

3000

Pilot only

2000 Fully fueled 1000 Pilot + 100% outer dome + 75% inner dome Pilot + 50% outer dome + 50% inner dome Pilot + 25% outer dome + 25% inner dome 0 1000 2000 T4adi (°R) 3000

0

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 problem 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 performed 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 availability 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 continues 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
Table 9.2.  Comparison of useful operating ranges for premixing modules and LDI injectors tested by GE in HSCT program (Kastl et al. 2005)
Mixer design Single cup test at P3 (atm) 1 IMFH Swirl IMFH Cyclone Swirler Swirl Jet 1 Swirl Jet 2 LDI Jet Mix LDI Multiple-Venturi X X X X X X X 4 X X X X X X 14–17 X X X Φ(NOx EI = 5) Φ(Comb. Eff. = 0.99) 1.4 1.3 1.6 1.2 1.1 1.3 P4 (atm) / T3 (R)

255

10 /1410 4 / 1410 10 / 1310 4 / 1460 4 / 1400 14 / 1520

X

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 useful. 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 temperature 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 currently formulated, the −80 percent NOx compared to year 2000 technology goal of ACARE or the −75 percent CAEP 6 of NASA’s ERA program, would result in a higher NOx emission at takeoff. Combined with the harsher conditions for premixing, full premixing and prevaporization in premixing channels no longer appear an option for subsonic turbofans. Accordingly, recent development has focused on partial premixing and prevaporizing lean burning combustors, presented in the remainder 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  Vaporization and Mixing in Premixing Ducts
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 reliable 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
50

Premixed Aero Engine Combustors

P = 3 bar Relative liquid mass flux (%) 40 P = 6 bar P = 9 bar P = 12 bar 30 P = 14.5 bar

20

10

0

0

20

40

60

80 100 x (mm)

120

140

160

Atomizer I, T=750 K, w=120 m/s, Ir=125 g/s/m

Figure 9.10.  Relative mass flux of kerosene in prevaporizer channel for different air pressures (Brandt et al., 1997).

atomization alike, which are the methods most commonly used in aero engines. Hence, measurements should be carried out at elevated pressure, posing an additional 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 airflow; hence, a typical cruise condition with similar air density and velocity was covered. Good atomization with Sauter diameters below 20 µm 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 additional 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 carried 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 channel 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
Relative liquid mass flux (%)

257

50 40 30 20 10 0 0 T= 750 K T= 850 K 20 40 T= 650 K

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 temperatures (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 engine cycles. 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 density and velocity were held constant (Hassa and Wiesmath, 2012). Because of the lower surface tension, the Sauter diameter drops from 20 to 15 µm. 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 temperature on vaporization rate was also demonstrated. For a residence time of 0.5 ms at the investigated condition, the temperature difference of 110°C 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 pressure loss without autoignition or flashback. During NASA’s 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

Airflow Fuel nozzle

Vortex Generators (8) 11.4

5.3

Low pressure Fuel tubes

Perforated 5.3 plate

Concept A – vortex generators Fuel Nozzles (4)

9.4 Concept C – perforated plate Rectangular duct (10.2 cm wide) Low pressure Injectors (4) 3.8 14.2 Concept D – annular duct

4 fuels tubes (1.75 cm I.D.) 18.2 Concept B – multi tube

5.3

Figure 9.12.  Premixing duct design concepts of GE (Dodds and Ekstedt, 1985).

1-Tip injector (D-1)

4-Tip injector (D-2) 0.56 mm I.D. tube

6-Tip injector (D-3)

Figure 9.13.  Annular duct fuel-injector configurations of GE (Dodds and Ekstedt, 1985).

the mixture should be between 0.4 < Φ < 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 injection 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 distribution 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 suppliers 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: 1. Spatial distribution of fuel by multiple injection points, as described in the preceding paragraph;

9.3  Development of Lean, Prevaporized Combustors with Premixing Ducts
Core flow

259

Secondary flow Air

Atomization occurs here

Air Fuel

Axial swirler

Radial inflow swirler

Figure 9.14.  Carburetor tube of Vorbix combustor (Zeisser et al., 1982).

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 premixing 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 centrally 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 homogenized 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
Fuel injection ports

Outer annulus

Air

Fuel

Air

Center body

Axial air swirlers

Radial air swirler

Figure 9.15.  Schematic of LPP module from Bittlinger and Brehm (1999).

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 liftoff, 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 channel 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 supported by the swirl of the annular channel. The other task of the swirl is the circumferential 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 atomization. Experiments performed at pressures up to 5 bar uncovered different operation

9.4  Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors
Separation bubble

261

Spray injection panel (a) (b)

Nodule (c)

Figure 9.16.  Airblast in cross-flow (a) discrete jet (b) atomized jet (c) atomized jet with liquid nodule (Leong et al., 2001).

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 considerable 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 droplets 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 atomization becomes smaller at higher pressure, differences in penetration at higher operating points will be also smaller, and if the system is only used for main injection, the independence of operating might be nearly achieved.

9.4 Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors
Full prevaporization and premixing in an aero engine is hard to realize, and even if it were realized easily, its inherent risks are hard to sell to aviation. As has been shown, for large engines, satisfactory premixing with liquid fuels is simply not possible in premixing passages alone. Common to the solutions is a lean burner AFR at full load, avoidance of premixing channels ending in flame holder configurations with an area constriction, and some degree of prevaporization and premixing before the flame. Premixing in the combustor before the flame is necessary. Given that conclusion, it is logical to try out how far one can get with premixing in the combustor only, which is the idea behind lean direct injection. First developments were undertaken for industrial turbines and gaseous fuels, but soon the applicability of the approach to liquid fuels was also investigated.

262

Premixed Aero Engine Combustors
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 configurations, also from Andrews and colleagues (1990), were tested at elevated pressures 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 module. 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 incomplete 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 diffusion 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).

9.4.1  NASA Lean Direct Injection Work 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 solutions 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 assembly 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 temperature resulted in a NOx increase from EI four to eight for a twenty-five-injector

264

Premixed Aero Engine Combustors
Number of Configuration Symbol injectors 25 25 25 25 36 A B C D E Swirl number 0.5 0.5 and 0.8 combination 0.8 Swirl direction All same direction All same direction All same direction

0.5 and 0.8 Alternating combination direction 0.5 and 0.8 All same combination direction

100 EINOx (g-NOx/kg-fuel)

10

1

1400

1600 1800 2000 Flame temperature (K)

2200

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 configurations 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 possible way for mixing, it is still far from a homogeneous reactor for the range of allowable 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 shrinking 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 temperature 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 potential 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 elegant solution for part load operation can be found. 9.4.2  Lifted Flames and Piloting of Lifted Flames The preceding sections demonstrated that for high-pressure ratio engines, full premixing 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.
9.4.2.1  Factors Influencing Flame Lift-off in Swirling Flows with Recirculation First, it is useful to become aware of the difference between a lifted jet flame and a lifted flame in a combustor with a swirling recirculation. Whereas the jet flame lifts

266
10

Premixed Aero Engine Combustors

8

6 Z /R 0 1 .4 1 .2 1 .0 0 .8 0 .6 0 .4 0 .2 –4 –2 4 2 0 X /R 0

2

4

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 stabilized 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 mixture has reached stoichiometry and burning in a diffusion mode will ensue. At constant 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

Pressure also has an effect on turbulent burning velocity as well as on atomization. For higher hydrocarbons, the decrease of the laminar burning velocity with pressure is overcompensated by the decrease of the thermal diffusivity and consequently the lift-off decreases. If we now look at the operational envelope of aero engines, the parameters are not independent. With higher power, global equivalence ratio, pressure, and preheat temperature all rise. We have seen that all factors lead to decreasing lift-off. Unfortunately, this is exactly the opposite of what would be desired for low emission and good operability, as lower lift-off produces higher temperatures and higher NOx at high power and higher lift-off is nearer to blowoff at part load. Hence, lifted flames are not fit to fulfill the aero engine requirements unless the lift-off can be manipulated by piloting.
9.4.2.2  Piloting of Lifted Flames From the previous discussion and the review of staged combustors with a pilot zone in Sections 9.2.3 and 9.3, it is clear that the heat released from the pilot must reach the stabilization zone of the main combustion without too much dilution, or the heat required for stabilization will give rise to an undesirable amount of NOx formation. Hence, an internal pilot is the method of choice, independent of the eventual existence of a pilot zone for staging. For main burners using swirl or annular airflows, a concentric pilot burner is one solution. The question of piloting with an inner or outer pilot seems trivial; almost all premixing combustors in industrial turbines use internal pilots or pilot zones; however, there is a trade-off with respect to ignition. The prevailing method used in aero engines is a spark igniter protruding through the outboard liner. Using an inner pilot means having to ignite a pilot spray shielded by the main airflow from the spark plug. Consequently, an outer pilot has been tried (Brundish et al., 2003). However, the difficulties of providing good circumferential homogeneity with very little pilot fuel and achieving good premixing for the main within a space, which is drastically limited by the surrounding pilot zone, seem more important such that the inner pilot is the method of choice. The simplest method to implement an inner pilot is the addition of a fuel injector in the center of the burner. The effect of piloting on the combusting flow field is illustrated with some results from a study using planar optical techniques (Hassa et al., 2005). A double swirl nozzle with radial swirlers was used. The main fuel was injected between the swirlers with a prefilmer; the pilot fuel was injected with a central pressure-swirl injector. Figure 9.21 shows an instantaneous temperature field for 10 and 20 percent piloting at an operating condition near idle. The 10 percent picture displays a broken flame front that allows fuel to escape unburnt; the picture for the 20 percent shows a somewhat more continuous high-temperature zone that still has some holes in the flame front. Not surprising, intolerable UHC emissions were measured. The effect of the degree of piloting can be seen in Figure 9.22, where liquid and gaseous kerosene was imaged with LIF and the reaction zone visualized with the deconvoluted OH* emission. The 20 percent pilot leads to a lifted and very elongated heat release, which is not finished at the upper end of the window at 55 mm from the

268

Premixed Aero Engine Combustors
10% pilot 20% pilot

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).
Fuel Reaction zone

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 bandwidth 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 knowledge 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 fraction 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. 9.4.3  Partially Premixing Combustors with Internal Pilots Compared to the supersonic engines, for a long time the driver of premixed combustion, 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 possible 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 architecture 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 cooling 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 premixed 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

Pilot flame 60.4 mm

68.5 mm

Figure 9.23.  Burner E of TechCLEAN program (Yamamoto et al., 2010).

9.4.3.1  The Japanese TechCLEAN Combustor The TechCLEAN engine is directed to a 50–100-seat airplane family with eight thousand to twelve thousand pounds of thrust (Futamura et al., 2009). The combustor is not the first lean SAC; however, information on its aerodynamics is published such that it can be used as an introduction to design issues. Testing was done for two cycles, the TechCLEAN and the ECO engine cycle, at a pressure ratio of twenty-five and seventeen, respectively (Yamamoto et al., 2010). The special conditions for lean combustion are those of a small engine with a small combustor and a low combustor inlet temperature. Single skin effusion cooling was realized for the sector tests. The burner schematic is shown in Figure 9.23 (Yamamoto et al., 2010). Main fuel injection is effected with jets spraying on a splash ring, but the pilot injector uses a more conventional prefilmer. Not less than five air swirlers and one swirling fuel path are used. Together with the swirler effective areas, the recess of the pilot injector, and the separation of the pilot and main airstreams, there are many design parameters. With the simplicity of the single annular combustor comes the complexity of the burner that makes it impossible to separate functions geometrically and satisfy one operational requirement at the time. In the beginning, it is a rather complex and large set of parameters that requires extensive work to understand and optimize, especially as we know that the elliptic nature of the recirculating combustor flows causes changes in one incoming airflow to change the others. This burner exhibits design features not yet introduced: the radial separation of pilot and main flows and the higher recess of the pilot zone. The direction of pilot swirl was always opposite the main flow, producing lower net swirl as with co-rotation. Hence there is some reattachment length of the flow behind the backward-facing step in the pilot recess, denoted by the circles in the scheme. The pilot flame is stabilized by the pilot recirculation. Figure 9.24 shows a computation of the isothermal flow of the nozzle. The global flow pattern suggests flame stabilization of the main combustion zone at the boundary of the inner recirculation. Hence, the heat release of pilot and main are geometrically separated, preventing a merging of the flames and diminution of the main flame lift-off as in Figure 9.22. This seems to be a common feature

9.4  Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors
y = 0, color : tangential velocity (m/s) : 50.0 (m/s)

271

50 -50

Figure 9.24.  Isothermal flow of burner D in single-sector combustor (Yamamoto et al., 2009).

Rich-lean combustor

Staging fuel nozzle

7% MTO 30% MTO 85% MTO 100% MTO

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 efficiency 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 wasn’t 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 condition in the ICAO landing and takeoff cycle for the RQL and premixed combustor.

272

Premixed Aero Engine Combustors

Cyclone Premixing cyclone flame zone Pilot Pilot/cyclone interaction zone Pilot recirculation zone

Figure 9.26.╇ Scheme of the TAPS burner (Dodds, 2005).

