The Effects of DTBP on the Oxidation of SI Primary Reference Fuels
- A Study in an HCCI Engine and in a Pressurized Flow Reactor
A Thesis
Submitted to the Faculty
of
Drexel University
by
Xiaohui Gong
in partial fulfillment of the
requirements for the degree
of
Doctor of Philosophy
August 2005
ii
© Copyright 2005
Xiaohui Gong All Right Reserved.
iii
DEDICATION
To my family
for their support, love and patience
iv
ACKNOWLEDGMENTS
I would like to take this opportunity to thank all of the people that have
contributed and supported me to complete my PhD Thesis. Of special note, I would like
to express my sincere and profound gratitude to my advisors Dr. Nicholas P. Cernansky
and Dr. David L. Miller for their support, guidance and encouragement throughout all
aspects of this effort. I will always remember and cherish their frankness, kindness and
friendship.
Special thanks go to Professor Kyung J. Choi for his support and guidance during
the first two years of my stay at Drexel University. A big credit goes to other faculty at
Drexel University, in particular, to Dr. Mun Choi, Dr. Alan Lau, and Dr. Gary Ruff for
their help during various stages of this work. Appreciation is also extended to other
committee members: Dr. Howard Pearlman, Dr. Tien-Min Tan and Dr. Stephen V. Smith.
Special thanks go to David B. Lenhert, Jincai Zheng, Weiying Yang, and Song
Liu for sharing their valuable time with me and for their assistance through many helpful
discussion and thoughtful comments. I appreciate all my friends and fellow researchers
in the Hess Lab for their contributions and supports to my work. In random order,
Richard Billmers, Rodney Johnson, Ashutosh Gupta, Robert Natelson, Matthew Kurman,
Jamie Lane, Charles Avila, Lin Lu, Yi Ma, Mike Foster, Giyoung Tak and Nanamid
Speed.
The past and present staff at the MEM department, Hess Laboratories and
Machine Shop: William Danley, Kathie Donahue, Stephanos Karas, Richard Miller,
Mark Shiber, Lou Haas, and others have been very helpful in the completion of this
study.
v
The financial support for this research has been provided by the National Science
Foundation (Grant # CTS-9910563), the Army Research Office (Contract # DAAD1903-1-0070), and Drexel University. The initial work on atomization of liquid jets in a
lean direct wall injection model for a new gas turbine combustor concept was supported
by NASA.
Thanks also go to my friends: Min Li, Dayong Yu, Huiling Chen, Zhiqing Huang,
Fenghua Liang, Yongzhong Wu, Houping Ying, Dawei Hu and Feng Dong for always
being around and giving support in various ways.
I am very grateful to all my family members who have taken good care of me
during the past years. Their love and support has been the source for me to overcome all
the bad times. My deepest appreciation goes to my wife, Shu Zhang. Without her
support, love and patience, it would have been impossible to finish this work. I greatly
thank my parents and parents-in-law for their selfless help and support in all aspects,
including coming to the USA to help take care of my daughter and son, Samantha and
Matthew.
vi
TABLE OF CONTENTS
ACKNOWLEDGMENTS ..................................................................................................IV
LIST OF TABLES………………………………………………………………..………IX
LIST OF FIGURES .............................................................................................................X
ABSTRACT...…………………………………………………………………………..XVI
CHAPTER 1
INTRODUCTION .......................................................................................1
1.1
Motivation.........................................................................................................1
1.2
Objective and approaches .................................................................................3
1.3
Accomplishments, contributions and recommendations ..................................5
1.4
Closure..............................................................................................................6
CHAPTER 2
BACKGROUND AND LITERATURE REVIEW .....................................7
2.1
Comparison of CI, SI and HCCI Engine Combustion and Emissions .............7
2.2
Homogenous Charge Compression Ignition (HCCI) Engines........................14
2.2.1 Introduction to HCCI ......................................................................................14
2.2.2 Advantages of HCCI......................................................................................20
2.2.3
The Importance of HCCI Research ................................................................23
2.2.4
Results Using Different Fuels.........................................................................27
2.3
Fuel Additives.................................................................................................32
2.4
Models of Hydrocarbon Oxidation Mechanisms............................................36
2.5
Low and Intermediate Temperature Regime Fuel Oxidation .........................42
CHAPTER 3
3.1
EXPERIMENTAL FACILITIES AND GENERAL TEST
METHODOLOGY ....................................................................................48
The Pressurized Flow Reactor Facility...........................................................48
3.1.1
Reactor Flow Systems ....................................................................................49
3.1.2
Sampling Method and Sample Analysis.........................................................52
3.1.3
PFR Experimental Methodology ....................................................................53
vii
3.2
Engine Facility..................................................................................................57
3.2.1
Intake Manifold ................................................................................................59
3.2.2
Exhaust Manifold..............................................................................................62
3.2.3
Engine monitoring and data acquisition system ...............................................63
3.2.4
Experiment Methodologies and Approaches....................................................65
3.3
Closure..............................................................................................................65
CHAPTER 4
THE EFFECT OF DTBP ON OXIDATION OF SI PRIMARY
REFERENCE FUELS IN A PRESSURIZED FLOW REACTOR ............66
4.1
Introduction.......................................................................................................66
4.2
Results and Discussion .....................................................................................70
4.2.1
Reactivity of the SI PRFs and Their Blends.....................................................70
4.2.2
Effects of DTBP on Fuel Oxidation .................................................................75
4.3
Closure..............................................................................................................79
CHAPTER 5
EFFECTS OF DTBP ON THE COMBUSTION OF SI PRIMARY
REFERENCE FUELS IN AN HCCI ENGINE...........................................81
5.1
Introduction.......................................................................................................81
5.2
Experimental Results ........................................................................................83
5.2.1
Operating Range Definition..............................................................................83
5.2.2
The Effect of Fuel on In-Cylinder Pressure......................................................85
5.2.3
The Effect of Equivalence Ratio on In-Cylinder Pressure ...............................93
5.2.4
The Effect of DTBP Concentration on iso-Octane.........................................104
5.2.5
The Effect of DTBP on Ignition Timing ........................................................107
5.2.6
Effect of DTBP on IMEP and Cycle to Cycle Variations ..............................109
5.2.7
Observation of a Unique Phasing Phenomenon .............................................122
5.3
Discussion.......................................................................................................123
5.4
Closure............................................................................................................126
viii
CHAPTER 6
DEVELOPMENT OF A SKELETAL KINETIC MODEL FOR
PREDICTION OF PREIGNITION REACTIVITY OF PRFS .................129
6.1
Introduction.....................................................................................................129
6.2
Skeletal Modeling Methods ............................................................................132
6.3
Current Model Development ..........................................................................137
6.4
Experimental Results ......................................................................................142
6.5
Model Validation ............................................................................................143
6.6
Closure............................................................................................................147
CHAPTER 7
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS .............148
7.1
Results and Conclusions .................................................................................148
7.2
Recommendations for Future Work ...............................................................152
LIST OF REFERENCES....................................................................................................155
APPENDIX A: HYDROCARBON OXIDATION AND AUTOIGNITION
CHEMISTRY ............................................................................................167
1.
Introduction.....................................................................................................167
2.
Mechanisms of Hydrocarbon Oxidation.........................................................169
APPENDIX B: ATOMIZATION OF LIQUID JETS IN SWIRLING FLOWS USING A
LABORATORY GAS TURBINE COMBUSTOR OPERATING IN
LEAN DIRECT WALL INJECTION MODEL........................................174
B.1.
Nomenclature..................................................................................................175
B.2.
Introduction.....................................................................................................175
B.3.
Experimental Apparatus and Instruments.......................................................178
B.4.
Results and Discussion ...................................................................................182
B.5.
Closure............................................................................................................194
B.6.