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 environmental concerns: local air quality around airports and atmospheric impact.
9.4.3.2╇ The GE Twin Annular Premixing Swirler (TAPS) Combustor Development for the TAPS combustor started in 1996 (Dodds, 2005). It is the only premixing combustor that has reached the TRL level 7 with flight tests of the GenX engine, and entered flight testing on the target airplanes Boeing 747–8F and 787. So far it is also the only premixing combustor for which the development of a family for different classes of thrust and pressure ratio has been announced, in addition to the GenX for the CFM 56 and recently the Tech X for the Bombardier Global 7000 business jet (Croft, 2010). Because of the competitive stage of the development, only limited information is available in the public domain. A scheme of the burner is shown in Figure 9.26. The pilot uses a simplex atomizer to spray the fuel onto a swirl cup from where it is atomized in an air blast mode between the two axial air streams. It has the same principal design features as the ECCP pilot injector of Figure 9.6. The main is developed from the cyclone swirler of Figure 9.17; the insert shows the spray issuing from the injector in stagnant air. The liquid jets enter the swirl channel within a coaxial air stream similar to the investigation of Leong and colleagues (2001). Again, the two flows are separated by a lip, which is cooled with bleed air from the outer pilot swirler. The interplay between the zones and the anticipated flame is shown in the scheme of Figure 9.27 (Dhanuka et al., 2009). With co-rotating swirl, the pilot generates a central recirculation zone, which stabilizes the pilot flame. The main flame stabilizes in the shear layer between pilot and main. The small recirculation behind the lip, termed LRZ, is important for the stabilization of the

9.4╇ Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors

273

Main air

CRZ

Main flame Mixing layer Pilot flame

Main fuel Pilot air Pilot fuel LRZ

PRZ

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 representation 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 conventionally cooled. A sketch from a patent (EP 1445 540 A1, 2004), Figure 9.29, displays 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
LTO NOx (g/kN)

50 45 40 35 30 25 20 20 30 40

CFM56-5B/P CFM56-7B CFM TAPS

50 60 70 LTO CO (g/kN)

80

90

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 combustors 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 necessary 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
Max smoke No.
CFM56-5B/P CFM56-7B CFM56-5B/PDAC CFM56-7BDAC CFM TAPS

10

5

0 20

25 30 Engine pressure ratio

35

Figure 9.31.  Comparison of ICAO LTO smoke for CFM 56 SAC, DAC and TAPS combustor (Mongia, 2007).
Takeoff Fueling mode Current RQL GEnx TAPS switch-over point Climb

EINOx(gNOx/lbFuel)

• Pilot + Main mode Cruise • Pilot only mode Approach Taxi

Combustor inlet temperature

Figure 9.32.  Staging diagram for the GenX TAPS combustor (Maurice et al., 2009).

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 staging 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.
9.4.3.3  The Rolls Royce Lean Combustor This section describes the Rolls Royce lean combustor. A sketch of the burner and the anticipated flow field is shown in Figure 9.33 to illustrate some of the general ideas behind the design (Nickolaus et al., 2002). It consists of three axial

276

Premixed Aero Engine Combustors

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 pressure 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

C-A

C-F

C-G

Figure 9.34.  RR lean burn rig fuel-injector configurations (Lazik et al., 2008).
∆ptot outer ann.

∆Ptot comb.

∆ptot inner ann.

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 isothermal flow and the pressure distribution in the combustor, illustrating the bifurcation 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 variant 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 60–70 percent is given for the burners and consequently 30–40 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 temperature 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.

9.4.4  Operability Aspects of Partially Premixed Aero Engine Combustors Although operability is more important than emissions, the latter drives premixed combustor development, and operability issues are only worth solving if the emission 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.
9.4.4.1  Emissions at Part Load Emissions at ground level have immediate impact on human health and safety in and around airports. Unburnt hydrocarbons and CO are regulated. Therefore, good combustion efficiency is required. New large engines with RQL combustors have UHC idle emissions near zero. CO is low enough to guarantee nearly 100 percent combustion efficiency. An obvious step for internally piloted lean burners would then be to take the RQL burner and put it into the premixed combustor as a pilot

9.4  Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors

279

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 temperature 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.
9.4.4.2  Stability and Lean Blowout Premixed nozzles have lower LBO margins than RQL combustors because of the dilution of the pilot flow by the main. Configurations with a pilot zone recess, as in Figures 9.23 and 9.26, have the advantage to limit that dilution near the stabilization region but need to cool the recess. This option is more purely exploited in the trapped vortex combustor (Sturgess et al., 2005). It is claimed that a smaller recirculation zone is needed for the same LBO and, therefore, lower part load NOx can be achieved. For an internal pilot surrounded by the main flow, there seems to be an advantage for a flame stabilization with lower swirl and outer recirculation of the pilot as in the bifurcated mode, compared to stabilization with an inner recirculation zone (Lazik et al., 2008); see again Figure 9.37 for a comparison of the flow fields. The same argument applies to the explanations of higher NOx at part load: less mixing with the main minimizes dilution and produces richer stoichiometry of the burning zone. On the other hand, if more than one staging point is allowed, making use of circumferential staging, the pilot effective area can be made smaller than a corresponding RQL nozzle and lower high-power NOx and higher LBO can be achieved at the same time at the expense of higher control complexity.

280

Premixed Aero Engine Combustors

Stability at idle is not the only requirement. Transient stability during slam deceleration must also be fulfilled. For premixing combustors, this obviously implies fuel staging. In addition, hail and rain ingestion during flight require a good stability margin 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 stoichiometry 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 tolerable 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 different phases of the ignition process (see Lefebvre, 1998):

1. 2. 3. 4. 5.

Formation of a flame kernel Flame propagation in primary zone Establishment of a stable flame at the burner Light around, that is, ignition of neighboring sectors Pull away, that is, establishment of steady idle operation

To achieve the formation of a flame kernel successfully, enough spray must be near the plasma at the moment of its formation. The fuel has to get to the spark igniter region. As in conventional combustors, the axial position of the igniter and the injector flow field influencing the spray cone need to be matched. Now, the spray has to cross the main flow field, as shown in the schematic of Figure 9.27 with the mixing layer between pilot flow and corner recirculation. With the corner recirculation, it can be transported to the spark plug region (see Figure 9.29). It is obvious that quick radial transport is necessary for the outer region of the spray. Not surprising, Lazik and colleagues (2008) report that the high swirl wide cone pilot on the right of Figure 9.37 gives better ignition results. One possible way to cross the mixing layer, then, is ballistic transport of big droplets. However, the mixing layer as, for example, depicted in Figure 9.27 is not thin. Therefore, the large-scale structures observed in almost all recirculating swirling flows (see Midgley et al., 2007) might play a decisive role in promoting the radial transport of spray through that mixing layer. As this process is intermittent and the plasma from the spark exists only a limited amount of time, repetition of the spark will greatly advance the probability of successful kernel formation. It has been also reported (Marchione et al., 2007) that turbulence can lead to a considerable elongation of the spark up to 20 mm, which enables penetration of energy through the forward flow into a central recirculation zone. Once ignition of the kernel is achieved, the flame needs to reach the injector. For that purpose, a recirculation is necessary (Marchione et al., 2007). The flame

9.4  Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors

281

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 penetrate 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 burners; 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 penetrate 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 spacing to satisfy the issue. For the development of the aerodynamics of internally staged premixing combustors, 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, understandable that all earlier developments tried to separate the critical function of staging into separate zones. The ability to mature internally staged premixers to airworthiness is also evidence for the coming of age of design methods and a more rational combustor design process.
9.4.4.4  Thermal Management The combustor inlet temperature at higher operating condition is considerably higher than the evaporation temperature of kerosene. Furthermore, heat from flame radiation also enters the nozzle. At temperatures reaching 200°C, pyrolysis reactions occur within the fuel that lead to the formation of deposits that will ultimately block the fuel lines. The repartition of the fuel will then produce hot spots and ultimately damage the liner. The problem also exists for unstaged burners, and a number of strategies have been developed to prevent coke formation (Lefebvre, 1998):

1. 2. 3. 4.

Reduction of fuel flow passage area to decrease hot section residence time Insulating air gaps Use insulating materials like ceramics 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 overheating 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
Mains air

Premixed Aero Engine Combustors
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.
9.4.4.5  Combustor Pressure Loss, Influence of Diffuser and Annulus The pressure loss together with combustion efficiency has a direct influence on the efficiency of the engine. However, because the pressure loss of the liner is the energy source driving the quick homogenization of fuel and air, which is, as we have seen, the main idea behind premixed combustion in aero engines, it does not make sense to try to achieve very low liner pressure losses. Reported values exceed 4 percent for premixing combustors. For premixing combustors with 60–70 percent of the combustion air passing through the burners, the flow from the diffuser to the burners and around the liners is largely different from the one in RQL combustors. The air comes from the compressor in a circumferentially homogeneous manner to enter the liner at discrete angular positions defined by the burner inlets. This entails a larger than before circumferential movement of the air from the sector boundaries to the burner. Also, the burner feed arm or the struts of the burner itself will distort the inflow to the liner. As we have seen, small variations in homogeneity can have nonlinear effects on emissions and especially on NOx. As an example of the order of magnitude of the effect, the experience of a double annular combustor is cited (Oda et al., 2003), where the feed arm to the pilot stage, positioned inboard, was at the same angular position as the centerline of the main burner. An increase of 80 percent against the version where the pilot feed arm did not obstruct the inflow to the main was measured. The size of the increase due to the installation suggests that some increase

9.4  Partial Premixing Combustors, Lean Premixing or Lean Direct Injectors

283

of NOx formation compared to an unobstructed flow is unavoidable, unless a compensation of the perturbations of the air feed can be achieved with the fuel system. Because of the multiple swirls in the injectors, this is a difficult proposal.
9.4.4.6  Combustor Outflow The premixing combustor must fulfill requirements of temperature profile and pattern factor. Of particular importance is the mixing of pilot and main efflux to combine to the desired temperature profile (see Figures 9.28 and 9.36). However, it is not probable that a premixed combustor would replicate the temperature profile of an RQL combustor. To tailor temperatures near hub and tip, some film cooling air near the combustor exit will be needed (Mongia, 2007). Since there are no more jets to mix in secondary or dilution air, eventual inhomogeneities of the primary zone temperature distribution will persist for a longer time. However, compared to the temperature distribution in the quench zone of an RQL combustor, premixed combustors should have an advantage to achieve lower pattern factor and indeed, expectations have been voiced that lower SFC could result with a higher turbine inlet temperature made possible by a lower PF (Maurice et al., 2009). Another difference to RQL combustors that cannot be avoided is the net swirl exciting an annular combustor dominated by the swirl of the injectors. 9.4.4.7  Combustion Oscillations Combustion oscillations don’t only occur in lean combustors; the engine rumble during acceleration to idle is common to conventional combustors. However, high-power oscillations as experienced during the introduction of dry low NOx industrial gas turbines in the field must not be repeated in aero engines for obvious reasons. Since the industrial gas turbine community has an advantage of fifteen years in service experience with premixed combustors, much more information is available on combustion oscillations in published form. Here, only some general differences to industrial engines relevant to combustion dynamics shall be enumerated. For a number of reasons, aero engine combustors are shorter and have less volume, such that longitudinal as well as circumferential acoustic modes reside in a higher frequency band. As described earlier, because of their higher exit temperature at full power, they use film cooling, which provides more damping than in convectively cooled premixed gas turbines. The liquid fuel evaporation adds another term to the total convective delay between burner and flame zone and forces additional dispersion on the distribution of delay times, which makes the heat release less concentrated. In addition, the contribution of the pilot to heat release will be higher. So far, all this is rather in the direction of a lower propensity to oscillations. On the other hand, the coherent turbulent structures, responsible for some of the positive mixing features presented, may also incite combustion oscillations (see Dhanuka et al., 2009). In general, it is the experience of the industrial gas turbine community that the propensity to oscillations of lean premixed combustion can be reduced with a higher pilot fuel flow. A combustor control method relying on this correlation has been implemented by Kokanovic and colleagues (2006), leading to higher NOx

284

Premixed Aero Engine Combustors

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.
9.4.4.8  Alternative Fuels Considering the recent media coverage and the vigorousness with which biofuels are advocated by aviation stakeholders, they or other alternative fuels need to be considered in the operability section. However, the present discussion will limit itself to the so-called drop in fuels, which satisfy all criteria formulated for kerosene. All other candidates are yet too far away from practical use in an aero engine to be examined in connection with the described premixing combustors. A considerable international research effort is under way, mostly directed to RQL combustors, to characterize the differences in the combustion behavior that emanate from those fuels. However, some preliminary conclusions toward premixing combustors can be drawn from the information currently available. Most of the drop in fuels that are discussed are products of Fischer-Tropsch synthesis. An advantage of some of those fuels is their very low aromatics content, which can lead to a reduction of soot emissions of up to 75 percent with reference to particle mass, as recently measured for a CFM 56 (Bulzan et al., 2010). This seems also possible for premixed combustors, only that the concept already widely reduces the soot emissions, such that the relative impact is reduced. Most synthetic kerosenes have smaller distillation curves with the advantage of faster evaporation, but also the possible disadvantage of reduced LBO and ignition at low temperatures, should the fuel only begin to vaporize at higher temperature. Other physical aspects of the fuel, namely surface tension and viscosity, influence atomization and with that again evaporation (Mondragon et al., 2011), but also, in case of droplet slip, fuel placement. As we have seen, the performance of current internally staged premixing combustors arises from a careful balance of flow features, which might be disturbed by whatever changes the flame location. On the contrary, the performance of RQL combustors is decided to a large part by what happens in the quench zone and the interaction of mixing jets with the main flow, such that they are less sensible to what happens in the primary zone. 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 relatively 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 placement, 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 combustion 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.

9.5 Summary
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 continue 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 sustained 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

temperature. This strongly favors single annular combustors. Autoignition times of kerosene at high-power conditions preclude full premixing in premixing channels before the combustor that could be safe from flashback. Premixing is therefore completed in a combustor with lifted flames. Flexible internal piloting has to ensure stability and the appropriate flame lift-off. A multitude of operability requirements, including ignition, absence of combustion oscillations, and thermal management, has to be solved before airworthiness is reached. All of those have a potentially detrimental effect on NOx reduction. Together with the moving target of engine pressures and combustor exit temperatures, this explains why it took so long for the technology to mature enough to be introduced replacing optimized RQL combustion with a step change in NOx emission. Having made the transition into service, it can be expected that further emission reductions will be possible with a further optimization using refined design tools and a penetration of lower thrust classes with premixing combustors can be expected.

REFERENcES

Altemark, D., and Knauber, R. (1987). “Ergebnisse von Untersuchungen an einem Vormischbrenner unter Druck mit extrem niedriger NOx Emission.” VDI-Berichte 645: 299–311. 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 90–020, 161–74. 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: 49–67 . 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 99–7008. 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: 815–21. Brandt, M., Hassa, C., Kallergis, K., and Eickhoff, H. (1994). “An Experimental Study of Fuel Injectors for Premixing Ducts.” ICLASS-94, Paper VII-9, 750–7 . Brundish, K. D., Miller, M. N., Morgan, L. C., and Wheatley, A. J. (2003). “Variable Fuel Placement Injector Development.” ASME GT-2003–38417 . 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 GT2010–23524. 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-79–1320.

References
Correa, S. M. (1992). “A Review of NOx Formation under Gas Turbine Conditions.” Combust. Sci. and Tech. 87: 329–62. Croft, J. (2010). “Engine Buzz: Blisks, Sabers and Safran.” Flight International (October 26–November 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: 2901–8. 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-87–2035. 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 2012–0936. 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., Kröner, M., and Sattelmayer, T. (2004). “Flashback in a Swirl Burner with Cylindrical Premixing Zone.” Journal of Engineering for Gas Turbines and Power 126(2): 276–83. Futamura, H., Fukuyama, Y., Nozaki, O., and Hayashi, S. (2009). “TechCLEAN Project: A Technology Development Program for Environmental Aspects of Small Passenger Aircraft.” ISABE-2009–1196. Greenfield, S. C., Herberling, P. V., Kastl, J., Matulaitis, J., and Huff, C. (2004). “HSCT Sector Combustor Evaluations for Demonstration Engine.” NASA/CR-2004–213132. Greenfield, S. C., Heberling, P. V., and Moertle, G. E. (2005). “HSCT Sector Combustor Hardware Modifications for Improved Combustor Design.” NASA/CR – 2005–213322. 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 Dörr, Th. (2005). “Validation Methodology for the Development of Low Emission Fuel Injectors for Aero-engines.” ISABE-2005–1143. 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, September 2–6. 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: 853–63, Springer-Verlag. Jones, R. E. (1978). “Gas Turbine Engine Emissions – Problems, Progress and Future.” Prog. Energy Combust. Sci. 4: 73–113.

287

288

Premixed Aero Engine Combustors
Kastl, J. A., Herberling, P. V., and Matulaitis, J. M. (2005). “Low NOx Combustor Development.” NASA/CR-2005–213326. Kokanovic, S., Guidati, G., Torchella, S., and Schuermans, B. (2006). “Active Combustion Control of NOx and Pulsation Levels in Gas Turbines.” ASME GT2006–90895. Lazik, W., Doerr, Th., and Bake, S. (2007). “Low NOx Combustor Development for the Engine 3E Core Engine Demonstrator.” ISABE-2007–1190. 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 GT2008–51115. 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: 542–6. 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” 1076–84. Lyons, V. J. (1982). “Fuel/air Nonuniformity – Effect on Nitric Oxide Emissions.” J. AIAA 20: 660–5. Maquisten, M. A., Holt, A., Whiteman, M., Moran, A. J., and Rupp, J. (2006). “Passive Damper LP Tests for Controlling Combustion Instability.” ASME GT2006–90874. 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 GT2007–27924. 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 2003–2657 . Mongia, H. (2007). “GE Aviation Low Emission Combustion Technology Evolution.” SAE 2007–01–3924. 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-2002–30409. 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: 193–203. 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: 98–120. Sokolowski, D. E., and Rohde, J. E. (1981). “The E3 Combustors: Status and Challenges.” AIAA-1981–1353.

References
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: 193–217 . Tacina, R. (1990). “Low NOx Potential of Gas Turbine Engines.” AIAA-90–0550. Tacina, R., Lee, Ph., Wey, Ch. (2005). “A Lean-direct-injection Combustor Using a 9 Point Swirl-venturi Fuel Injector.” ISABE-2005–1106. 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-2004–53263. Tacina, R., Mao, Ch-P., and Wey, Ch. (2003). “Experimental Investigation of a Multiplex Fuel Injector Module for Low Emission Combustors.” AIAA 2003–0827 . 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 – 2002–211347 . Umeh, Ch., Kammer, L., and Barbu, C. (2007). “Active Combustion Control by Fuel Forcing at Non-coherent Frequencies.” ASME GT2007–27637 . 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, 1–11. 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 GT2009–59852. 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 GT2010–23206. Zeisser, M. H., Greene, W., and Dubiel, D. J. (1982). “Energy Efficient Engine Combustor Test Hardware Detailed Design Report.” NASA CR-167945, PWA-5594–197 .

289

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 environmental 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 second part, where several specific design solutions are presented. These are grouped into different classes of combustors according to fuel type and the pollutant emission 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 drivers and how they interact with the electrical grid as a whole, the power plant, and the gas turbine are also included in Chapter 1. Beyond Energy Conversion Taking the thermodynamic cycle of gas turbines into consideration, the main function of the combustor is to provide a high-temperature fluid flow for driving the subsequent turbine. For older gas turbine designs dating from before 1980, the requirements of the gas turbine combustion system were quite moderate and so the development effort was kept low. Pollutant emissions were of only peripheral concern, as long as soot and yellow smoke from nitrogen dioxide formation were below the limits of visibility. These requirements were met using fairly simple combustors, which were easy to operate without sophisticated control and periphery. Because of the enormous increase of turbine inlet temperatures in high-efficiency gas turbines and the stricter emissions regulations, the design of combustors for gas turbines has become a key technology. Since 1980, the abatement of pollutant emissions has entirely driven combustor development for industrial gas turbines. This
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, allowing operation of the engines on at least two fuels for standard operation or one standard plus one backup fuel, respectively. Fuel switchover capability between primary 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. Operating Characteristics and Combustion Efficiency The wide operating range of stationary gas turbines and the transient processes at startup are important combustion technology challenges and, indeed, control the complexity of today’s solutions. During startup and at low loads, the temperature levels are insufficient for complete oxidation of the fuel, combustion intermediates are emitted in substantial concentration, and combustion efficiency drops below 100 percent. Industrial gas turbines in normal service are not usually operated in this load range. Traditionally, the development of new combustion technologies focused on providing low NOx emissions at high load while increasing the combustion temperature. Recently, however, additional requirements increasingly determining the development targets of new combustion systems have to be fulfilled. These are caused by the changing operational mode foreseen for gas-turbine-powered combined cycle plants. Because of the increasing share of renewables in power production, mainly created by wind energy, combined cycle power plants increasingly operate at intermediate load and with daily load cycles. In the framework of the improvement of the load-following capability of single as well as combined cycle gas turbine plants, the development of combustors with better combustion efficiency and lower emission of combustion intermediates at low load has become an important development target for future combustion systems. 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 compounds (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 challenges 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 current 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 challenging 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 thermoacoustic 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. Safety and Reliability Thermal integrity of the combustion chamber is an important safety aspect during normal operation. To save cooling air and to achieve optimum oxidation of carbon monoxide, the design temperature of combustor liners approaches the maximum allowable material temperature in normal operation (1100 K–1200 K). This is achieved by careful tailoring of the cooling to the local heat load. For low-emission combustion systems with premix burners, flashback and autoignition in the premixing zones have to be avoided at all costs. Achieving safe and reliable operation of premixed systems at increasing cycle pressure ratios and flame temperatures and for fuels of higher reactivity remains a challenge in combustion engineering for gas turbines. Combustor Integration The overall engine design imposes numerous geometric constraints that limit the combustion chamber topology and length and that influence the thermoacoustic system properties. • One of the very few constants in the historical evolution of gas turbine combustors is the design of the aerodynamics of combustion chambers for low-pressure losses not exceeding 3–4 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 capabilities of the gas turbine with steep load gradients up to 50 MW/min during transient 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 systems 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 concern. 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 system stimulates continuous design efforts aiming at decreasing the complexity of current solutions and at using components that can be manufactured with better economy.

10.2  Flame Types
The optimum combustion technology in gas turbine combustors is largely determined by the type of fuel used and emissions to be achieved. Ultimately, all combustion technologies are based on premixed, non-premixed, or partially premixed flames, representing the blend of the two limiting cases. In early gas turbines, non-premixed combustion was employed because of its inherent reliability, its wide operating range, and its lower system complexity. However, because of the high system specific air excess, premixed combustion at low temperatures below 2000 K is particularly suitable for low-emission gas turbines. Non-premixed combustion can only be used in the rare cases when higher emission of nitrogen oxides are permitted or if the heating value of fuel is so low that the maximum flame temperature does not reach the range of intense formation of oxides of nitrogen. The non-premixed combustion with an additional injection of inert substances such as nitrogen or water (in liquid or gaseous phase) in the ignition zone of the flame is of higher technical significance because this injection leads to additional expansion work in the turbine and augments power output. Since the injection of these substances primarily leads to delay and to dispersion of the heat release into better-mixed areas of lower fuel concentration, similarities with partially premixed flames are observed regarding the formation of oxides of nitrogen. This means that the thermal effect of the reduction of the peak temperatures in the primary zone is believed to dominate over chemical effects from local fuel gasification reactions in water-rich zones with low oxygen content.

294

Industrial Combustors

10.2.1  Influence of Fuel Properties on Combustion Technologies In particular, the increasing requirements regarding the environmental impact of thermal machines will continue to require the development of new and optimization of combustion techniques. Because of the widely different characteristics of gas turbine fuels, expanding the fuel flexibility in future systems is a challenge. In the past, fuel-specific combustion techniques were successively refined and combinations of several techniques were employed in engines with multi-fuel capability. However, recent research attempting to substantially widen fuel specifications for burners using gaseous fuels may result in higher fuel flexibility with respect to chemical composition.
10.2.1.1  Gaseous Fuels Most industrial gas turbines and gas turbines for electric power generation burn gaseous fuel when in regular operation. Although natural gas continues to dominate gas turbine fuels, other gaseous fuels are of increasing interest. A request on combustion systems emerging over the past decade is the combustion of various types of gaseous fuels ranging from different blends of natural gases to syngases and even hydrogen. A classification of the different fuels burned in modern gas turbines is shown in Figure 10.1. The fuels are characterized by their Lower Wobbe Indices Woi:

Woi = LHVvol

ρ Air ρ Fuel

(10.1) 

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 heating 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 τChem:



τ Chem ∝

a sl 2 

(10.2)

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
Chemical time scale τchem (ms)

295

1

Blast furnace gas Syngas(diluted)

Natural gas

0.1

Syngas (undiluted) Carboncapture process gas LBtu Fuels 0 10 MBtu Fuels 20 30 40 Hydrogen

0.01

0.001

HBtu Fuels 50

Lower Wobbe number : Woi (MJ/m3)

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), premixing 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 propensity 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 residues gasification is bound in hydrogen and carbon monoxide. Their inert components, 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 formation is almost fully suppressed because the peak temperature in the flame is naturally 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

of concern. With regard to fuels with fuel-bound nitrogen or oxides of nitrogen (from, for example, air-blown gasification), non-premixed combustion is advantageous because partial reburning of these compounds to molecular nitrogen takes place in understoichiometric zones of the combustor. However, similarly favorable nitrogen oxide emissions as in the premixed combustion of natural gas are not achieved.
10.2.1.2  Liquid Fuels There are only niche markets for gas turbines burning crude oil, naphtha, and heavy fuel oil. Although liquid fuel is seldom used for regular operation, light fuel oil is often used as backup fuel, and these gas turbines then must often meet emission targets when running on backup fuel. In most cases, liquid fuels such as light fuel oil and naphtha are burned in non-premixed mode with water injection for NO abatement, although the premixing of such liquid fuels has a similarly high potential as natural gas with regard to the reduction of nitrogen oxide emissions if they are free of fuel-bound nitrogen. However, the simultaneous optimization of atomization, evaporation, and premixing is particularly challenging because of the low self-ignition temperatures and the short self-ignition times of such fuels. When working with conventional atomization techniques, achieving reliable operation above an upper limit of about 2 MPa chamber pressure becomes increasingly difficult because the available mixture preparation time in the burners falls below approximately one millisecond. There is no low NOx combustion process available for combustion of heavy fuel oil in gas turbines, in spite of the long tradition of combustion of this fuel type in gas turbines. Basically, the high quantity of fuel-bound nitrogen typically found in heavy fuel oils requires a combustion method with air staging, which can be realized to some extent by means of non-premixed flames. However, the typical high concentrations of sulfur, vanadium, and sodium are known to cause problems in the hot gas path stemming from hot gas corrosion and from deposits from slag in a temperature range of about 1100 K–1300 K.