Literature Cited...............................................................................................195
VITA ..............….................................................................................................................197
ix
LIST OF TABLES
Table 2-1. Primary sources for hydrocarbon emissions in SI engines............................. 11
Table 2-2. Categories of chemical kinetic models (Zheng et al., 2004) .......................... 37
Table 3-1. Bead heaters temperature set points ............................................................... 51
Table 3-2. Cooperative Fuel Research engine geometry ................................................. 59
Table 4-1. Pressurized flow reactor test conditions ......................................................... 67
Table 5-1. Engine test conditions..................................................................................... 83
Table 5-2. DTBP effect on ignition timing.................................................................... 107
Table 5-3. DTBP effect on COVIMEP ............................................................................. 121
Table 6-1. Pressurized flow reactor test conditions ....................................................... 131
Table 6-2. Skeletal chemical kinetics model of Li et al. [1996] .................................... 133
Table 6-3. Skeletal model for low temperature, NTC and intermediate temperature
regions by Zheng et al. [2002] ..................................................................... 136
Table 6-4. Active species of current model ................................................................... 137
Table 6-5. Current skeletal model.................................................................................. 138
Table 6-6. Key fuel specific reaction parameters in current skeletal model.................. 139
Table B-1. Injector configurations…………………………………….……………......180
Table B-2. Swirler configurations..…………………………...…….…………..….......181
Table B-3. Experimental conditions……………………………………………….…...182
x
LIST OF FIGURES
Figure 1–1. Flow chart of Ph. D. work and thesis organization……...………………….3
Figure 2–1. Comparisons of SI, CI and HCCI combustion processes .............................9
Figure 2–2. Typical SI engine envelope of end gas temperature and pressure
histories leading up to the point of knock (Wang, 1999)............................43
Figure 3–1. Schematic of the Pressurized Flow Reactor facility....................................50
Figure 3–2. Typical fuel reactivity map..........................................................................55
Figure 3–3. Schematic of Cooperative Fuel Research (CFR) engine facility.................58
Figure 3–4. Intake system schematic ..............................................................................60
Figure 3–5. Exhaust system schematic ...........................................................................62
Figure 3–6. Engine data source map...............................................................................64
Figure 4–1. DTBP thermal decomposition (Griffiths et al., 1990).................................69
Figure 4–2. Reactivity maps for n-heptane, PRF20, PFR50, PRF63, PRF87,
PRF92 and iso-octane from CCD experiments in a PFR ............................71
Figure 4–3. Reactivity maps for PRF87 and n-heptane at a constant n-heptane
concentration as listed in Table 4-1, cases M and L ...................................72
Figure 4–4. Reactivity maps for n-heptane, PRF 20 and n-heptane at the
PRF20 level as listed in Table 4-1, cases F and B ......................................72
Figure 4–5. Branching pathways for hydrocarbon oxidation at low and
intermediate temperature.............................................................................73
Figure 4–6. Reactivity maps for n-heptane at different concentration as listed
in Table 4-1, cases A and B.........................................................................74
Figure 4–7. Reactivity maps for iso-octane and iso-octane + 1.5% DTBP as
listed in Table 4-1, cases R and S................................................................75
Figure 4–8. Reactivity maps for PRF92 with varying levels of DTBP additive
as listed in Table 4-1, cases O, P and Q ......................................................76
Figure 4–9. Reactivity maps for PRF87 and PRF87 + 1.5% DTBP as listed in
Table 4-1, cases M and N............................................................................76
xi
Figure 4–10. Reactivity maps for PRF63 and PRF63 + 1.5% DTBP as listed in
Table 4-1, cases J and K..............................................................................77
Figure 4–11. Reactivity map of PRF50 and PRF50 + 1.5% DTBP as listed in
Table 4-1, cases H and I ..............................................................................77
Figure 4–12. Reactivity maps for PRF20 and PRF20 + 1.5% DTBP as listed in
Table 4-1, cases F and G .............................................................................78
Figure 4–13. Reactivity maps for n-heptane with varying levels of DTBP
additive as listed in Table 4-1, cases A, C, D and E ...................................78
Figure 5–1. Typical pressure traces for HCCI operation with the different test
fuels at: (a) φ = 0.42 and Tin = 410 K; (b) φ = 0.42, Tin = 410 K
and 1.5% DTBP...........................................................................................85
Figure 5–2. Two stage ignition of n-heptane and n-heptane + 1.5% DTBP at
φ = 0.39 and Tin = 410 K .............................................................................86
Figure 5–3. Two stage ignition of PRF20 and PRF20 + 1.5% DTBP at
φ = 0.28 and Tin = 410 K .............................................................................87
Figure 5–4. Two stage ignition of PRF50 and PRF50 + 1.5% DTBP at
φ = 0.42 and Tin = 410 K .............................................................................87
Figure 5–5. Two stage ignition of PRF63 and PRF63 + 1.5% DTBP at
φ = 0.35 and Tin = 410 K .............................................................................88
Figure 5–6. Two stage ignition of PRF87 + 1.5% DTBP and single stage
ignition of PRF 87 at φ = 0.57 and Tin = 410 K ..........................................89
Figure 5–7. Two stage ignition of PRF87 + 1.5% DTBP at φ = 0.39 and
Tin = 410 K ..................................................................................................89
Figure 5–8. Two stage ignition of PRF87 + 1.5% DTBP and single stage
ignition of PRF87 at φ = 0.35 and Tin = 450 K ...........................................90
Figure 5–9. Two stage ignition of PRF87 + 1.5% DTBP and single stage
ignition of PRF87 at φ = 0.57 and Tin = 450 K ...........................................90
Figure 5–10. Single stage ignition of PRF92 + 1.5% DTBP at φ = 0.42 and
Tin = 410 K ..................................................................................................91
Figure 5–11. Single stage ignition of PRF92 + 1.5% DTBP and PRF92 at
φ = 0.49 and Tin = 450 K .............................................................................91
xii
Figure 5–12. The effect of DTBP concentration on iso-octane autoignition at
φ = 0.57 and Tin = 450 K .............................................................................92
Figure 5–13. The effect of equivalence ratio on PRF0 autoignition at
Tin = 410 K ..................................................................................................93
Figure 5–14. The effect of equivalence ratio on PRF20 autoignition at
Tin = 410 K ..................................................................................................94
Figure 5–15. The effect of equivalence ratio on PRF50 autoignition at
Tin = 410 K ..................................................................................................95
Figure 5–16. The effect of equivalence ratio on PRF63 autoignition at
Tin = 410 K ..................................................................................................96
Figure 5–17. The effect of equivalence ratio on PRF87 autoignition at
Tin = 410 K ..................................................................................................97
Figure 5–18. The effect of equivalence ratio on PRF92 autoignition at
Tin = 410 K ..................................................................................................98
Figure 5–19. The effect of equivalence ratio on PRF100 autoignition at
Tin = 410 K ..................................................................................................99
Figure 5–20. The effect of equivalence ratio on PRF87 autoignition at
Tin = 450 K ................................................................................................100
Figure 5–21. The effect of equivalence ratio on PRF92 autoignition at
Tin = 450 K ................................................................................................101
Figure 5–22. The effect of equivalence ratio on PRF100 autoignition at
Tin = 450 K ................................................................................................102
Figure 5–23. The effect of equivalence ratio on PRF100 autoignition at
Tin = 500 K ................................................................................................103
Figure 5–24. The effect of DTBP addition on iso-octane autoignition at
Tin = 410 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP..............................104
Figure 5–25. The effect of DTBP addition on iso-octane autoignition at
Tin = 450 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP..............................105
xiii
Figure 5–26. The effect of DTBP addition on iso-octane autoignition at
Tin = 450 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP..............................106
Figure 5–27. The effect of DTBP addition on ignition timing reduction for neat
iso-octane at selected φ’s and Tin = 500 K ................................................108
Figure 5–28. The effect of DTBP addition on ignition timing at selected φ’s for
PRF0, PRF20, PRF50 and PRF63at Tin = 410 K ......................................109
Figure 5–29. The effect of equivalence ratio on IMEP for PRF0, PRF20,
PRF50 and PRF63 at Tin = 410 K..............................................................110
Figure 5–30. Pressure variation for eight consecutive cycles for PRF100 at
φ = 0.49 and Tin = 450 K ...........................................................................111
Figure 5–31. Pressure variation for eight consecutive cycles for PRF100 +
1.5% DTBP at φ = 0.49 and Tin = 450 K...................................................112
Figure 5–32. Pressure variation for eight consecutive cycles for PRF92 at
φ = 0.49 and Tin = 450 K ...........................................................................113
Figure 5–33. Pressure variation for eight consecutive cycles for PRF92 + 1.5%
DTBP at φ = 0.49 and Tin = 450 K ............................................................114
Figure 5–34. Pressure variation for eight consecutive cycles for PRF87 at
φ = 0.42 and Tin = 410 K ...........................................................................115
Figure 5–35. Pressure variation for eight consecutive cycles for PRF87 +
1.5%DTBP at φ = 0.42 and Tin = 410 K....................................................116
Figure 5–36. Comparison of peak pressure variation for eight consecutive
cycles for PRF100 and PRF100 + 1.5%DTBP at φ = 0.49 and
Tin = 450 K ................................................................................................118
Figure 5–37. Comparison of peak pressure variation for eight consecutive
cycles for PRF92 and PRF92 + 1.5%DTBP at φ = 0.49 and
Tin = 450 K ................................................................................................119
Figure 5–38. Comparison of peak pressure variation for eight consecutive
cycles for PRF87 and PRF87 + 1.5%DTBP at φ = 0.42 and
Tin = 410 K ................................................................................................120
Figure 5–39. Examples of cylinder pressure “phasing” during iso-octane start
up at Tin = 450 K and φ = 0.57 ..................................................................122
xiv
Figure 6–1. Reactivity maps for n-heptane, PRF20, PRF63, PRF92 and
iso-octane from CCD experiments in a PFR .............................................142
Figure 6–2. The plug flow reactor geometry for CHEMKIN calculations...................143
Figure 6–3. Comparison of n-heptane reactivity measured experimentally and
predicated using detailed and skeletal models ..........................................144
Figure 6–4.
Comparison of PRF20 reactivity measured experimentally and
predicated using detailed and skeletal models ..........................................145
Figure 6–5.
Comparison of PRF63 reactivity measured experimentally and
predicated using detailed and skeletal models ..........................................145
Figure 6–6.
Comparison of PRF92 reactivity measured experimentally and
predicated using detailed and skeletal models ..........................................146
Figure 6–7. Comparison of iso-octane reactivity measured experimentally and
predicated using detailed and skeletal models ..........................................146
Figure B-1.
Schematic of model gas turbine combustor facility……………………...179
Figure B-2. Test section detail………………………………………………………...179
Figure B-3. The effect of injection angle• on atomization at •three different axial
locations with SN = 0.86, M air = 0.889 N and M jet = 0.086N……………184
Figure B-4.
The effect of liquid jet momentum and air momentum on the mixing
•
(D = 0.840 mm, SN = 0.86, θ = 35˚, x/X = 0.2). (A) M air = 0.889 N,
•
•
•
•
M jet = 0.093 N; (B) M air = 0.889 N, M jet = 0.125 N; (C) M air = 0.889 N,
•
•
•
•
= 0.135 N; (D) M jet = 0.118 N, M air = 0.622 N; (E) M jet = 0.118 N,
•
M air = 0.889 ……………………………………………………………..185
M jet
Figure B-5.