10.2.2  Flame Characteristics The reaction progress in premixed as well as non-premixed flame fronts is governed by the turbulent transport of energy (temperature) and species, which are completely different in both cases. In non-premixed combustion, the chemical reaction takes place in a layer between the fuel and air. The formation of the combustible mixture is governed by the diffusion of fuel and air toward each other and toward the reaction zone formed in zones of favorable equivalence ratio. Although in turbulent flames turbulent diffusion due to turbulent mixing processes dominates on the large scales, molecular diffusion leads to fine-scale mixing required for reaction. It can be shown that locally a pseudo-laminar character of the flame front is preserved to a large extent (“flamelets”). The opposite direction of the diffusion of air and hot gas on the air side leads to an increase in the

10.2  Flame Types

297

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 emitted 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 operating 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 diffusion 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 essential. 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 turbine combustors has always been restricted by the trade-off between emissions and complexity that cannot be circumvented.

298

Industrial Combustors

10.2.3  Flame Stabilization Optimization of flame stabilization is an essential task, particularly in lean-­ premixed combustors for fuel mixtures with low burning speed. High power densities can be realized using flows with recirculation zones because additional convective upstream transport leads to the interaction of hot combustion products with the unburned mixture directly at the burner outlet. These contact zones provide a permanent, self-regenerating source of ignition from which the flame propagates into the unburned mixture. This principle allows much higher flow velocities in the downward-flowing region and a high power density of the combustor. Therefore, a major task of the gas turbine combustor is the generation of a large-scale backflow with high velocities against the main flow direction. Swirling flows are particularly suitable for this purpose because vortex breakdown occurs at sufficiently high swirl, manifested by intense reverse flow in the core area of the combustion chamber. A typical example of such a flow field is shown in Figure 10.2. Figure 10.3 shows an illustration of flame stabilization. The gray scale is a measure of the OH radical concentration in the central plane of a square combustion chamber downstream of a premix burner measured by planar laser-induced fluorescence. The unburned mixture is free of OH radicals (dark area) as these radicals are only present in high concentrations in the reaction zone (light zone). After completion of the reaction, the OH radical concentration drops from the high values in the reaction zone toward the value in thermodynamic equilibrium (medium gray area). In the large image in Figure 10.3, the mean of the OH radical concentration is shown. This has been obtained from the averaging of a large number of single shot images. The bright area on the left covering the burner exit region is the ignition zone; in the center of the swirling flow, hot combustion products are transported backward toward the burner outlet. A major reason for the exclusive use of swirl burners in combustion chambers of industrial gas turbines is that flame holders without swirl and based solely on the sudden area expansion exhibit insufficient convective transport of combustion products upstream and power densities for fuels with low burning speeds. For fuels of high reactivity, this rule does not apply because precaution against flashback is another criterion to be considered. The minimum achievable temperature of premixed combustion in gas turbine combustion chambers at the lean extinction limit is of great technical importance because it is closely related to the issue of the reduction of NOx formation. This limit only depends to a certain extent on the swirl intensity of the burner as long as a recirculation zone is provided. It is also almost constant within the technically reasonable limits of the burner size, although the extinction limit of very small burners is observed at slightly higher temperatures. As a common conceptual model, it is assumed that this effect is caused by the smaller ratio of the mean residence time in the primary zone and the time required for chemical reaction. At the lean extinction limit, even a considerable reduction of the burner size requires only a moderate increase of adiabatic flame temperature to compensate for the unfavorable ratio of

10.2  Flame Types

299

0 20

m/s
40 20 0
60 50 40 30 20 10 0 –10 –20 –30 –40

Premix burner

–20

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.

Premix burner

Instantaneous at lean blowout

-

Average at full load

Combustion chamber

Instantaneous at full load

Figure 10.3.  Laser-induced fluorescence of OH radicals in a swirl-stabilized premixed methane flame burning in the mid-plane of a square combustion chamber.

time scales and to keep flame stabilization intact. Within the permitted range for the burner pressure drop (2–4 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 ignition 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 turbine burners. For standard gas turbine fuels, the ignition zones are generally found in areas where the local equivalence ratio is not too far from stoichiometric. If a high static stability of the flame is required, the fuel is injected in a classical manner from a central lance into the shear layer between the incoming air

300

Industrial Combustors

and the recirculated hot products to provide maximum stability of ignition. A moderate reduction of NOx emissions and soot formation at the expense of deteriorating static and sometimes also dynamic stability is achieved by fuel injection toward larger radii, but only if this shifts the main reaction zone downstream into better-mixed areas. 10.2.4  Heat Release and Burnout In swirl-stabilized flames, the heat release in the combustion chamber is initiated at the fuel jets near the fuel injector and the shear layers between reactants and combustion products, respectively. From there, the flame propagates with the local turbulent burning velocity relative to the flow velocity, which usually varies from 50 to 100m/s. An essential aim of the design of gas turbine burners and combustion chambers for fuel with low reactivity is the acceleration of this process that leads to shorter combustion chambers. This is achieved by the significant increase of the turbulent burning speed st with respect to the laminar flame speed sl, which does not exceed a few meters per second for typical gas turbine operating conditions. In the lean range near the extinction limit, laminar flame speeds barely reach 1 m/s for fuels such as natural gas or premixed diesel oil. An important macroscopic measure for the influence of the turbulence of the flow on the heat release is the turbulent burning speed st (Bradley, 1992). At low turbulence intensity, the turbulent burning speed increases almost linearly with turbulence expressed in terms of the characteristic root mean square value of the velocity fluctuation. At high turbulence intensity, a maximum is reached before quenching becomes important, and finally sudden extinction due to high turbulent stirring and stretching is observed. As a rule of thumb, the turbulent burning speed can be increased by about one order of magnitude compared to the laminar flame speed with an appropriate level of turbulence. In premixed and non-premixed combustion, another important task of gas turbine burners is the generation of turbulence at the burner outlet to augment the subsequent reaction in the combustor by increasing turbulent mixing and turbulent flame speeds, respectively. The turbulence production due to the inflow of air or the air-fuel mixture, respectively, through the burners occurs almost exclusively in the shear layers of the turbulent jet. Swirl generation in the burners usually leads to the production of turbulence in the burners and to an increase of the turbulence level of the inflow. Typical levels of turbulence in gas turbine burners employing swirl reach 10–20 percent. With these values, turbulent flame speeds on the order of 10 m/s are typically achieved for fuels with low laminar flame speed. The two single shot images in Figure 10.3 show that, in the first place, the increase of the flame speed of premixed flames due to turbulent stirring is achieved by the generation of flame surface. The turbulent vortices create a wrinkled and corrugated flame surface and, particularly at lower flame temperatures near the extinction limit, the flame also exhibits zones separated from the main flame front. Subsequently, these zones react detached from the main flame front.

10.2  Flame Types

301

1100

1500

K 1700 1500
1700

Premix burner
900 1300

1300 1100 900 700 Combustion chamber

Figure 10.4.  Time-averaged temperature distribution in a swirl-stabilized premixed flame in the mid-plane of a square combustion chamber.

At moderate turbulence and at a sufficient distance from the flame extinction limit, the local processes in the turbulent flame front can be approximated by laminar flames assuming effects such as local flame curvature and flame stretch remain disregarded. This is of particular importance because it permits the study of, for example, NOx formation, to some extent using simple theoretical models such as one-dimensional laminar flames. In a rough approximation, the increase in flame speed is proportional to the increase of flame surface by turbulence. Since the production of flame surface by turbulence in reality also depends on the laminar flame speed, the result is not a simple proportionality between the increase of flame surface and the turbulent fluctuation velocity. No simple universal law exists that would provide precise results for premixed flames in gas turbine combustors, as correlations for turbulent flame speeds provide very different results for the influence of the laminar flame speed on the turbulent burning rate. Even at high turbulence intensities at the combustor inlet, the turbulent burning speed of hydrocarbon mixtures is much smaller than the mean inflow velocity, which means the reaction front can propagate only at a small angle with respect to the approach flow. The mixture is continuously consumed with increasing distance from the burner in the shear layers of the swirling jet. A typical example of the temperature distribution in the mid-plane of a square combustor is illustrated in ­ Figure 10.4. In the example case, which is typical for premixed flames in gas turbine combustors, the heat release takes place largely within one burner diameter. For the design of combustors, however, the time until complete CO oxidation has been achieved is relevant. For premixed hydrocarbon flames, it can be shown by detailed kinetic modeling that the thermodynamic equilibrium concentration is reached approximately 1 ms after the onset of reaction in the pseudo-laminar local flame fronts at typical gas turbine pressures and temperatures. It must be concluded from this result that the burnout length is mainly limited by the turbulent burning rate and therefore depends largely on the burner size. Assuming simply that the turbulent burning rate does not depend on the burner size, the result is the proportionality between the flame length and the burner diameter. In reality, the burner diameter influences turbulent length scales and the turbulent spectrum. Therefore, there are deviations from strict

302

Industrial Combustors

proportionality. Because of the rapid burnout of homogeneously premixed flames near the lean extinction limit, they typically emit only very low carbon monoxide emissions. Only if they are exposed to additional air entrainment during combustion, which leads to pockets of mixtures too lean to burn and to local quenching, is an increase of the monoxide emissions observed. These effects cannot be avoided at low loads or near idle of gas turbines with premixed combustors in fuel-staged operation (see Section 10.4). So far, it has been assumed that the premixed fuel-air mixture is supplied to the combustor at a temperature below its self-ignition limit. However, in reheat combustors and conventional combustors at part load, self-ignition is of particular importance for the reaction progress. As shown in Section 10.4, the mixing of the fuel-air mixture with burned combustion products leads to fast ignition if mixture temperatures exceed roughly 1200 to 1300 K. For this reason, the emission of combustion intermediates such as carbon monoxide in larger quantities is observed only in cases when the combustion temperature falls below these values. This can happen if colder streaks are present in the downstream part of the combustor: because the local temperature is dominated by the local concentration of exhaust gases and the exhaust gas temperature, combustion intermediates can only survive if downstream from the primary zone, areas with a low concentration of exhaust gases exist due to insufficient mixing or if the average combustor exhaust temperature is too low to reach the temperature threshold of 1200 to 1300 K. In combustors with premix burners operated at high loads, this problem appears only if turbulent mixing and flame propagation are insufficient within the combustor length. At part load with fuel staging (see Section 10.4), the interaction of air from burners not supplied with fuel with the burning mixture stemming from the burners in operation automatically leads to a low average temperature level in the downstream part of the combustor, promoting the emission of combustion intermediates, independent of the level of mixing in the combustor. Early mixing is not necessarily advantageous because this leads to quenching of the reactions in the mixture provided by the burners fed with fuel. However, selecting low mixing intensities or delayed contact between the flows generated by the individual burners also cannot fully solve the part load emission problem, because the formation of shear layers containing partially reacted mixture diluted with air cannot be easily avoided. In principle, liner cooling air can be another source of combustion intermediates in the exhaust gas, but premixed systems generally use very low coolant mass flows to save air required for achieving sufficient air overshoot in the burners. Combustion zones with heat loss are also not very relevant in gas turbine combustion, because convective heat extraction through the combustion chamber walls is low compared to the heat released in premixed combustors. In non-premixed flames, reaction progress is determined principally by the intensity of the turbulent diffusive mixing processes and the reaction rate. To put these two effects in relation to each other, a Damköhler number, Da, can be defined

10.2  Flame Types

303

that sets the time for the turbulent mixing of fuel and air in relation to a characteristic time required for reaction: Da = tMix tChem  (10.3)