Optimum atomization at the same air-liquid momentum rate
•
•
ratio ( M air = 9.54 M jet , D = 0.60 mm, SN = 0.86, x/X = 0.2,
•
•
•
θ = 35˚). (A) M air = 0.889 N, M jet = 0.093 N; (B) M air = 0.854 N,
•
M jet
•
M jet
Figure B-6.
•
•
•
= 0.088 N; (C) M air = 0.753 N, M jet = 0.079 N; (D) M air = 0.622 N,
•
•
= 0.066 N; (E) M air = 0.504 N, M jet = 0.054 N...................................186
Correction between nozzle diameter and air-liquid momentum rate
ratio with (A) SN = 0.49, (B) SN = 0.86, (C) SN = 1.48……………….188
xv
Figure B-7.
•
Effect of injector diameter (SN = 0.86, θ = 35˚, M air = 0.889 N,
•
x/X = 0.2): (A) D=0.344mm, M jet = 0.068 N; (B) D=0.515mm,
•
M jet
•
M jet
•
= 0.088 N; (C) D=0.60mm, M jet = 0.093 N; (D) D=0.84mm,
•
= 0.118 N; (E) D= 1.19mm, M jet = 0.15N .........................................189
Figure B-8.
Modified Correction between nozzle diameter and air-liquid
momentum rate ratio……………………………………………………..192
Figure B-9.
Generalized correlation of air/liquid momentum rate ratio for
optimum atomization…………………………………………….............193
xvi
Abstract
The Effects of DTBP on the Oxidation of SI Primary Reference Fuels
- A Study in an HCCI Engine and in a Pressurized Flow Reactor
Xiaohui Gong
David L. Miller Ph.D. and Nicholas P. Cernansky Ph.D
A promising new engine operating mode, Homogeneous Charge Compression
Ignition (HCCI), does not use traditional Spark Ignition (SI) or Compression Ignition (CI)
-combustion control systems. Instead it relies completely on the inherent preignition
chemistry of the cylinder charge to control combustion phasing and ignition timing. The
subsequent HCCI combustion process determines the rate of heat release, the reaction
intermediates and the ultimate products of combustion. Therefore, understanding the
ignition and oxidation chemistry of potential HCCI fuels is particularly important.
One option for ignition control of HCCI engines is to use small amounts of
ignition-enhancing additives to alter the ignition properties. Di-tertiary Butyl Peroxide
(DTBP) is one such additive and it has demonstrated its capacity to improve ignition of
fuels in diesel engines.
In this study, the oxidation of SI primary reference fuels (PRFs) and their blends,
and the effects of the additive DTBP on their ignition and oxidation behavior were
investigated experimentally in both an engine operating in the HCCI mode and a
Pressurized Flow Reactor (PFR). The effect of DTBP on iso-octane in the PFR shows
evidence of reactivity promotion by a chemical effect rather than just a thermal effect.
Experimental results in the engine show an ignition delay time reduction of at least
3 CAD for all tested fuels; COVIMEP improvement to <10% (a 37.5% reduction) for
PRF92 at inlet temperature of 450 K and equivalence ratio of 0.49; and extension of
xvii
stable HCCI operations for relatively high RON fuels to a broader equivalence ratio
range and to lower inlet temperatures.
In parallel to these experimental studies, an initial modeling effort was
undertaken to modify and reformulate a skeletal chemical kinetic model for the SI PRFs
and their blends. The model was developed as an extension of our previous preignition
model by modifying several reactions to incorporate recent advances in our
understanding of the relevant chemistry.
The model was also reformulated to be
compatible with the standard CHEMKIN simulation package. In general, the updated
skeletal model successfully predicted the reactivity behavior of the fuels tested over the
600-800 K experimental range of this study.
1
CHAPTER 1.
1.1
INTRODUCTION
Motivation
In the last century, the development of Internal Combustion (IC) engines has
achieved a high level of success. These engines have been gradually optimized for best
performance and emissions. In the early years, increasing engine power and reliability
were the most important goals. Within the past five decades, however, the regulation of
exhaust emissions and the decline of petroleum resources have focused attention on
development of clean and efficient engine designs.
New regulations introduced by the United States Environmental Protection
Agency (USEPA) require vehicles and powered equipment in the U.S. to significantly
reduce carbon monoxide (CO), unburned hydrocarbons (UHC), nitrogen oxides (NOx),
and particulate matter (PM).
For example, for light-duty vehicles, new emission
standards, which will be in full effect by 2009, require a 93% (for diesel engine) or
82.5% (for gasoline engine) reduction in NOx and a 87.5% reduction in PM, to levels of
NOx
< 0.07 g/mile and PM < 0.01 g/mile;
for heavy-duty diesel engines, new
regulations require a reduction of 90% for both PM (< 0.01 g/bhp-hr) and NOx (< 0.2
g/bhp-hr),
which
will
take
effect
in
2007
and
2010,
respectively
[http://www.dieselnet.com/standards/us/light.html#tier2].
Fuel efficiency continues to be a major area of public and policy interest due to its
direct relation to carbon dioxide emissions, which is the pollutant most often associated
with global warming. Light vehicles contribute about 20 percent of all U.S. carbon
dioxide emissions and approximately 40 percent of all U.S. oil consumption. Crude oil,
from which nearly all vehicle fuels are made, is a finite natural resource. Fuel efficiency
2
is also directly related to the cost of operating a vehicle and becomes increasingly
important when oil and gasoline prices rise, as has happened recently.
Tremendous effort has been devoted to improving performance and reducing
emissions of current engines, such as employing 3-way catalyst, sacrificing some engine
performance to get lower emissions, etc.
It is clear that conventional IC engines
encounter some difficulties in improving efficiency while reducing emissions and can not
meet the stringent regulations to be enforced in the next few years.
As these techniques approach their limits, new technologies are getting more
attention.
One example is a mode of operation termed Homogeneous Charge
Compression Ignition (HCCI). HCCI has been the subject of many experimental and
theoretical investigations beginning in 1979 [Onishi et al., 1979] and continuing until the
present with well over 100 papers presented so far in calendar 2005.
The Homogeneous Charge Compression Ignition (HCCI) concept promises the
advantages of compression ignition (CI) engines and spark ignition (SI) engines.
However, there are still several technical barriers that need to be overcome before HCCI
can be widely used. Improving ignition timing control and expanding operation range of
HCCI are two of the main issues.
These issues are controlled by the autoignition
chemistry which involves low and intermediate temperature reactivity and by the
subsequent high temperature oxidation chemistry. Clearly, a fundamental understanding
of the relevant hydrocarbon autoignition and oxidation processes is essential if this
advanced engine concept is to become a reality.
3
1.2
Objective and approaches
The overall purpose of this study was to improve our understanding of fuel
oxidation chemistry and to provide better control methods applicable to HCCI operation.
One option for ignition control of HCCI engines is to use small amounts of ignitionenhancing additives to alter the ignition properties. To this purpose experiments were
conducted using our pressurized flow reactor (PFR) and cooperative fuel research (CFR)
engine facilities to investigate the effect of the additive di-tertiary butyl peroxide (DTBP)
on oxidation of SI primary reference fuels. A flow chart of my work and this thesis is
provided in Fig. 1-1.
Effect of DTBP on Fuel Oxidation
Modeling
Experiment
Conduct SI PRFs + DTBP
Oxidation Experiments on
CFR Engine Facility
Extend and Reformulate
Skeletal Model for PRFs
Conduct SI PRFs + DTBP
Oxidation Experiments on
PFR Facility
Test and Refine the New
Model
Explore Interaction
between n-Heptane, isoOctane and DTBP
Figure 1–1. Flow chart of Ph. D. work and thesis organization
4
Specifically, the main activities of this study were as follows.
(1) Investigated experimentally the oxidation of SI PRFs and their blends, PRF20,
PRF50, PRF63, PRF87, and PRF92, with and without the addition of DTBP in a
CFR engine and in a PFR over the temperature range of 600 - 1000 K at elevated
pressures. The detailed description of the facilities and experimental methodology
are given in Chapter 3. The experimental procedure and results from the PFR and
engine efforts are presented in Chapters 4 and 5, respectively.
(2) Conducted mechanistic analyses, developed a skeletal chemical kinetic model
compatible with the standard CHEMKIN simulation package, and validated the
model using PFR data.
Numerical modeling is important to identify the key
reactions in the oxidation mechanism. The skeletal model was developed as an
extension of our previous preignition model [Li et al., 1996; Zheng et al., 2001 and
2002a and b] by modifying several reactions to incorporate recent advances in our
understanding of the relevant chemistry. The objective in this work is part of a long
term effort to apply skeletal models to a broad range of fuels. The modeling work
and results are described in Chapter 6.
(3) Examined and compared the current detailed model for n-heptane, iso-octane and
their mixtures with experimental results from the PFR. The detailed mechanism
had been developed by Curran et al. [1998, 2002] at Lawrence Livermore National
Laboratory (LLNL).
This comparison between the experimental and detailed
modeling results is also presented and discussed in Chapter 6.
(4) Summarized findings and made recommendations for future work. These
observations and conclusions are presented in Chapter 7.
5
A general background and literature review is provided in Chapter 2 and an expanded
discussion of hydrocarbon oxidation and autoignition chemistry is included as
Appendix A.
Before beginning the studies on the DTBP and primary reference fuels, an
experimental study on the atomization of liquid jets in swirling flows in a laboratory gas
turbine combustor operating in a lean direct wall injection (LDWI) mode (a new ultralow-emission gas turbine combustor concept) was carried out. As the first step toward
understanding the combustion phenomena in a LDWI mode, the hydrodynamic behavior
of wall-injected liquid jets in confined cold swirling air flows was investigated. As this is
separate project from the DTBP studies, it is briefly described and reported in
Appendix B.
1.3
Accomplishments, contributions and recommendations
The main contributions of this study were as follows:
(1) Elucidated DTBP’s mode of action.
(2) Provided detailed experimental data for oxidation of primary reference fuels with or
without DTBP in both PFR and engine. These experimental data were used during
the mechanistic analysis phase of this study;
(3) Expanded engine stable HCCI operations for relatively high RON fuels to a broader
equivalence ratio range and to lower inlet temperatures; and
(4) Improved and reformulated existing pre-ignition skeletal chemical kinetic models to
be compatible with the standard CHEMKIN simulation package and successfully
6
predicted the reactivity behavior of the fuels tested over the 600-800 K
experimental range of this study.
In addition, as part of an early initial study, the hydrodynamic behavior of wall-injected
liquid jets in confined cold swirling air flows was investigated and the initial breakup and
subsequent jet atomization of liquid jets in the swirling airflows was characterized.
Collectively, these efforts, which have been documented in several papers and
presentations, represent my unique contribution to the area.
The following future work is recommended:
(1) Measure species evolution information from both engine and PRF with the addition
of DTBP in order to further determine the effect of DTBP;
(2) Incorporate new advances, such as the mechanism of high temperature oxidation,
into the skeletal kinetic model and extend its range of applicability;
(3) Examine the effect of DTBP on additional fuels, including non PRF alkanes,
alkenes, aromatics and real fuels.
1.4
Closure
This introduction has provided an overview of the motivation, the research
objectives and study methodology, and primary accomplishments and contributions of
the present work. Recommendations for additional work have been provided as well.
The next chapter, Chapter 2, provides a general background and review of
literature pertinent to the research work prior to describing and discussing the present
experimental and modeling efforts.
7
CHAPTER 2.
BACKGROUND AND LITERATURE REVIEW
Many practical problems in engine operation and performance are controlled by
autoignition chemistry. Classic examples are knock in spark ignition engines and cold
start in diesel engines. The chemistry that controls autoignition in HCCI combustion is
the same as that which leads to knock in SI engines. Studies of autoignition began in the
early 1900’s when knock was first realized as a limitation on engine output and fuel
efficiency. Thus, all of the previous research work devoted to knock chemistry in SI
engines over the last 100 years is directly applicable to HCCI combustion, and there is a
wealth of literature that can be used to guide our research.
Instead of reviewing all aspects of hydrocarbon combustion, this chapter provides
background and reviews the past work related to the scope of this research program. First,
a comparison between the CI, SI and HCCI combustion processes and emissions is made.
Second, the history of HCCI engines is reviewed. Then, related research work on fuel
additives and an introduction to their effects on hydrocarbon oxidation is presented.
Finally, previous research on hydrocarbon autoignition and oxidation and on kinetic
mechanism development is discussed.
2.1
Comparison of CI, SI and HCCI Engine Combustion and Emissions
Emissions from the combustion of hydrocarbons in internal combustion engines
are major sources of pollution throughout the world. Regulations introduced by the
Environmental Protection Agency (EPA), California Air Resources Board (CARB), and
international regulatory agencies are requiring vehicles and off-highway powered
equipment to substantially reduce emissions. Significant reductions in carbon monoxide
8
(CO), unburned hydrocarbons (UHC), nitrogen oxides (NOx) and particulate matter (PM)
will be required in almost all classes of engines. Also, fuel efficiency continues to be a
major area of public and policy interest due to its direct relation to carbon dioxide
emissions, which is the pollutant most often associated with global warming.
The simplest way to improve the efficiency of an engine is to increase the
compression ratio.
However, high temperatures and pressures caused by high
compression ratios are normally associated with high NOx. Also, in spark ignition
engines, the high temperature end gases promote autoignition and knock, which limits the
maximum engine compression ratio.
Therefore, the way to improve efficiency by
increasing compression ratio also increases NOx emissions. Also, combustion strategies
that reduce NOx emissions invariably result in increased HC and PM emissions, and
conversely, strategies that reduce HC emissions almost always increase NOx emissions
[Borman, 1980; Turns, 1999; Heywood, 1988].
Generally, due to high temperatures and heterogeneous combustion of the
atomized fuel, Compression Ignition (CI) engines are very efficient, but emit a large
amount of NOx and PM and only small amounts of CO and UHC.
Modern well
controlled catalyst equipped SI engines are modest emitters of CO, UHC and NOx, and
very small emitters of PM, but are less efficient. They also require more refined fuels
than CI engines. A comparison of CI, SI and HCCI engine combustion processes and
emissions is presented in Fig. 2-1.
9
(a)
Spark Ignition:
• spark-ignited
• flame propagation
• premixed combustion
• throttled
• port-injection
• stoichiometric
(b)
Compression Ignition:
• auto-ignition
• flame propagation
• premixed and diffusive
combustion
• unthrottled
• direct-injection with swirl
• Variable stoichiometry
(lean to rich)
(c)
HCCI:
• auto-ignition
• no flame propagation
• premixed volumetric
combustion
• unthrottled
• port or direct-injection
• lean/ dilute stoichiometry
Figure 2–1. Comparisons of SI, CI and HCCI combustion
processes (Figures from Ogink, 2004)
In Spark Ignition (SI) engines, the fuel is mixed with air in the intake manifold to
form a premixed charge with equivalence ratio around stoichiometric. When the spark
plug fires, a flame kernel is formed and a flame propagates through the homogenous
charge. As flame propagation occurs, the temperature at the front --- a thin zone of
intense chemical reaction --- is high, and significant NOx formation occurs in the postflame, hot combustion products. Stratified charge SI engines, while attempting to avoid
this high temperature region, still have problems with high emissions [Aoyama et al.,
1996].
10
The thermal efficiency of SI engines depends on the compression ratio.
Unfortunately, the compression ratio is limited by autoignition of the unburned gases.
Severe autoignition leads to knock and limits engine efficiency and thereby increases
emissions. The homogenous premixed combustion in the SI engine contributes to its
very low PM emissions. The NOx formed in the flame front and post flame regime is
primarily NO. The most significant reaction mechanism forming NO is the Zeldovich
[Miller and Bowman, 1989].
CO, a primary intermediate of HC combustion, is
invariably formed and in untreated exhaust CO concentration is the highest of all
emissions.
For all types of engines, hydrocarbon emissions result from the presence of
unburned fuel in the engine exhaust. In SI engines, about 9% of the fuel supplied to an
engine is not burned during the initial flame propagation event. However, most of this
unburned fuel is consumed as a result of post combustion oxidation processes during the
power expansion stroke, including oxidation in the exhaust port during the blow down
process. Ultimately, about 2% of the total fuel flow into the engine will leave with the
exhaust, including partial reaction products, such as acetaldehyde, formaldehyde,
1, 3 butadiene, benzene, etc. [Cheng et al., 1993]. As hydrocarbon emissions represent
lost chemical energy, the UHC emission also represents a decrease in the thermal
efficiency.
There are six primary mechanisms believed to be responsible for hydrocarbon
emissions from SI engines, Table 2-1.
11
Table 2-1. Primary sources for hydrocarbon emissions in SI engines
(Cheng et al., 2003)
% fuel escaping
normal combustion
% contribution to the 2% of
unburned fuel after burnout
Crevices
Oil layers
Deposits
Liquid fuel
Flame quench
Exhaust valve leakage
5.2
1.0
1.0
1.2
0.5
0.1
38
16
16
20
5
5
Total
9.0
100
Source
•
Crevices – these are narrow regions in the combustion chamber into which the
flame cannot propagate because they are smaller than the quenching distance.
Crevices represent about 1 to 2% of the clearance volume.
•
Oil layers - Since the piston ring is not 100% effective in preventing oil migration
into the cylinder above the piston, an oil layer exists within the combustion
chamber that absorbs fuel.
•
Deposits – Carbon deposits build up on the valves, cylinder and piston crown.
These deposits are porous with pore sizes smaller than the quenching distance so
trapped fuel cannot burn.
•
Liquid fuel – For some fuel injection systems there is a possibility that liquid fuel
is introduced into the cylinder past an open intake valve. The less volatile fuel
constituents may not vaporize (especially during engine warm-up) and be trapped
in the crevices and carbon deposits.
12
•
Quenching – Most of the hydrocarbon contained in the wall quench layer diffuse
into the hot combustion products outside the layer and get consumed during the
post combustion oxidation processed. However, bulk gas quenching can occur
during the decompression and blow down processes when the temperature drops
to a low enough level.
•
Exhaust valve leakage- Exhaust valves which are normally closed may leak
UHC’s directly into the exhaust port.
In CI engines the liquid fuel is injected at high pressure directly into the
combustion chamber near Top Dead Center (TDC). The atomization, vaporization and
mixing of fuel spray with the swirling compressed air in the cylinder occurs in a hightemperature and high-pressure environment. When the in-cylinder temperature is above
the autoignition temperature of the fuel, the mixture will spontaneously ignite following
an ignition delay period.
Subsequently, any vaporized premixed charge with
stoichiometry within the flammability limits will be rapidly consumed.
Ultimately,
mixing controlled combustion dominates the remainder of the combustion process. The
inhomogeneous mixture and high combustion temperature in CI engines produces NOx
in the oxygen-rich and stoichiometric regions, and particulate in the fuel-rich regions.