In the hypothetical case of extremely small Damköhler 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 Damköhler numbers usually larger than 1. This indicates that mixing intensity governs the average reaction progress 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 achieving 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 Damköhler 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 monoxide 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 thermodynamic 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 reference, the CO equilibrium is approximately three orders of magnitude higher for stoichiometric 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 breakdown 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 kinetically 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 combustion chamber, the equilibrium CO concentration decreases but, simultaneously, 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 turbine, 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 concentration 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 several 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 industrial 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 viewpoint of combustion in industrial gas turbines, the following aspects are of particular design relevance:
THERMAL NO PAtHWAY.  With increasing temperature, the oxygen radicals become increasingly effective in splitting the molecular nitrogen in the air. The interaction of these radicals with the nitrogen leads to nitrogen oxide and a nitrogen radical, which reacts in a second step with molecular oxygen or a hydroxyl radical to nitrogen oxide under generation of an oxygen radical that closes the loop. The reason for the strong temperature dependence of the thermal formation route is the high activation energy of the initiation reaction of the oxygen radicals with molecular nitrogen, and the strong increase of the concentration of oxygen radicals in flames with increasing temperature. As their concentration in the flame front is far above the thermodynamic equilibrium, the NO production rate found there is much higher than in the zone downstream of the flamelets where their values are closer to equilibrium. This difference is essential for understanding the effect of post-flame residence time in combustors. Only if the post-flame residence time is long with respect to the kinetic time required for the reaction does post-flame NO formation lead to substantial additional NO, despite the relatively low oxygen radical concentration found in the combustion products. PRoMpt NO PAtHWAY.  The prompt NO mechanism contributes to the NO formation

of gas turbines utilizing hydrocarbon containing gas turbine fuels. In hydrocarbon flames, not only O radicals but also reactive fragments of such fuels can split molecular nitrogen in the air. Since these species are predominantly found in the fairly thin flamelets, the formation appears almost instantaneous or “prompt.” The contribution of the prompt NO formation pathway to the NO emissions of premixed flames is low in gas turbine conditions.
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 blowout limit of premixed flames. NO emitted by combustion processes is oxidized in nature to NO2, the emission of NO2 from gas turbine combustors is particularly undesirable because of the yellowish color of exhaust gases containing higher amounts of NO2, which leads to public concerns. For intensely mixing combustion chambers without overcooled areas operated at high flame temperatures,
FoRMAtion of NitRoGEn DioXiDE (NO2).  Although

306

Industrial Combustors

NO2 is converted to NO and cannot persist in the combustion chamber. However, if temperatures below approximately 1200 K exist at low load in lean zones containing NO, the undesirable conversion of NO to NO2 takes place. With increasing NO concentration in these areas (see also Section 10.3.2.1), the risk of generating high concentrations of NO2 leading to yellowish plumes increases, especially in non-premixed flames with air excess in the primary zone. In principle, the problem also exists in staged operation of premixed flames, although it can be solved easily by means of a suitable staging strategy (see Section 10.4), which reduces the NO production at low load enough so that visible NO2 emissions are avoided. 10.3.1  NO Formation in Non-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 situation 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 combustion temperature and load. However, the emissions remain far beyond the commonly 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 sufficient above a minimum flame temperature and load, but that both emission limits cannot be reached simultaneously without applying additional NOx abatement techniques, such as the injection of water in liquid or gaseous phase into the primary 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 conversion 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 Chapter 4. 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
NO/NOlimit CO/COlimit CO/COlimit

307

NO/NOlimit 1 Limits

Water injection

0% Increasing flame temperature

100% Power output

Figure 10.5.  NOx and CO emissions of non-premixed flames as function of the flame temperature 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 reaction 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 configuration. 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 instabilities 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. 10.3.2  NO Formation in Premixed Flames Modern natural gas-fired engines with premixed combustors operating at combustion temperatures below 1800°C achieve emission levels below 10 ppm. As a rule of thumb, increasing combustion temperatures by 70 K will double the NOx emissions of practical systems. The trade-off between NOx and CO emission shown in Figure 10.5 for non-premixed flames does not exist for premixed flames (see Section 10.2.4), which are not exposed to mixing with cold air or with very lean mixtures typical for staged part-load operation (see Section 10.4). With increasing air overshoot, unstaged premixed flames show a simultaneous drop in the NOx as well as the CO emissions until the lean blowout limit is approached. High CO emissions are only observed near the lean extinction limit of premixed flames. Compared with the very strong temperature effect, the NO production of premixed hydrocarbon flames at a given flame temperature exhibits much weaker sensitivity to fuel composition, as long as fuel-bound nitrogen is not present. This also applies to completely prevaporized and premixed liquid fuels.

308

Industrial Combustors

10.3.2.1  Asymptotic Case of Perfectly Premixed Flames and Minimum NO Limit Nitrogen oxide emissions from premixed flames may be influenced by the interaction between the turbulence and chemistry in the flame fronts. In gas turbine burners, large vortices are generated that interact with each other and produce a cascade of ever-smaller vortices. The size of the smallest eddies is of the same order of magnitude as the thickness of the flamelets representing flame front. Accordingly, it is possible that flame stretch through local velocity gradient leads to quenching and reignition phenomena and that local stirring influences the chemical reaction and therefore also the NO formation. To date, NO formation in turbulent gas turbine flames under pressure has been neither experimentally nor theoretically investigated in full detail. However, simple flames models have been combined with detailed kinetics to estimate the potential influence of the individual turbulence effects on NO formation. They have also been used to validate the findings from studies with NOx emission data from large experiments operated at gas turbine conditions using perfectly premixed fuel-air mixtures. The rationale behind this approach is the fact that the flame front of premixed gas turbine flames is adequately represented by an ensemble of thin flamelets, which are exposed to the turbulence effects mentioned earlier. Comparative kinetic studies in which a one-dimensional laminar flame was used as reference showed a remarkably low sensitivity of NO formation to high turbulent diffusion, to flame stretch, to afterburning of quenched zones, and to stirring of the flame front. With regard to gas turbine combustion technology, the strong temperature dependence of the nitrogen oxide formation clearly dominates over potential effects stemming from interactions between turbulence and chemistry (Sattelmayer et al., 1998). The two single shot images in Figure 10.3 showing instantaneous OH concentrations reveal that a clearly defined transition exists between the unburned mixture and the exhaust gas in premixed turbulent flames of gas turbine burners. Because of the observed low sensitivity of the NO formation on the local turbulence structure, it can therefore be assumed that the one-dimensional laminar flame is an acceptable model adequately representing local conditions in turbulent premixed flames in gas turbine burner chambers with respect to NO formation. Figure 10.6 plots the NO formation across a one-dimensional flame front where the local temperature has replaced the axial flame coordinate. Only the concentration of the oxygen radical is shown because the concentrations of the H and OH radicals, equally important for NO formation, follow roughly the trends observed for the O radical. At high pressure, the recombination of radicals is increased, so the mole fraction of the O radical is much lower. Figure 10.6 reveals that NO production is promoted by the high radical concentrations in the flame front and then drops because of the lower radical concentration in the combustion products. The pressure dependency of the different NO formation pathways of nitrogen oxides, as well as the different pressure scaling of reaction density, thermal diffusion, and load for preheating of the reactants, leads to higher nitrogen oxide concentrations at lower temperatures at gas turbine pressure. By comparing the nitrogen oxide

10.3  NO Formation
Molar No production rate, 10–3/S 6 Mole fraction γ, 10–3 5 4 3 2 1 0 γCO/10 1 0.1 0.01 0.001 (Tproducts – T)/(Tproducts – Treactants) –1 γO 0.1 MPa γCH4/10NO prod γ NO*1000 3 2 1 0 4 6 Mole fraction γ, 10–3 5 4 3 2 1 0 1 γCO/10 γO 0 γCH4/10NO prod 0.5 MPa γ NO*1000 3 2 4

309

Molar No production rate, 10–3/S

–1 1 0.1 0.01 0.001 (Tproducts – T)/(Tproducts – Treactants)

Figure 10.6.  NO formation in laminar flames (TProducts = 1840 K). The chemical reaction proceeds 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 temperature 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 examples 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 combustion 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 formation 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 CO (vol %) NO (ppmv) 1.5 MPa CO (vol %) 5

T(K) 0.1 MPa T(K) 1.5 MPa 2000 1500

T(K)

1000 500 0

3 2 1 0 1 2 3 t (ms) 4 0 5

Figure 10.7.  Comparison of the time required for heat release in laminar flames (TProducts = 1840 K).

also in combustors with long residence times because post-flame NO formation is low. However, with the expected further rise of flame temperatures in the future, the reduction of the residence times achieved either by the reduction of the flame sizes or the further augmentation of mixing will become increasingly important. Technical solutions that allow getting closer to the minimum chemical time required have the potential to reach very low nitrogen oxide emissions even at extremely high flame temperatures. Experiments with turbulent perfectly premixed systems have shown for a wide range of flame holders that NOx emissions are not influenced by the specific flame holder design and that values below the 20 ms curve in Figure 10.8 can be obtained for typical engine pressures of 1 – 2 MPa (see, for example, Lovett and Abuaf, 1992), whereas older premix burner designs produce substantially more NOx because of imperfect fine-scale mixing. This leads to the conclusion that the improvement of mixing of the fuel with the air in premix burners and the reduction of the residence time in the combustor to the minimum required for CO conversion to CO2 are the keys to achieving minimum NOx emissions in combustors. Other parameters such as flame structure, turbulence level, and so forth are of minor influence and not particularly relevant for the design of practical systems. In principle, lean-premixed combustion can also be fuel staged. In this case, the premixed first stage is ideally operated at flame temperatures of approximately 1700 K to largely eliminate the formation of nitrogen oxides. Subsequently, the temperature must be sufficiently decreased by air admixing or partial expansion so that the remaining fuel can be fully mixed with the exhaust gas before the onset of the secondary reaction due to self-ignition. Kinetic studies predict that the NO emissions of such an idealized process is approximately proportional to the share of the fuel injected in the second stage. This means that the benefit of the staged process with respect to the unstaged reference increases with increasing temperature of the combustion

Mole fraction γ

4

10.3  NO Formation
100 100

311

t-tminimum
0 ms 5 ms 10 ms 20 ms 40 ms 80 ms

0.1 MPa

t-tminimum
0 ms 5 ms 10 ms 20 ms 40 ms 80 ms

NO (ppm, 15% O2)

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 formation of perfectly premixed flames.

products upstream of the second fuel injection. However, achieving complete mixing 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 1200–1300 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 formation 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 containing nitrogen leads to almost full conversion to NO in lean, premixed flames.
10.3.2.2  Influence of Mixing Quality on NO Formation Assuming again that premixed flames in the normal operating range of gas turbines consist of an ensemble of wrinkled and corrugated flamelets (see Section 10.2.4), and considering that, because of the high concentration of combustion intermediates in the heat release zone of these flamelets the NO formation rate is also highest there, it can be concluded that the overall NOx emission of premixed flames primarily depends on the spatial and temporal uniformity of the mixture in the flame. In the very lean range, a simplified model for the NO production in imperfectly premixed flames can be made, assuming that reaction takes place in an ensemble of laminar flamelets of different equivalence ratio and assuming that the mixing state during the reaction in which the majority of the nitrogen oxide is formed is frozen. This model captures the aspect of mixing and NOx formation adequately, although neglecting post-flame mixing overestimates NOx production in the hot gas, since, in reality, turbulent mixing will continuously reduce temperature peaks. The modeling results for the nitrogen oxide emissions plotted in Figure 10.8 indicate the lower limit of the attainable values in gas turbines with perfect mixing of

312

Industrial Combustors

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 connection 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 premixing ducts unsuitable for gas turbine combustors, as they would make the combustion 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 turbine 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 (multiple 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 pressures 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 systems 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 mixture 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

Fuel mass fraction f

0

0.2 0.4 0.6 0.8

1

0 0

Fuel mass fraction f

0.2 0.2 0.4 0.4 0.6 0.6 0.8 0.8 1 1

Fuel mass fraction f

0

0.2 0.4 0.6 0.8

1

Air Fuel Fine scale mixing

4

5

Fuel mass fraction f

0

0.2 0.4 0.6 0.8

1

0

Fuel mass fraction f

0.2 0.4 0.6 0.8

1

Figure 10.9.  Dispersion of the fuel in the air and reduction of the mixture fluctuations in a co-flow of fuel and air.

using the capability of the air for the promotion of large-scale mixing is an alternative 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 vortices. 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 qualitatively 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 temporal mixture fluctuations is analytically described by the β-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 direction (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 fuelrich 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: s= g f  (10.4)



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 perceptible 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 mixing 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 deviation 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: U= g f (1 − f )  (10.5)

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 ≈ 10–5 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 mixing 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 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0

0.06 0.05 0.04 0.03 0.02 0.01 0

Figure 10.10.  Mixing quality in the exit plane of two premix burners (laser induced fluorescence, water channel experiments, left: instantaneous concentration, right: temporally averaged fuel concentration).

of gas turbine burners. Providing additional mixing length between the fuel injectors 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 substantial increase in the NO emissions if reactive components are added to the fuel. Figure 10.10 shows a comparison of the fuel concentrations of premix burners 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 combustor 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 considerations 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 additional flexibility with regard to the suppression of thermoacoustic instabilities.

10.4 Staging at Part Load and Idle
Since the performance of industrial gas turbines is controlled by the amount of fuel, and the flow rate of the air can be influenced only within narrow limits using variable guide vane control, gas turbine combustion chambers with premix burners must provide suitable fuel or air staging methods allowing the operation at different fuel-air ratios with favorable emissions. Since staging strategies based on air bleed from the compressor produce large efficiency losses at part load, these are not competitive with fuel and air staging in combustors. Only the classic design of gas turbine combustors without premixing of air and fuel has the inherent capability for broadband operation without staging, providing the required control of the fuel, or air distribution. With decreasing global equivalence ratio, the reaction zones in non-premixed flames shift dynamically toward the fuel injectors and adapt to the current fuel-air ratio (see Section 10.2.4). At the same time, temperature will fall in the burnout zone and the oxidation of CO slows. After falling below a minimum temperature, therefore, the CO emissions in the exhaust begin to rise. Further, dropping flame temperatures leads to partial quenching of the hydrocarbon oxidation and combustion intermediates (VOC) are emitted. Finally, at excessive excess air near idle, the fuel in the leanest zones of the combustor is no longer ignited and the fuel concentration in the exhaust gas rises (UHC). If water is used for the reduction of nitrogen oxides, a compromise must often be found between burnout and nitrogen oxide emissions even at high load. The transition from non-premixed to premixed combustion requires control measures for the adjustment of the fuel or the air distribution in the combustion chamber, adapting the combustion process to the requirements of the thermodynamic process at part load down to idle. As premixed hydrocarbon flames achieve the desired low emissions only in a small window of the fuel-air ratio, their inherent properties are largely incompatible with the gas turbine process. This difficulty cannot be overcome without substantial growth in complexity of the fuel or air supply systems. An advantage of using gas turbines in combined cycle results from the desire for a high-exhaust gas temperature over the entire load range to transfer sufficient heat to the steam process. As long as the exhaust temperature can be kept constant by reducing the amount of air using a compressor with an adjustable guide vane, this also leads to a relatively constant flame temperature and premixed combustion chambers can be operated in this area without any staging. Basically, the choice of the flame temperature far above the temperature at lean blowout of the combustor at full load is favorable because a wider load range can be covered without staging. However, with increasing demands concerning the abatement of NOx emissions, the flame temperature at full load moves closer to the lean extinction limit of the premixed flames and this degree of freedom in combustor design vanishes. This requires

10.4  Staging at Part Load and Idle

317

8 Tproducts 8 6 1500 K 1550 K 1600 K 1650 K 1700 K

Ignition delay time (ms)

6

4

4

2

2

0 0.3

0.4 0.5 0.6 0.7 0.8 0.9 Mole fraction products γproducts

1

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” recirculated 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 methane chemistry. Interesting, when the results are plotted over the mixing temperature (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 similar 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 1200–1300 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 potentially leading to the emission of unburned hydrocarbons or combustion intermediates 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 combustion 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 burnout depends on pressure. For the part-load staging of premixed combustion chambers, a variety of technical solutions was found. These will be described later in the case studies in more detail. However, all variants have common underlying principles, which are schematically shown in Figure 10.12. In principle, two essentially different approaches can be distinguished, depending 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 flashback 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 combustors 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 principles are commonly applied:
SWitCHinG to Non-pREMiXED CoMBUStion oR StABiLiZAtion of tHE PREMiXED

method requires premix burners with an additional fuel supply for the generation of hotter combustion zones directly downstream of the burner exit. These stabilize the flame for global equivalence ratios below the lean blowout limit of the premixed flame. In the simplest case, all fuel is switched to direct injection bypassing the premixing zone in the burners when the lean blowout limit of the premixed stage is

FLAME BY A PiLot FLAME.  This

10.4  Staging at Part Load and Idle

319

Air staging
Burner Combustor

Fuel flow Air flow & Hot gas flow

Fuel staging Premix stage Non-premixed / Pilot stage
Burner Combustor

Without longitudinal staging
Burner
Group1 Group2 Combustor

Longitudinal staging
Burner
Group1

Combustor
Group2

Figure 10.12.  Staging methods employed in combustion chambers with premix burners.
NO Variable guide vane control or Compressor bleed

0%

100%

0%

100% 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 premixed 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 premix 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 incomplete 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 development 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 stabilization. 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 generation 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 interaction between burners is used for staging.
FUEL SWitCHinG to InDiViDUAL GRoUpS of BURnERS.  Another

strategy takes advantage of the fact that low-emission combustion chambers are equipped with a larger number of premix burners. Here the individual burners do not require a second fuel supply. The fuel supply to the burners is split into several groups that can be switched on and off separately using individual stop valves downstream of the one common control valve. A more complex variant is that each burner has its own stop valve. In the classical method, the fuel supplies to the several groups of burners are successively opened with increasing load, that is, the fuel is allocated in steps to a larger number of burners. This leads to an oscillation of the burner

10.4  Staging at Part Load and Idle
Combustion efficiency 100% CO & UHC

321

0% NO

100% Variable guide vane control or Compressor bleed

0%

Power output

100%

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 emissions count.
PiLotinG of BURnER GRoUpS BY ADjACEnt PREMiX BURnERS.  The

disadvantage of the staging methods described previously does not exist with this type of staging, which is similar to the last method with respect to the required fuel system topology, the main difference being that all group valves must have the capability to control the amount of fuel flow supplied to each burner group. If the amount of fuel for the next burner group is increased gradually with the load so that the other fueled burners are not affected, almost constant nitrogen oxide emissions can be obtained over the normal load range. In this case, the last group acts as a premixer that does not generate a separate flame but that produces a uniform lean mixture that reacts in an afterburning mode after mixing with the hot combustion products of the premix burner operated. At low power, this can lead to the deterioration of the combustion efficiency with respect to the previous staging method if the afterburning of the lean mixture does not reach completion. Once the equivalence ratio of the piloted burners exceeds the lean extinction limit, they stabilize their own flames and fuel can be supplied to the next group of burners. With such piloted concepts, complete combustion and low CO emissions are achieved for bulk hot gas temperatures above 1400 K in the primary zone without significant increase of the NO

322
Combustion efficiency 100% CO & UHC

Industrial Combustors

0% NO

100% Variable guide vane control or Compressor bleed

0%

Power output

100%

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 usually 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 conflicting 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 interaction 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 longitudinal staging requires air supply to both longitudinal stages, making combustors more complex. Furthermore, each stage requires its own fuel injection systems and premixing 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.
SEQUEntiAL CoMBUStion.  All

methods depending on the reaction (afterburning) of lean mixtures during or after mixing with hot products generated by burners providing flame stabilization can be viewed as sequential combustion. Also, gas turbines

10.5  Case Studies

323

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 combustor is reduced by partial expansion and, therefore, primarily fuel is injected in the second combustor.

10.5 Case Studies
10.5.1  Non-premixed Combustors 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 criterion for the evolution of gas turbine combustors.
10.5.1.1  Natural Gas and Liquid Fuels Can-annular combustors are the most widely used combustors in heavy-duty gas turbines such as the General Electric (GE), Mitsubishi Heavy Industries (MHI), and Siemens/Westinghouse engines (Siemens). An important advantage of this type of combustor is that single elements can be tested under engine conditions in large-scale test rigs. Can-annular combustion systems consist of a number of nominally cylindrical combustors mounted in a ring inside a common, annular pressure vessel, or casing. Each of these combustors consists of fuel nozzles and air inlets (usually with swirlers), forming the head end, or burner, a cylindrical combustion chamber liner, and a transition piece that routes the combusted fuel air mixture to the turbine inlet. We will first overview the General Electric systems. The first General Electric gas turbine for electrical service with axial flow compressor, can-annular combustion system, and axial flow turbine was derived from the TG-100 turboprop engine in the 1940s, and the same can-annular architecture is still employed in all large engines for power generation. GE Energy diffusion flame systems as used on Frame 3, 5, 6, 7, and 9 gas turbines can burn a wide range of fuels from residual liquids, such as crude oil, to lighter liquids (diesel fuel and naptha), conventional natural gas, and a great variety of process gases. The can-annular layout and relatively simple construction of the “Standard” designs are shown in Figure 10.16 with the nozzle, cap, flow sleeve, liner, and transition piece assemblies identified. Fuel is supplied through a single nozzle positioned at the forward end of the liner. The fuel injector assembly includes a single large axial air swirler, and may have several gas and/or liquid fuel circuits, as well as provisions for injecting diluents for emissions control or power augmentation, including liquid water or steam. Gases are typically injected through a series of passages in the hub side of the axial swirler, mixing with the compressor discharge air as it flows through each axial swirler passage. Additional fuel gases or smaller pilot fuel circuits for ignition and part-load operation may be provided at the center of the fuel tip.

324
Combustion cover

Industrial Combustors
Retractable Combustion Combustion Turbine spark Flow wrapper shell liner plug sleeve Combustion air Film cooling air

Fuel nozzle Combustion casing Compressor Support discharge casing clamp Transition piece

Figure 10.16.  Schematic of the general electric standard non-premixed combustor.

Combustion liner Body Water injection connection Nozzle body Cap Cowl Gas fuel jet Atomizing air passage Inner oil tip Air swirler Water spray nozzle (1 of 12)

Fuel oil connection

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 atomizing 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 assembly 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 circumference 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 zirconia 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 geometry, 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 diameter but otherwise very similar to the original MNQC parts. The resulting combustor was indeed quiet, reducing dynamic pressure amplitudes from several psi peak-topeak 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 reduction 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 continues 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 surrounding the transition piece, taking a pressure drop of 1 to 2 percent relative to the compressor 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

Figure 10.20.  Gas injector of the General Electric MNQC combustor.

Figure 10.21.  Impingement-cooled transition piece of the General Electric MNQC combustor.

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 80°C 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 operation in 1939 by BBC Brown Boveri (now Alstom) has one single horizontally positioned 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 entering 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
Fuel gas Main swirler

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 established 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 products, the fuel burns as a typical non-premixed flame. Such diffusion flames usually burn in the zones with near stoichiometric air-to-fuel ratio (ϕ =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 augmentation 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 control. 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 combustor 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
10 1 2 3 9 4

Industrial Combustors
7 5 8 6

1 2 3 4 5 6 7 8 9 10

Annular combustor Inner combustor wall Outer combustor wall Burner Fuel supply Compressor diffuser Inner extraction diffuser Outer extraction diffuser Ignition torch Sight glass

Quench air Gas injector Swirler

Figure 10.25.  Compact non-premixed low NOx combustor.

full burnout. Because of the air overshoot, the temperature drops and NOx formation is reduced. Although this type of combustor showed an NOx reduction potential of a factor 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 temperatures. However, full burnout was reached in the load range covered by inlet guide vane control of the compressor (see Section 10.4).
10.5.1.2  MBtu Fuels The Alstom MBtu Burner represents a modified version of the EV Burner, described in Section 10.5.2. It is adapted for operation with highly reactive, fast-burning gases such as MBtu syngas or other gases with high hydrogen content (Reiss et al., 2002). Injection of these fuels along the air inlet slots is not an option because of the high flame velocity, which might lead to flashback into the burner cone. In addition, the higher fuel volume flow resulting from the lower heating value of syngas would lead to distortions in the incoming air profile if the fuel entered through the air slots. These constraints are overcome by combining the aerodynamics of the standard EV Burner with a fuel injection method suitable for reactive fuels. To deal with lower heating values and faster combustion velocities, the MBtu Burner has larger fuel orifices located close to the exit of the conical swirler (Figure 10.26). These are all located circularly around the burner exit and inject the fuel radially inward. Although there is no classical premixing zone for MBtu fuels, a large degree of premixing is achieved

10.5  Case Studies
Vortex breakdown Gas injection holes Combustion air

331

Flame front

MBtu-Fuel

Figure 10.26.  Alstom MBtu burner for the GT13E2 annular combustor.

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 combustors, 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 distance 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 emissions 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 7–8 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
Fuel oil (diffusion) Natural gas (diffusion) Air Natural gas (premix) Air

Industrial Combustors

Fuel oil Syngas 2/ Natural gas Syngas 1 Fuel oil (premix) Air Air Air

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 the exit 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.
10.5.1.3  LBtu Fuels The use of LBtu gases, such as blast furnace gas as a by-product of iron production, requires modifications of the gas turbine with respect to the natural gas engine. In addition to modifications to the compressor to accommodate increased fuel mass

10.5  Case Studies

333

Figure 10.29.  Alstom silo combustor with LBtu single burner.

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 reactivity 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 optimum 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

A

F

A

F

LBtu outer channel

Figure 10.30.  Alstom LBtu burners.

despite the slow mixing of the air with the fuel. This slow mixing is a characteristic 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 operation, some form of NOx abatement is needed because of higher NOx values stemming from the higher flame temperatures. The application of General Electric’s 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 oxidizer 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 governed 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. 10.5.2  Premixed Silo Combustors for Natural Gas and Liquid Fuel with Water Injection Silo combustors are predominantly applied to gas turbines with moderate firing temperatures 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

φ=1

%O2 21 15 10 5 0

Figure 10.31.  Oxygen distribution in a General Electric LBtu combustor.