NOx is formed in the high temperature regions where both oxygen and nitrogen are
available, and in the post combustion hot gas regions [Miller and Bowman, 1989]. As
temperature is proportional to load in a CI engine, more NOx is formed as the load
increases. Due to the diffusive combustion process and the presence of very rich mixture
13
regions, PM formation is unavoidable. Some of the PM is destroyed in the flame by
oxidation and the unoxidized PM becomes an exhaust emission [Schommers et al., 2000].
It is difficult to reduce both NOx and particulate simultaneously. In CI engines,
the new rules will require electronic engine controls, exhaust gas recirculation (EGR),
and improvements in after-treatment (particulate filter, NOx trap or DeNOx) to reduce
NOx and particulate levels.
The following two factors are believed to be additional sources for UHC emission
in CI engines:
•
Undermixing of fuel and air - Fuel leaving the injector nozzle at low velocity, at
the end of the injection process, cannot completely mix with air and burn.
•
Overmixing of fuel and air - During the ignition delay period evaporated fuel
mixes with the air, regions of fuel-air mixture are produced that are too lean to
burn. Some of this fuel makes its way out the exhaust. If ignition delays are
excessively long more fuel becomes overmixed.
Since in-cylinder temperatures are higher in CI engines, UHC emissions are usually
significantly less than in SI engines.
HCCI engines utilize homogeneous charge as in SI engines; however, the charge
is compressed to ignite as in CI engines. This new combustion concept provides the high
fuel efficiency of CI engines and the lower NOx and PM emissions of SI engines. Key to
the application of HCCI is to create a charge that produces a smooth heat release profile
across the entire operating ranges.
This usually requires a dilute, lean charge that
produces maximum temperatures low enough that thermal NOx emissions are
14
dramatically reduced. Due to lean, premixed operation the PM emission is lower too.
High efficiencies are achieved by operating unthrottled with high compression ratios as in
compression ignition engines. There are some other benefits with HCCI engines as well,
such as the capability of using multiple fuels.
However, due to the low combustion temperature, particularly at lower load
conditions, excess CO and UHC emissions are found in HCCI [Dec, 2002; Christensen et
al., 2001; Easley et al., 2001]. A detailed discussion of HCCI engines is given in the
following section.
2.2
Homogenous Charge Compression Ignition (HCCI) Engines
The first HCCI concept was proposed in the late 1970’s. This idea has drawn
major attention in the last decade due to the urgency to meet stricter regulations on NOx
and PM emissions. Although tremendous experimental and modeling efforts have been
brought to bear on HCCI phenomena in the past several years, only the recent advent of
electronic sensors and controls has made HCCI engines a potential practical reality
[Epping et al., 2002]. This section provides a brief history of HCCI studies, an overview
of the current state-of-the-art in HCCI technology, and a list of the R&D barriers that
must be overcome before HCCI engines can be considered for commercial application.
2.2.1
Introduction to HCCI
HCCI is an alternative piston-engine combustion process that can provide
efficiencies as high as compression-ignition (CI) engines while producing ultra-low
15
oxides of nitrogen (NOx) and particulate matter (PM) emissions, unlike CI engines.
HCCI engines operate on the principle of having a dilute, premixed charge that reacts and
burns volumetrically throughout the cylinder after compression by the piston. HCCI
incorporates the best features of both spark ignition (SI) and compression ignition (CI).
As in an SI engine, the charge is well mixed, which minimizes particulate emissions, and
as in a CI engine, there are no losses due to inlet throttling, the charge is ignited by the
high ambient pressure and temperature produced by compression, and the load is
determined by the amount of fuel in the charge, which leads to high efficiency. However,
unlike either of these conventional engines, the combustion occurs simultaneously
throughout the volume rather than in a flame region. This important attribute of HCCI
allows combustion to occur below typical flame temperatures, dramatically reducing
NOx emission. The resulting disadvantage of HCCI operation is that the engine may be
hard to start and the combustion process requires new control methods.
These
disadvantages presently restrict the application of HCCI engines. However, the potential
of the HCCI concept has motivated studies designed to understand the ignition and
oxidation chemistry of possible fuels.
The first HCCI operation was reported by Onishi et al. [1979] who measured a
unique combustion behavior they called “Active Thermo-Atmosphere Combustion
(ATAC)”, which was intermediate between SI and CI.
Achieved on a two-stroke
gasoline engine under relatively lean conditions, the ATAC process obtained lower fuel
consumption and low emissions in the region of light and medium loads, with less noise
and vibration. High speed Schlieren photographs showed that ATAC was initiated by a
multipoint autoignition without discernable flame propagation.
16
Later the same year, Noguchi et al. [1979] reported similar self-ignited
combustion in a two-stroke gasoline engine. They named the combustion process “TS
(Toyota-Soken) combustion”. High levels of HCO, HO2, and O radicals were observed
within the cylinder prior to autoignition, which demonstrated that pre-ignition chemical
reactions had occurred and that these reactions certainly contributed to the autoignition.
In a traditional SI engine, these preignition radical species are primarily associated with
end-gas autoignition, namely knock. After autoignition took place, H, CH, and OH
radicals were detected, which were indicative of high-temperature chemical reactions.
Also, the combustion process seemed to start at lower temperatures and pressure than
those for conventional CI combustion.
Following these two pioneering studies, the operating mode, renamed HCCI, has
been demonstrated on a number of two-stroke engines by several researchers. Lida [1994]
broadened the stable two-stroke ATAC combustion range by using methanol as the fuel.
Later, other alternative fuels such as dimethyl ether, ethanol, and propane were also
tested by Lida [1997] to investigate the fuel sensitivity of HCCI operation on two-stroke
engines. Honda has proven the reliability of the concept for a production two-stroke
engine by placing 5th overall in the Granada-Dakar desert race with a pre-production
motorcycle [Yamaguchi, 1997]. A pre-production two-stroke engine employing HCCI
has also been shown by Duret and Venturi [1996]. In both cases, HCCI was used to
improve combustion stability, reduce HC emissions and improve fuel economy at part
load.
Honda has a 2-stroke cycle, single-cylinder HCCI engine that operates on
gasoline and powers a motorcycle [Ishibashi and Asai, 1996]. This engine operates in
17
HCCI mode at low to moderate loads, and switches to conventional SI operation at high
loads. Even though HCCI is used over only part of the duty cycle, the engine has
demonstrated considerable advantages in fuel economy, which is 27 percent better than a
regular 2-stroke cycle engine under "real-life" riding conditions. Hydrocarbon emissions
are also reduced by 50 percent with respect to a regular 2-stroke cycle engine. However,
without emission controls, hydrocarbon emissions are still very high compared to the
current emissions standards.
While efforts on two-stroke HCCI engines have made significant progress, the
efforts on four-stroke HCCI engines have achieved only marginal success.
The
inherently high Exhaust Gas Recirculation (EGR) rate of two-stroke engines helps to
control the rate of heat release, and thus the knock intensity of the engine. For a fourstroke engine, controlling the rate of heat release with little or no EGR while maintaining
the engine performance is an obstacle to achieving HCCI operation.
The first success in applying HCCI combustion to a four-stroke engine was
achieved by Najt and Foster [1983]. They successfully conducted HCCI experiments
with blends of paraffinic and aromatic fuels over a range of engine speeds and dilution
levels in a four-stroke CFR test engine with a variable compression ratio. The intake air
was heated to a high level to achieve HCCI operation and mimic the benefit of high
internal residuals present in two-stroke engines. Ignition and smooth energy release were
obtained by varying the engine operating parameters, such as equivalence ratio, inlet
temperature, and EGR rate. They used global autoignition chemistry and kinetics to
analyze the experimental results. It was concluded that HCCI ignition is controlled by
low temperature (below 950 K) hydrocarbon oxidation (and they recommended the use of
18
a skeletal reaction model proposed by Shell-Thornton Research Labs), and that the
energy release process is controlled by the high temperature (above 1000 K) hydrocarbon
oxidation kinetics as characterized by Dryer and Glassman [1978].
An empirical
equation was also developed based on Dryer and Glassman’s global kinetics, and
successfully predicted the average rate of energy release.
An early effort to determine the permissible operating parameters of a four-stroke
HCCI engine was conducted at Southwest Research Institute by Thring [1989] using
gasoline as fuel. Using a Labeco CLR engine, Thring mapped the HCCI operating range
by varying equivalence ratio, EGR rate, engine speed, and inlet temperature. In this work,
HCCI combustion could only achieve stable operations at conditions of low speed and
low load in a four-stroke engine, and the overall HCCI operating range was very narrow.
Diesel engine like fuel economy was achieved under selected conditions (ISFC in the
range of 180 to 200 g/kWh). High EGR rates (in the range of 13 to 33 percent) and high
intake temperatures were necessary for HCCI operation.
It is widely accepted that HCCI combustion is dominated by chemical kinetic
reaction rates [Najt and Foster, 1983], with no requirement for flame propagation. This
notion has been supported by numerous studies, which indicated that the order of radical
formation in HCCI combustion corresponds to that of self-ignition instead of flame
propagation [Noguchi et al., 1979; Oguma et al. 1997]. Experimental [e.g., Shimazaki et
al., 1999] and modeling [e.g., Aceves et al., 1999, 2000, 2001a and b] efforts have also
supported this idea.
Recent chemical kinetics modeling of HCCI combustion has concluded that HCCI
ignition is controlled by hydrogen peroxide (H2O2) decomposition. Hydrogen peroxide is
19
formed as a result of low temperature chemical reactions in the engine charge and at a
high enough temperature it decomposes into two OH radicals, which are very efficient at
attacking the fuel and releasing energy. Hydrogen peroxide decomposition occurs over a
temperature range of 1050-1100 K at the elevated in-cylinder pressures after compression.
This fundamental chemistry of HCCI autoignition and combustion is identical to the
chemistry of knock in spark-ignition engines. With high-octane fuels, little heat is
released prior to this main ignition event at 1050-1100 K; however, with low-octane fuels
(e.g., diesel fuel) significant heat-producing reactions begin at temperatures of about 800
K [Kelly-Zion and Dec, 2000]. Although the amount of energy liberated is too small to
be considered ignition, these low-temperature reactions quickly drive the mixture up to
the 1050-1100 K temperature necessary for H2O2 decomposition and main ignition. It is
this effect that requires HCCI operating parameters to be adjusted with changes in fuel
type [Kelly-Zion and Dec, 2000]. Active radicals (i.e., reactive chemical compounds,
such as H, OH; HO2) present in the exhaust gases do not survive the exhaust and intake
strokes and play a very minor role in starting HCCI combustion; however, partial
oxidation products formed during fuel decomposition can be carried over and, under
proper conditions, sensitize the incoming charge and initiate early pre-ignition reactions.
While the HCCI process has been studied intensively over the past several years,
the chemical mechanisms that control the combustion process are still far from being
completely understood. This statement is based on the observation that no universal and
reliable model has been developed for HCCI prediction, in spite of the tremendous efforts
that have been made towards these objectives. This is not to say that progress has not
been made. Multi-zone detailed chemical kinetic models coupled with CFD codes [Kraft
20
et al., 2000; Aceves et al., 2000] have shown progressively better ability to predict the
heat release rate and the onset of HCCI ignition in engines. While these studies are very
encouraging, they are limited with respect to experimental conditions and fuels, because
the chemical mechanism used in these models were developed under conditions not
directly applicable to HCCI conditions.
The development of chemical mechanism
information for hydrocarbon oxidation under the highly dilute, large percentage of
Exhaust Gas Recirculation (EGR), and pre-heated inlet charge conditions expected in
HCCI engines should improve these chemical kinetic models.
2.2.2
Advantages of HCCI
Relative to SI gasoline engines, HCCI engines are more efficient, approaching the
efficiency of a CI engine due to the following three factors: (1) the elimination of the
throttling losses, (2) the use of high compression ratios (similar to a CI engine), and (3) a
shorter combustion duration (since it is not necessary for a flame to propagate across the
cylinder). HCCI engines also have lower engine-out NOx than SI engines. Although
three-way catalysts are adequate for removing NOx from current-technology SI engine
exhaust, low NOx is an important advantage relative to direct-injection, spark-ignition
(DISI) technology, which is being considered for future SI engines.
Relative to CI engines, HCCI engines have substantially lower emissions of PM
and NOx. Emissions of PM and NOx are the major challenges for CI engines to meet
future emissions standards, and hence controlling these emissions is the focus of
extensive current research. The low emissions of PM and NOx in HCCI engines are a
result of the dilute homogeneous air and fuel mixture in addition to low combustion
21
temperatures. The charge in an HCCI engine may be made dilute by being very lean, by
using Exhaust Gas Recirculation (EGR), or by some combination of the two. Because
flame propagation is not required, dilution level can be much higher than the levels
tolerated by either SI of CI engines. Combustion is induced throughout the charge
volume by compression heating due to the piston motion, and it will occur in almost any
fuel/air/exhaust-gas mixture once the 800 to 1100 K ignition temperature (depending on
the type of fuel) is reached. As combustion occurs, the temperature will rise above the
ignition temperature, but complete combustion can be achieved at temperatures below
those at which significant NOx is produced. In contrast, in CI engines, minimum flame
temperatures are 1900 to 2100 K, high enough to make unacceptable levels of NOx
[Flynn et al., 2000]. Additionally, the combustion duration in HCCI engines is much
shorter than in CI engines since it is not limited by the rate of fuel/air mixing. This
shorter combustion duration gives the HCCI engine an efficiency advantage.
Another advantage of HCCI combustion is its fuel-flexibility. HCCI operation
has been demonstrated for a wide range of fuels [Oguma et al., 1997; Christensen et al.,
1997; Gray and Ryan, 1997]. HCCI engines can operate on gasoline, diesel fuel, and
most alternative fuels, such as methanol, ethanol, LPG and natural gas etc. However,
gasoline is particularly well suited for HCCI operation. High efficiency CI engines, on
the other hand, cannot run on gasoline due to its low cetane number.
HCCI is potentially applicable to virtually every size-class of transportation
engines from small motorcycles to large ship engines which certainly encompasses
automobiles and trucks. In fact, the smallest commercially available engines, those for
model airplanes, are actually HCCI engines [Heywood and Sher, 1999]. HCCI is also
22
applicable to reciprocating engines used outside the transportation sector such as those
used for electrical power generation and pipeline pumping.
If we assume that vehicles with HCCI engines would be 25 percent more efficient
than their non-HCCI counterparts, large reductions in the demand for petroleum are
possible [Epping et al., 2002]. (The 25 percent difference seems reasonable given that
current diesel versions of vehicles use 40 percent less fuel than their gasoline
counterparts).
Even if HCCI engines were to achieve only a 25 percent market
penetration, the savings in oil consumption would be significant. Additional savings may
accrue from reduced refining requirements for fuels for HCCI engines relative to gasoline
for conventional SI technology.
HCCI is a potential low emission alternative to CI engines in light-, medium- and
heavy-duty applications. Even with the advent of effective exhaust emission control
devices, CI engines are currently seriously challenged to meet the future emission
standards.
Although the actual cost and fuel-consumption penalties of CI emission
controls are uncertain, the use of HCCI engines or engines operating in HCCI mode for a
significant portion of the driving cycle could significantly reduce the overall cost of
operation, thus saving fuel and reducing the economic burden of lowering emissions.
SI engines for automotive applications also require intensive design efforts to
improve overall vehicle fuel efficiency. It appears that SI engines will require advanced
NOx emission control devices similar to those being developed for CI engines.
While HCCI engines have several inherent benefits as replacements for SI and CI
engines in vehicles with conventional powertrains, they are particularly well suited for
use in internal combustion engine/electric series hybrid vehicles.
In these hybrids,
23
engines can be optimized for operation over a fairly limited range of speeds and loads,
thus eliminating many of the control issues normally associated with HCCI, creating a
highly fuel-efficient vehicle. In addition to the on-highway applications discussed above,
it should be noted that the benefits of HCCI engines could be realized in most other
internal combustion engine applications such as off-road vehicles, marine applications,
and stationary power applications. The resulting benefits would be similar to those
discussed previously.
2.2.3
The Importance of HCCI Research
Although stable HCCI operation and its substantial benefits have been
demonstrated at selected steady-state conditions, several technical barriers must be
overcome before HCCI engines can be widely used. The main disadvantages of HCCI
and efforts to overcome these technical barriers are briefly listed below:
•
Hard to control ignition timing and combustion rate
HCCI ignition is determined by the charge mixture composition and its
temperature history (and to a lesser extent, its pressure history). Changing the power
output of an HCCI engine requires a change in the fueling rate and, hence, the charge
mixture.
As a result, the temperature history must be adjusted to maintain proper
combustion timing. Similarly, changing the engine speed changes the amount of time for
the autoignition chemistry to occur relative to the piston motion. Again, the temperature
24
history of the mixture must be adjusted to compensate. These control issues become
particularly challenging during rapid transients.
Several potential control methods have been proposed to adjust operational
parameters for changes in speed and load. Some of the most promising include varying
the amount of hot exhaust gas recirculation (EGR) left in the cylinder after combustion,
using a fuel additive to enhance ignition [Flowers et al., 2000; Olsson et al., 2001; Flynn
et al., 1999], using a Variable Compression Ratio (VCR) mechanism to alter TDC
temperatures [Christensen et al., 1997, 1999; Flynn et al., 1999; Sharke, 2000], and using
Variable Valve Timing (VVT) to change the effective compression ratio and/or the
amount of hot residual retained in the cylinder [Theobald and Henry, 1994; Kaahaaina et
al., 2001; ]. VCR, VVT, and fuel additives are particularly attractive because their time
response can be made sufficiently fast to handle rapid transients.
Although these
techniques have shown strong potential, they are not yet fully proven, and cost and
reliability issues must be addressed.
The possibility also exists to control HCCI
combustion by controlling the temperature, pressure, and composition of the mixture at
the beginning of the compression stroke. In this methodology, thermal energy from
exhaust gas recirculation (EGR) or compression of the inlet charge is used to vary charge
inlet (and subsequent in-cylinder) conditions [Martinez-Frias et al., 2000]. The main
advantage of this method is its simplicity. The disadvantage of this method is that it may
be too slow to react to the rapidly changing conditions that typically exist in
transportation applications. A full transient response analysis of this type of system has
yet to be performed and would depend on the specific system used.
25
•
High CO and UHC emissions, particularly at lower load conditions
HCCI engines have inherently low emissions of NOx and PM, but relatively high
emissions of hydrocarbons (HC) and carbon monoxide (CO). Some potential exists to
mitigate these emissions at light load by using direct in-cylinder fuel injection to achieve
appropriate partial-charge stratification. However, in most cases, controlling HC and CO
emissions from HCCI engines will require exhaust emission control devices. Catalyst
technology for HC and CO removal is well understood and has been standard equipment
on automobiles for many years. However, the cooler exhaust temperatures of HCCI
engines may increase catalyst light-off time and decrease average effectiveness. As a
result, meeting future emission standards for HC and CO will likely require further
development of oxidation catalysts for low-temperature exhaust streams. However, HC
and CO emission control devices are simpler, more durable, and less dependent on scarce,
expensive precious metals than are NOx and PM emission control devices.