Figure 10.32.  Alstom’s 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 Alstom’s premixed silo combustors consisted of a large number of small premix burners, replacing the single swirler of non-premixed combustors, 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 improvement 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 operation. At low load, the fuel flow to groups of burners is sequentially shut off in accordance with the fuel staging principle shown in Figure 10.14.

336
Combustion Spray air evaporation

Industrial Combustors
Vortex breakdown

Gas

Gas

Atomization

Gas Flame Ignition injection front holes

Figure 10.33.  Illustration and photograph of the second-generation Alstom premix burner, referred to as the EV burner.

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 hundred 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.

The second-generation Alstom (Sattelmayer et al., 1992) premix burner consists of a swirler with two offset half cones, forming two air inlet slots, as shown in Figure  10.33. This burner was first introduced in the silo combustors of GT11N1 engines. Gaseous fuel is injected along the two slots through multiple holes. The dual fuel version of this burner has a simple pressure atomizer with low injection angle and also a water nozzle for NOx abatement. In the cone, fuel and air are mixed in the swirling flow. This swirling flow breaks down at burner exit because of the sudden expansion into combustor and generates an annular swirling jet with a recirculating flow on the centerline, which then stabilizes the flame in free space. The second-generation premix silo combustor of the GT11 gas turbines is shown in Figure 10.34. Combustors of this type are equipped with either individual fuel valves for each burner or a smaller number of valves feeding groups of burners. The common staging principle is switching burners off at part load (see Section 10.4 and Figure 10.15). The flame images in Figure 10.35, taken from downstream toward the burner exit plane of a downscaled laboratory combustor, show the switching pattern for a particular operational concept from idle over part load up to base load. Fully premixed combustion is possible throughout the entire operation range. Silo combustors from Alstom do not utilize internal diffusion piloting. Instead, individual burner switching and external piloting, with the help of two

338
Fuel oil (diffusion) Natural gas (diffusion) Air Natural gas (premix) Air

Industrial Combustors

Fuel oil (premix) Air

Figure 10.36.  Siemens hybrid burner for silo combustors.

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 SGT5–2000E series engines utilize two silo combustors, and its scaled version SGT6–2000E for the 60 Hz market. The SGT5–2000E and the SGT6– 2000E are equipped with one large vertical silo combustors on each side of the turboset. 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 components concentrically arranged with respect to each other. The central part consists 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 operation, 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

Compressor exit diffuser

Annular combustor

Turbine inlet guide vanes

Figure 10.37.  Siemens annular combustor with hybrid burners (SGT5–4000F & SGT6–4000F).

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/m³, as well as all different types of liquid fuels including heavy distillates, 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, especially at elevated flame temperatures. Typical annular combustors have average residence times of approximately 25 ms. Such residence times are also achieved with can-annular systems as explained in Section 10.5.4.
10.5.3.1  Single Stage Combustors An annular combustor with Siemens hybrid burners (see Section 10.5.2) is shown in Figure 10.37 (SGT6–4000F and SGT5–4000F). Downstream of the compressor diffuser, the air is guided toward the twenty-four burners mounted along the circumference. The flow pattern in the annular combustor is also shown with arrows:

340
Main air

Industrial Combustors
Main air

Main fuel Pilot air Pilot fuel

Main fuel Pilot air Pilot fuel

Hybrid burner with diffusion pilot

Hybrid burner with premixed pilot

Figure 10.38.  Development of burner for the annular combustors of the SGTx-4000F engines.

twenty-four large recirculation zones are generated downstream of the burner exits by the co-swirling burner flows, which provide flame stabilization around the circumference over a wide operation range. The hybrid burner used in the annular combustor is derived from the older version 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 emissions, 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 mixing 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 passage 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 Conical oil spray diffusion stage

Figure 10.40.  Oil injectors of the Siemens hybrid burner.

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 swirling 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 passage. When this premix stage is switched on at higher loads, lower NOx emissions

342

Industrial Combustors
Ceramic heat shields

Carrier structure Metallic tile holder

Purge air

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 evaporation 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 measure 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 critical 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 optimized use of the air from the compressor, with the aim of maximizing the fraction

10.5  Case Studies
Control valve Main burner group

343

Control valve

Lean burner group

Gas Fuel distribution pipe work

Figure 10.43.  GT13E2 premix burner groups.

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 SGT5–4000F uses ceramic heat shields to protect the sustaining metallic structure from hot combustion gases. Since the ceramic heat shields are made of Al2O3 and can sustain temperatures of at least 1500°C, 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 combustor 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 annular combustors. The GT13E2 gas turbine was the first engine with this combination (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 (Döbbeling et al., 2007). The

344
107 111 203 205 207 209 208 210 201 202 109 112 110 108

Industrial Combustors

105

103 101 102 611 612 610 608 606 605 604 603 609 607

106 104

204 206

211 212 301 302 303 304 305 306 308 307 309 310

View against flow direction

602 601 512 511 510 509 508 506 505 507

504 502 503 312 402 410 412 501 311 404 406 408 401 409 411 403 405 407

1/4-Premix group 3/4-Premix group

18 burners 54 burners

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 combustion 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 (30–50 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
GT casing

345

Cooling shirt Vane 1

Zone 1 Zone 2 Front segments

EV burners

Combustor hood

Compressor airflow

Figure 10.45.  Annular combustor of the Alstom GT13E2 gas turbine.

Temperature (heat release)

1600 1300 900 1700

Axial velocity

T (K)
1700 0.0 1.0 2.0 3.0

1000 1500

Swirler

Mixing tube

Oil nozzle Gas Air inlet injector

Combustor

Figure 10.46.  Alstom AEV burner prototype.

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
SEV-Combustor

Industrial Combustors
Fuel injector EV-Combustor EV-Burner

LP Turbine

HP Turbine

Compressor

Figure 10.47.  Sequential combustion system of GT24 and GT26 gas turbines.

EV outer segment EV outer liner

EV front plate

EVburner

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 momentum 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 beneficial 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.
10.5.3.2  Two Stage Combustors The Alstom GT24 and GT26 gas turbines utilize a sequential annular combustor system consisting of a high-pressure premix combustor (EV combustor) followed by an intermediate pressure reheat combustor (SEV combustor) (Joos et al., 1996). The main components of these gas turbines are the axial compressor with twenty-two stages, the first combustor equipped with EV burners, the single-stage high-pressure turbine, the second combustor with SEV burners, and a four-stage low-pressure turbine. A cross-section of the sequential combustion system of the GT24 and GT26 gas turbines is presented in Figure 10.47. The combustion of the fuel in the second stage in the hot gases from the primary stage leads to very favorable NOx emissions (Eroglu et al., 2009). Additionally, owing to the sequential nature of the combustion system, gas turbines employing the reheat cycle provide unique low part-load operation with respect to emissions.

10.5  Case Studies
Single annular SEV burners combustion zone Fuel lances Vortex generators

347

(Fuel lance moved)

Figure 10.49.  Alstom SEV burner and GT24 and GT26 combustor.

The EV combustors of GT24 and GT26 gas turbines utilize internal staging features (staged premix) whereby each burner is connected to two separate fuel supply 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, external 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 pressure, 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 plenum 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 combustor 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
Primary operation (diffusion flame) Fuel – Ignition to 20% load 100% compressor air Lean-Lean operation Fuel – 20 to 50% load 60% Diffusion Partially premixed

40%

Primary zone

Secondary zone

Premixed operation Second-Stage buring - 50 to 100% Load - Transient during transfer to premixed - Primary zone converted to premixer - Primary zone extinguished Fuel Fuel 81% 100% 19% Partially premixed Fully premixed

Low NOx/CO Fuel staging

Figure 10.50.  General Electric DLN 1 fuel staging and combustion modes.

10.5.4  Premixed Can-annular Combustors
10.5.4.1  General Electric DLN 1 Combustors General Electric’s DLN-1 system evolved from the “Multi-Nozzle Quiet Combustor” discussed in Section 10.5.1.1. DLN 1-style combustors have been used in a variety of E class combustors, including the 7E and EA, 9E, 6B, 52D, 51P, and 32J gas turbines. As illustrated in Figure 10.50, the DLN-1 system is a multi-mode, axially staged, lean premixed, dry low NOx can-annular combustor, typically operating at a peak exit temperature of less than 1300°C. NOx emissions of less than 5 ppm are achieved in the most recent versions. The main component of the DLN 1 combustor is a cap and liner assembly that has room for six (or five in the 6B) primary fuel nozzles and a centerbody to accommodate a secondary fuel nozzle with a sub-pilot to stabilize a flame at its tip (Figure  10.51). The liner also includes a venturi assembly that serves as a flame stabilizer. The throat of the venturi is located just downstream of the end of the centerbody. The liner has dilution holes downstream of the venturi. The dilution holes are a fixed air bypass sized appropriately to target a combustion zone temperature. From a combustion point of view, the liner can be divided into two different zones. The area upstream of the venturi is the primary zone. This zone has numerous holes on the liner – similar to the liner on a diffusion combustor. This zone has a flame in it in lower operating modes, but in premixed low emissions mode, it serves as the premixing zone. The fuel-air mixture from the fuel nozzle mixes with the jets

10.5  Case Studies
Primary FNs Secondary FN Dilution

349

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 distinct 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 baseload 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 maintained 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 provides a stabilizing feature for the secondary flame and suppresses dynamics amplitudes. 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 circuit 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 combustor 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 downstream end of the centerbody. The cavity and the two swirlers, visible in Figure 10.52, 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 section 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 primary gas tips, operating the machine only on the partially premixed central secondary zone in each chamber. Fuel flow is then restored to the primary and premixed operation in emissions compliance resumes.
10.5.4.2  General Electric GE DLN 2 Combustors As gas turbine cycle pressure ratios and firing temperatures increased to produce higher thermal efficiency, and the scale of the machines for power generation also increased with the global availability of natural gas and increased electrical demand, pollutant emission limits remained static or were reduced. The GE 7F gas turbine was one of the first of these new machines operating at significantly higher pressures and temperatures, but subject to the emissions expectations established by the recently fielded General Electric DLN 1 system and other DLN

10.5  Case Studies

351

7F DLN2.0 Combustion System Fuel nozzle and Quaternary combustion fuel injection Forward Combustion Flow cover assembly casing Aft liner sleeve casing

Wrapper

Cap a nd pr em tube 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 transition 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 operating 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 continuously 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 1300°C, with a combustor exit temperature of over 1400°C. The fuel staging required to maintain peak temperatures below the NOx threshold as firing temperature is increased from ignition through full load was accomplished by manipulating 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 control 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 1600°C. 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 combustor 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 configuration 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 100°C 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
Fuel stage sequence Mode 1 Fuel/Air (ratio

Industrial Combustors

Mode 2

Mode 3

Mode 4 Mode 6.3 Local fuel/air Lean blow out Power output PM3 PM3 PM1 PM2 PM1

Overall

fuel/air

Fuel distribution (%)

PM1

PM2 PM2 PM1

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 differently 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 temperature to over 1470°C. 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 nozzle 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 section 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 geometry 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 simplified 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.
10.5.4.3  Siemens DLN Combustor The Siemens SGT6–5000F engine utilizes a can-annular combustion system. The lean-premixed technology of Siemens employed in can-annular combustors in the E and F class originates from joint engine development programs between MHI and Westinghouse Electric Corp. (now part of Siemens Energy Inc.). The original lean, premixed system is called the dry low NOx (DLN) combustor and was predominantly developed by MHI in the 1980s (Aoyama et al., 1991) and early 1990s (Matsuzaki et  al., 1984). Subsequent work by Westinghouse and then Siemens extended this technology to an ultralow NOx (ULN) design for application to SGT6–5000F, SGT6–3000E, and W501D5 engines. A later step was the further extension and scaling for application to SGT5/6–8000H engines. The DLN combustor consists of a central, non-premixed pilot zone with a pilot cone to provide stabilization and a number of partially premixed, main nozzles surrounding the central pilot. The pilot nozzle has a moderate to high swirl number (0.8 to >1.2), thereby creating a recirculation zone, while the main nozzles have a low swirl number (>0.4). As a result, the pilot flame is compact and intense while the main flame zone consists of a group of “long, lazy flames.” As a result, the pilot flame, utilizing less than 10 percent of the air and fuel, stabilizes the partially premixed main flames at a lower swirl number than would otherwise be possible, enabling a longer main flame with a large distribution of time lags. This distribution of time lags results in a flame less susceptible to coherent interaction with acoustic waves and, therefore, less likely to exhibit high thermo-acoustic amplitudes. Figure 10.56 shows a schematic drawing of the DLN system. The main nozzles are attached to the support housing and are divided into two groups, an “A” and a “B” stage, with every other nozzle belonging to the same stage. This staging was designed to enable operation during startup and loading in

356

Industrial Combustors
Table 10.1.  DLN staging as a function of engine load
Load range Ignition to 30% load 30 to 50% load 50% to 100% load Stages active Pilot, A stage Pilot, A + B stage Pilot, A, B + C stage

Pilot

Support housing C-stage

Transition Basket

Figure 10.56.  Siemens DLN system for can-annular combustors.

combination with the pilot nozzle. The typical loading sequence for the DLN combustor 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 airflow. 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 support 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

Figure 10.57.  Siemens DLN fuel nozzle for can-annular combustors.