Thus,
simultaneous chemical oxidation of HC and CO (in an HCCI engine) is much easier than
simultaneous chemical reduction of NOx and oxidation of PM (in a CI engine).
•
Relatively narrow operating range
Although HCCI engines have been demonstrated to operate well at low-to-
medium loads, difficulties have been encountered at high-loads. Combustion can become
very rapid and intense, causing unacceptable noise, potential engine damage, and
eventually unacceptable levels of NOx emissions. Expanding the controlled operation of
an HCCI engine over a wide range of speeds and loads is a big challenge for HCCI.
HCCI starts having NOx problems as load increases (φ = 0.5 to 0.6), and will likely
26
require transitioning to conventional operation at high load. Thus, the biggest problem
for HCCI may be the control of the transitions into and out of HCCI.
Preliminary research indicates the operating range of HCCI can be extended
significantly by producing a broad temperature distribution inside the cylinder and/or by
partially stratifying the charge mixture (i.e., SCCI combustion) at high loads to stretch
out the heat-release event. Several potential mechanisms exist for achieving this partial
charge stratification, including varying in-cylinder fuel injection, injecting water, varying
the intake and in-cylinder mixing processes to obtain non-uniform fuel/air/residual
mixtures, and altering cylinder flows to vary heat transfer. The extent to which these
techniques can extend the operating range is currently unknown, and R&D will be
required. Because of the difficulty of high-load operation, most initial concepts involve
switching to traditional SI or CI combustion for operating conditions where HCCI
operation is more difficult. This dual mode operation provides the benefits of HCCI over
a significant portion of the driving cycle but adds to the complexity by switching the
engine between operating modes.
•
Difficulty with cold start and light load
At cold start, the compressed-gas temperature in an HCCI engine will be reduced
because the charge receives no preheating from the intake manifold and the compressed
charge is rapidly cooled by heat transfer to the cold combustion chamber walls. Without
some compensating mechanism, the low compressed-charge temperatures could prevent
an HCCI engine from firing. Various mechanisms for cold-starting in HCCI mode have
been proposed, such as using glow plugs, using a different fuel or fuel additive, and
27
increasing the compression ratio using VCR or VVT.
Perhaps the most practical
approach would be to start the engine in spark-ignition mode and transition to HCCI
mode after warm-up. For engines equipped with VVT, it may be possible to make this
warm-up period as short as a few fired cycles, since high levels of hot residual gases
could be retained from previous spark-ignited cycles to induce HCCI combustion.
Although solutions appear feasible, significant R&D will be required to advance these
concepts and prepare them for production engines.
2.2.4
Results Using Different Fuels
One of the advantages of HCCI combustion is its intrinsic fuel flexibility. The
literature shows that HCCI can be achieved with a range of hydrocarbons [Oguma et al.,
1997; Christensen et al., 1997, Gray and Ryan, 1997], including gasoline, diesel fuel,
propane, natural gas, and neat or binary mixtures of the SI engine primary reference fuels
(PRF), iso-octane and n-heptane.
HCCI combustion has little sensitivity to fuel
characteristics such as lubricity and laminar flame speed. Fuels with any octane or cetane
number can be burned, although the operating conditions must be adjusted to
accommodate different fuels, which can impact efficiency, as discussed below. An HCCI
engine, in principle, can operate on any hydrocarbon or alcohol liquid fuel, as long as the
fuel is vaporized and mixed with the air before ignition.
The applicability of typical fuels to HCCI engines is discussed below. Other fuels
(methanol, ethanol, and acetone) have also been tried in experiments, but with
inconclusive results.
28
Gasoline:
Gasoline has several advantages as an HCCI fuel, one being a high Octane
Number (ON). ON is used to indicate the resistance of a motor fuel to knock, and it is
based on a scale in which isooctane is 100 ON and n-heptane is 0 ON. ON’s are typically
in the range of 87 to 92 in the U.S. and up to 98 in Europe, which allows the use of
reasonably high compression ratios in HCCI engines. Actual compression ratios for
gasoline-fueled HCCI engine data vary from 12:1 to 21:1 depending on the fuel octane
number, intake air temperature, and the specific engine design (which may affect the
amount of hot residual naturally retained). This compression-ratio range allows gasolinefueled HCCI engines to achieve relatively high thermal efficiencies (in the range of
diesel-fueled CI engine efficiencies). A potential drawback to higher compression ratios
is that the engine design must accommodate the relatively high cylinder pressures that
can result, particularly with high engine loads. Additional advantages of gasoline include
easy evaporation, simple mixture preparation, and a ubiquitous refueling infrastructure.
Gasoline is a complex mixture of hundreds of hydrocarbons. While the majority
of research engine tests utilize full-boiling range fuels, often it is desirable to limit the
chemical and/or physical complexity of the fuel to generate insight and understanding
into the underlying fundamental processes.
This parallels the problem with
computational chemistry models of the combustion processes.
The need exists for
models of the chemistry of real fuels; unfortunately, it is currently not possible to
represent the chemistry of all these complex hydrocarbon mixtures with detailed
chemical kinetic models. Consequently, it is advisable to develop computational models
29
for simpler mixtures (validated against experimental data) before moving to the
complexity of real fuels.
The simplest surrogate fuels for gasoline consist of single components, e.g., the
use of iso-octane as a gasoline surrogate. Binary blends of n-heptane and iso-octane, the
octane rating scale primary reference fuels (PRFs), also find wide-spread use as
convenient surrogates for variable RON/MON fuels. Mixtures of these two PRFs are
used to define the octane number (ON) scale, specifically by the volumetric percentage of
iso-octane in the mixture.
Therefore, the present work concentrated on the SI PRFs and their mixtures. nHeptane, which is also used as a representative diesel fuel component, and iso-octane
have quite different oxidation chemistries. Studies show that n-heptane autoignition
occurs in two stages, while iso-octane autoignition happens in a single stage at higher
temperature [Epping et al., 2002]. Further experiments show that HCCI combustion of
PRFs and PRF blends in engines is usually characterized by a two-stage heat release
process due to the separate contributions of low temperature reactions (LTR) and high
temperature reactions (HTR) [Rao et al., 2004].
Research also shows that HCCI
operation with pure n-heptane requires a compression ratio of about 11:1 to phase
autoignition at TDC without inlet air preheating, while iso-octane and high octane
gasoline (RON 98) require compression ratios of 21.5:1 and 22.5:1, respectively
[Christensen et al., 1999].
30
Diesel Fuel:
The HCCI combustion of diesel type fuels can be more easily achieved than with
gasoline type fuels because of diesel fuels’ lower autoignition temperature. However,
overly advanced combustion phasing can cause low thermal efficiency. In addition,
mixture preparation is a critical issue. There is a problem getting diesel fuel to vaporize
and premix with the air due to its low volatility [Christensen et al., 1999; Peng et al.,
2003]. Therefore, to obtain premixed HCCI combustion using diesel fuel, the air-fuel
mixture must be heated considerably to evaporate the fuel, and the compression ratio of
the engine must be very low (8:1 or lower) to obtain satisfactory combustion, which
results in a low engine efficiency. Alternatively, the fuel can be injected in the intake
port or in-cylinder but, without air preheating, temperatures are not sufficiently high for
diesel-fuel vaporization until well into the compression stroke. This strategy often results
in incomplete fuel vaporization and poor mixture preparation, which can lead to PM and
NOx emissions. However, one concept for direct injection of diesel fuel, involving late
injection (after TDC) with high swirl, has been successful at thoroughly vaporizing and
mixing the fuel before ignition at light to moderate loads. Using this method, diesel-like
compression ratios of 15:1 to 16:1 can be used resulting in high efficiency. This mode of
operation is used in the Nissan MK engine [Kimura et al., 1999 and 2001]. Like gasoline,
diesel fuel has an extensive refueling infrastructure.
The HCCI operation using diesel fuel was extensively tested at Southwest
Research Institute (SwRI) by Ryan and Callahan [1996] and Gray and Ryan. [1997]. For
the first time, Knock Intensity (KI) was used to trace knock and determine the acceptable
HCCI operating range. According to the definition, knock that is just marginally audible
31
is used to define a KI of 5 on a scale from zero to ten. The rate of pressure rise is
measured and used to yield a KI. A KI of 4 was used to identify permissible HCCI
operations. The HCCI operating range was tested by varying EGR rate, compression
ratio, and inlet temperature. They found that management of EGR rate and equivalence
ratio was critical to achieving HCCI. Under 50 percent EGR rate and stoichiometric
fresh charge condition, the engine would produce acceptable power output with near total
elimination of smoke.
A simple empirical model was also proposed by SwRI to predict HCCI ignition
delay time:
td = 0.021*(O2)-0.53*(Fuel)0.05*(ρ)0.13*exp(5914/T)
Where td is the ignition delay time (ms), O2 is the oxygen molar density (moles/m3), Fuel
is the fuel molar density (moles/m3), ρ is the density (kg/m3), and T is the air temperature
(K). However, the compression ratio had to be lowered from 16:1 to 8:1 to achieve
HCCI operation, and the unburned hydrocarbons were very high. Also, they found that it
makes little difference whether the dilute mixture was achieved by going very lean (e.g.,
below the equivalence ratio in which a flame can propagate, φ ~ 0.6 or 0.7) or by adding
exhaust gas recirculation.
32
Propane:
High efficiencies can be achieved with propane-fueled HCCI engines because
propane has a high octane number (105). In addition, because propane is used as a gas, it
can be easily mixed with air. Some infrastructure also exists for propane and it has a high
energy density during storage, as it is a liquid at moderate pressures.
Natural Gas:
Because natural gas has an extremely high octane rating (about 110), natural gas
HCCI engines can be operated at very high compression ratios (15:1 to 21:1), resulting in
high efficiency.