Figure 10.58.  Siemens DLN liner for can-annular combustors.

through an exit hole. The combination of the inlet hole and exit hole serves to regulate 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 temperatures at the same NOx levels.
10.5.4.4  Siemens ULN Combustor The ULN combustor was developed to meet the single-digit NOx emissions requirements for F-class engines. It is also used to maintain emissions at below 25 ppm levels with increased firing temperatures. The basic architecture of the ULN system is the same as the DLN system, with a pilot nozzle, two main stages, and a C stage. However, in the ULN system the fuel is injected through the swirler vanes instead of through the sidewalls of the fuel nozzles. In addition, the ULN system has a two-stage

358

Industrial Combustors
Table 10.2.  ULN fuel staging as a function of engine load
Load range Ignition to sync speed Synchronization to 25% load 25 to 45% load 45% to base load Stages active Pilot, A stage Pilot, A + D stage (premix pilot) Pilot, A, B + D stage Pilot, A, B, C + D stage

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 emissions 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 profile 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 Figure 10.60. 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

CO, corrected normalized to peak

15 ppm design

9 ppm design

10%

20%

30%

40%

50%

Gas turbine load

Figure 10.60.  ULN CO-emissions versus load.

10.5.4.5  Siemens PCS Combustor The ULN technology was also scaled up by a factor of ~1.2 to the Siemens H class engines. This stage of the combustion system evolution is called the platform combustion system (PCS), since it is sized to be applicable to a wide range of Siemens engines. The basic technology behind the PCS system is similar to the ULN system, with many parameters optimized for the higher firing temperature of the H class engines and to further improve the manufacturability, starting reliability, and robustness of the system. The main focus of Siemens with respect to the aerodynamic design was to optimize the swirl number, swirl profile, and mixing profile of the PCS system to achieve the required combustion firing temperature and NOx emissions levels while avoiding combustion dynamics throughout the entire engine-operating regime. Based on the experience with the ULN combustors, the thermal design was optimized from the beginning, as were the combustor and transition seals. This system was first demonstrated in 2007 at the E.ON Irsching power plant near Ingolstadt, Germany (Gruschka et al., 2008).

10.6  Nomenclature
a Da f f g g LHVvol s s1 st T U Woi m2/s thermal diffusivity Damköhler number mixture fraction average mixture fraction variance of the fuel mixture fraction pdf standard deviation of the mixture fraction pdf volumetric lower heating value normalized standard deviation laminar flame speed turbulent flame speed temperature unmixedness parameter Wobbe index

MJ/m3 m/s m/s K MJ/m3

360

Industrial Combustors

Greek ϕ γ ρ τ Subscripts Air Chem Fuel Mix Products Reactants Unmixed Abbreviations HAP pdf UHC VOC
REfEREnCES

kg/m3 s

equivalence ratio mole fraction Density characteristic time air chemical fuel turbulent mixing of fuel and air combustion products reactants, mixture of fuel and air entirely unmixed flow state of fuel and air hazardous air pollutants probability density function unburned hydrocarbons volatile organic compounds

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, 247–62. 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. Döbbeling, 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., Döbbeling, 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: 41–9. 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 GT2009–59540. Gruschka, U. et al. (2008). “ULN System for the New SGT5–8000H Gas Turbine: Design and High Pressure Rig Test Results.” Proceedings of the ASME/IGTI TurboExpo, Paper No. GT2008–51208.

References
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 GT2011–45923. 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: 247–54. 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 GT2011–45373. 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: 795–800. Payrhuber, K., Jones, R. M., Scholz, M. H. (2008). “Gas Turbine Flexibility with Carbon Constrained Fuels.” Proceedings of the ASME/IGTI TurboExpo, Paper GT2008–50556. 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): 118–25. Sattelmayer, T., Polifke, W., Winkler, D., and Döbbeling, K. (1998). “NOx-Abatement Potential of Lean-Premixed GT-Combustors.” Transactions of the ASME, Journal of Engineering for Gas Turbine and Power 120: 48–59. 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 GT2011–45944. 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.

361

Index

absorption chilling, 33 acoustically excited vortex, 56 aero gas turbines emission impact, 81–3 peaking applications, 30, 58 weight constraints, 85 air separation unit, 75 air toxics, 96 alternative fuels, 64 ammonia slip, 39 anaerobic digestion, 72 annular combustors architecture, 60 premixed, 339 single stage, 339 two stage, 346 autoignition, 54, 245 biofuel, 65 bioreactor, 71 bitumen, 65 bottoming cycles, 37 can annular combustor architecture, 59 cap and liner assembly, 325 dual fuel standard nozzle assembly, 324 carbon dioxide capture and storage, 99, 212, 214 capture vs. transport/storage, 215 emission issues, 213 impurity issues, 214 specifications, 214 emissions, 83 carbon monoxide emissions, 5, 41, 222, 226, 303 formation effect of pressure, 195, 196 NOx vs. CO divergence, 31, 89, 98 fuel-air ratio effect, 181 non-premixed flame, 307 oxidation, 180

carbonic acid corrosion, 214 chlorofluorocarbons, 97 clean coal technology. See integrated gasification combined cycle climate change, 96 coal gasification, 35, 73 coalbed methane, 68 cogeneration challenges, 34 cogeneration vs. separate production, 115 district energy systems, 34 importance, 33 combined heat and power. See cogeneration combustion instabilities conditions, 56 fuel staging, 14 heat release oscillations, 55 hydrodynamic instabilities, 56 combustor architecture can, 59 can annular, 59 full annular, 60 blowoff, 50 characteristics, 4 combustion regime staged vs. non-staged, 44 design priorities, 238 design requirements, 83, 292 dry low NOx, 36. See lean premixed combustion emissions. See emissions carbon monoxide. See carbon monoxide exhaust gas recirculation, 217 lean premixed vs. RQL, 248 nitrogen oxides. See nitrogen oxides oxyfuel combustion, 225 partial premixing, 278 particulate matter. See particulate matter prevaporization, 242 exit temperature profile, 18 flame stabilization, 298 flashback, 51, 246

363

364

Index
combustor (cont.) boundary layer, 52 core flow, 51 vortex breakdown, 54 fuel flexibility, 291 fuel staged, 13, 248 approaches, 61 combustion efficiency, 14 combustion instability, 14 fuel coking, 14 heat load, 13 lean premixed, 310 mixing issues, 302 part load, 318 methods, 318 fuel switching to individual burners, 320 longitudinal fuel staging, 322 piloted premixed flame, 318 piloting by adjacent burners, 321 soot emissions, 13 stability risk, 13 fuel-air mixing. See fuel-air mixing heat release and burnout, 300 liner cooling, 18, 21 emissions, 22 designs, 21–2 fatigue cracks, 21 heat loads convection, 21 radiation, 20 non-premixed. See non-premixed combustor operating limits, 6–8, 291 premixed, 237 , See premixed combustors lean premixed combustion pressure losses, 4, 282 rich quench lean, 8–13 advanced designs, 11 challenges, 12–13 liner durability, 12 smoke control, 12 designs, 8 emission challenges, 9–10 fuel-air ratio lapse, 9 safety, 292 stability, 243, 279, 292 turbulence production, 300 variable geometry, 252 condensation nucleus counter, 126 critical velocity gradient, 52 cross-fire tubes, 59 crude oil, 64 Damkohler number, 50, 302 diesel fuel, 64 differential mobility analysis, 126 district energy systems, 34, 43 dual fuel capability. See backup fuels electrostatic separator, 126 emission policies Canada, 105 European countries, 106 other countries, 109 United States, 102 emission trading scheme, 108 emissions, 29, 57 , 82 air pollutants, 97 air toxics, 96 carbon monoxide. See carbon monoxide measurements, 100 nitric oxides. See nitrogen oxides particulate matter, 82, See particulate matter SOx, 82, 97 , See gaseous aerosol precursors, species, SOx unburned hydrocarbons, 82 effect of power, 6 emerging issues, 91 impact, 37 limits, 89 low vs. high altitude, 82 measurement, 87 policy considerations, 100 reduction, 39 national activities, 95 valuation, 114 energy efficient engine program, 248 enhanced coal bed methane recovery, 213 enhanced oil recovery, 213, 214 environmental assessment, 110 EPA SPECIATE database, 162 exhaust gas recirculation applications, 209 carbon dioxide capture, 213 chemical kinetics effect, 211 CO emissions, 222 combustion impact, 210 combustor considerations, 216 emission requirements, 213 molecular oxygen, 213 nitrogen oxides, 213 sulphur dioxide, 213 external vs. internal, 216 NOx emissions, 218 fuel-air mixedness, 223 mechanistic pathways, 219 oxygen concentration, 219 water and carbon dioxide effects, 221 radiation heat transfer, 210 transport properties, 210 experimental clean combustor program, 248 first commerical power plant, 32 Fischer-Tropsch synthesis, 73 flame holders, 244 fuel cell, 48 fuel coking, 14 fuel injectors, 15, 16, 19

Index
air blast atomization, 16 pressure atomization, 15 fuel staging. See combustor, fuel staged fuel-air mixing Damkohler number, 302 jet penetration, 18–19 mixing enhancement methods using air flow, 312 using fuel jet momentum, 312 unmixedness parameter, 314 fuel-air ratio vs. load, 43 fuels alternative fuels, 64 comparison, 65 gas, 65 associated gases, 70 backup fuels, 67 classification, 294 coke oven gas, 71 landfill gas, 71 natural gas applications, 295 liquefied natural gas, 69 natural gas vs. coal technology, 99 non conventional sources, 68 non-premixed combustors, 323 pipeline natural gas, 65 effect of composition, 67 sulphur, 68 variability, 66. See peak shaving implications, 66 premixed combustors, 334 non associated gases, 70 peak shaving, 67 propane, 67 reactivity, 294 refinery gas, 70 syngas, 73 compositions, 74, 75 future scenarios, 75 sources, 73 synthesis, 73 uses, 73 synthesis gas. See syngas liquid fuel, 63, 296 crude oil, 64 diesel, 64 diesel fuel, 64, 65 distillates, 64 jet fuel composition, 176 non-premixed combustors, 323 premixed combustors, 334 usage, 65 surrogate fuels, 177 full fuel cycle analysis, 113 gas pipeline industry, 36 gas research institute, 66 gaseous aerosol precursors, 154 evolution, 155 open questions, 170 species, 156 organics, 156 contribution to particulate formation, 161, 162 control, 157 alternative fuels, 164 flame ionization detector, 157 limitations, 158 emission profile, 162 formation, 167 mechanisms, 168 oxidative route, 167 pyrolysis route, 167 fuel impacts, 168 measurement challenges, 158 SOx, 156 formation, 165 fuel impacts, 167 fuel source, 160 measurement, 166 regulations, 157 sulphur dioxide, 160 sulphur trioxide, 160 sulphuric acid, 160 General Electric full burner staging, 62 green house gases carbon dioxide, 83 hydrochloroflourocarbons, 97 methane, 99 methane vs. carbon dioxide, 36 nitrous oxide, 97 reduction, 99 Grenelle environment forum, 109 ground based gas turbines combined cycles, 31 advanced designs, 45 issues condensers, 31, 41 noise impacts, 31, 42 thermal pollution, 31, 41 vapour plumes, 31 selective catalytic reduction, 32 cycle enhancements reheat, recuperation, intercooling, 26 designs aeroderivative, 57 heavy duty, 57 plug and play, 57 dispatchable power source, 29 flexibility, 28 fuels, 62 grid interaction, 28 microgrids, 29 remote power generation, 64 repowering, 32 advantages, 38, 99

365

366
ground based gas turbines (cont.) synchronous operation issues, 28 tandem operation, 28 water addition, 27 , 41

Index
emissions, 244 effect of wall temperature, 245 cooling methods, 251 emissions comparison, 248 NOx, 238 effect of pressure, 239 mixture inhomogenity, 241 stoichiometry, 240 exit temperatures, 240 flame holding, 244 bluff body, 244, 260 perforated plate, 252 swirl stabilization, 244, 260 flashback, 246 lean premixed prevaporized combustor, 247 NASA funded programs, 248 Vorbix combustor, 250 stability, 251 partial premixing, 250, 261 combustor operability issues alternate fuels, 284 combustion oscillations, 283 emissions, 278 exit temperature, 283 ignition, 280 pressure loss, 282 stability, 279 thermal management, 281 GE twin annular premixing swirler combustor, 272 Japanese TechCLEAN combustor, 270 lifted flames, 265 jet flame vs. bluff body