However, similar to gasoline or propane, the engine design must
accommodate the relatively high cylinder pressures that can result. Natural gas is widely
available throughout the U.S.
2.3
Fuel Additives
Fuel additives can be grouped into different categories based on functions, such as
engine performance, fuel stability, and fuel handing and contaminant control. Engine
performance additives discussed here are a class of additives that can improve engine
performance usually by changing autoignition characteristics.
Historically, the study of fuel ignition-enhancing (or suppressing) additives was
motivated by the need for a cetane (or octane) number improver. Cetane Number (CN) is
a rating scale used for diesel engines to indicate the tendency of a fuel to autoignite. The
33
rating compares a fuel’s performance in a standard engine with that of a mixture of
cetane (CN = 100) and alpha-methyl-napthalene (CN = 0).
While the major source of
diesel fuel has been straight-run distillates separated from crude oil; however, the
increase in market demand for diesel fuel has led to the position where oil refiners are
incorporating more cracked distillates into diesel fuels.
Diesel fuels derived from
cracked distillates generally have a relatively low cetane rating (i.e., poor ignition
quality), as the cracking processes result in higher proportions of aromatic molecules in
the product. Such fuels normally require higher temperatures to ignite than their straightrun counterparts. In diesel combustion, this results in extended ignition delay periods and
faster initial burn and rate of pressure rise, with the consequent effects of greater noise
output and rough running.
In the 1940s and 1950s, a number of investigations into the potential of additives
for diesel fuel ignition quality improvement were carried out [Bogen and Wilson, 1944;
Robbins et al., 1951; Anderson and Wilson, 1952; Brien, 1956; Hurn and Hughes, 1956].
During the period of plentiful petroleum supplies (approximately 1950-1970), there
seemed to be little need for fuels research and the literature reflects this with relatively
few publications. Some of the studies that were performed were engine based [McGreath,
1971; Kamel, 1984] while others were of a more fundamental nature [Salooja, 1962;
Dunskus and Westwater, 1961; Satcunanathan and EI-Nesr, 1972; Kirsch et al., 1981].
However, with the oil crisis of the seventies and the then growing use of cracked distillate
fuels, there was renewed interest in developing suitable ignition promoting additives [Li
and Simmons, 1986; Pishinger et al., 1988; Inomata et al., 1990; Clothier et al., 1990].
34
The effects of additives on knock in SI engine were also studied by Downs et al.
[1951]. Alkyl peroxides, aldehydes and hydrogen peroxide were investigated. The
results demonstrated the key role of alkyl peroxides in the knock process. Formaldehyde,
acetaldehyde, propionaldehyde, and butylaldehyde were added in molar concentrations of
5% or more to a full boiling gasoline. Interestingly, the formaldehyde acted as an antiknock and the other aldehydes were only slightly pro-knock.
As noted, ideally, autoignition in HCCI should occur at the point where the piston
reaches top dead center to provide optimum power and efficiency; therefore the timing of
the autoignition is critical. Without the help of an external triggering event, HCCI has a
problem in controlling the ignition timing. One option for ignition control is to use small
amounts of ignition-enhancing additives to alter the ignition properties slightly.
Although a large number of ignition additives have been shown to be effective in
enhancing the ignition quality of the parent fuel to which they are added, their precise
role in promoting ignition remains uncertain. One school of thought is that they modify
the physical processes contributing to the delay period.
For example it has been
suggested that ignition additives may act in much the same way as water, when the latter
is introduced into diesel fuels in the form of an emulsion. Being at supercritical pressure,
additives may evaporate instantaneously; thereby shattering fuel droplets and assisting
atomization [Valdmanis and Wulfhorst, 1970]. Others claim that additives act as heatflux improvers, considerably increasing the heat transfer rate in nucleate boiling and so
reducing evaporation time [Dunskus and Westwater, 1961; Satcunanathan and EI-Nesr,
1972]. Another school of thought maintains that the main effect of additives is that of
accelerating the autoignition chemistry. Most effective additives are thermally unstable
35
and their thermal decomposition could be expected to yield free radicals. It has been
suggested that these are effective in enhancing the chain branching reactions leading to
ignition [Hurn and Hughes, 1956; Salooja, 1962]. It has also been claimed that the local
temperature rise, caused by the heat release from the thermal decomposition of the
additive may be of equal importance in stimulating the autoignition of the fuel [Inomata
et al., 1990].
Di-tertiary Butyl Peroxide (DTBP), (CH3)3COOC(CH3)3, is one such additive and
it has been suggested as a commercial cetane number improver in diesel engines. DTBP
was selected in this study because (1) it is readily available; (2) it has been used by a
number of previous workers, permitting cross reference to other tests; and (3) its structure
and decomposition mechanism is reasonably well understood.
Like most effective
additives, DTBP is thermally unstable and its thermal decomposition liberates heat and
yields free radicals.
Experiments show that DTBP can reduce ignition delay both in rapid compression
machines for SI PRFs and PRF blends [Inomata et al., 1990; Tanaka et al., 2003] and in
engines for diesel fuels [Ai-Rubaie et al., 1991]. However, it was unclear whether the
effect of DTBP is thermal or chemical. The present work is aimed at determining its the
mode of action and its effects on HCCI operation.
DTBP was adopted as the “reference” additive and the performance of other
additives and mixtures were compared to it by Ai-Rubaie et al. [1991]. Experiments
show that DTBP can reduce ignition delay both in rapid compression machines for SI
PRFs and PRF blends [Inomata et al., 1990; Tanaka et al., 2003] and in engines for diesel
fuels [Ai-Rubaie et al., 1991]. In rapid compression studies, the oxidation of 1% DTBP
36
alone in air is capable of raising the compressed gas temperature by 35ºC [Inomata et al.,
1990], and with 2% addition of DTBP to PRF90, the ignition delay time was cut in half
[Tanaka et al., 2003]. Addition of 1% by volume of DTBP to a diesel fuel with cetane
number 40 at an injection temperature of 880 K in an engine caused a 14% reduction of
the ignition delay [Ai-Rubaie et al., 1991].
The experimental investigations of DTBP described above were conducted at
relatively high temperatures and varying conditions in engines, shock tubes and rapid
compression machines. In spite of this work, it remains unclear whether the effect of
DTBP on the lower temperature autoignition processes was thermal or chemical. Thus,
part of the present work was aimed at elucidating DTBP’s mode of action. The work was
carried out primarily in a Pressurized Flow Reactor (PFR) which can effectively simulate
the conditions occurring during the critical preignition regime of hydrocarbon oxidation,
and involved measurement of CO concentration, serving as a reactivity marker, at 8 atm
and 650 K < T < 900 K. Measurements were made for seven PRF fuel blends, with and
without DTBP addition. Additional tests were carried out in a CFR engine. The effects
of DTBP on primary reference fuels in the PRF and in an engine are reported in Chapters
4 and 5, respectively.
2.4
Models of Hydrocarbon Oxidation Mechanisms
Although hydrocarbon combustion properties have been studied for over a
century, numerical combustion modeling work did not become an essential part of
combustion research and development programs until the 1980s [Dryer, 1991] with the
37
development and advance of computers.
There have been two distinctly different
approaches to the numerical modeling of hydrocarbon oxidation. One involves explicitly
accounting for all possible chemistry detail and the other elects to account for a minimal
set of features.
Generally mechanisms are classified as detailed, lumped, reduced,
skeletal, or global [Zheng et al., 2004].
Detailed models try to include all of the
important elementary reactions and individual species using the best available rate
parameters and thermochemical data. The other four model types are all driven by the
desire to minimize the model size. Their general characteristics are shown in Table 2-2.
Table 2-2. Categories of chemical kinetic models
(Zheng et al., 2004)
Category
Description
Species
Reactions
Detailed
the latest “comprehensive” reaction set
100’s
1000’s
Lumped
uses a lumped description for larger species
100’s
1000’s
Reduced
a subset of the detailed model
10’s
10’s- 100’s
Skeletal
employs class chemistry and lumping concepts
10’s
10’s
Global
utilizes global reactions to minimize reaction set
<10
<10
Regardless of the type of mechanism, each reaction requires the associated
reaction rate coefficients and species thermodynamic properties. Accurate estimation of
heats of formation for all radicals and stable species are needed to identify possible
pathways and to estimate activation energies and the rates of reversible reactions. As the
fuels of interest increase in size and complexity, the estimation becomes more difficult.
A detailed description for each category is given below.
38
Detailed Mechanism
A detailed mechanism, as its name implies, includes almost all of the important
elementary reactions and individual species with available rate parameters and
thermochemical data.
As the level of understanding and the size of the molecules
increase, detailed mechanisms become extremely large. A detailed C7 hydrocarbon
mechanism may contain thousands of reactions and hundreds of species [Curran et al.,
1998, 2002]. Coupled with CFD and if used in an engine simulation, such a model would
require tremendous computational resources.
In 1984, Westbrook and Pitz [1984] introduced a comprehensive chemical kinetic
mechanism for the oxidation and pyrolysis of propane and propene. This model was later
extended to lower temperatures [Smith et al., 1985] and to much more complex fuels
[Westbrook and Pitz, 1987]. In a later work by Westbrook et al. [1991], they included
low temperature reaction paths involving alkylperoxy radical isomerization in the
program and examined the chemical kinetic process leading to knocking in spark-ignition
internal combustion engines. Since then there have been efforts to develop detailed
models for butane [Green et al., 1987a, b], pentane [Wang et al., 1999] and heavier
hydrocarbons such as n-heptane and iso–octane [Curran et al., 1998 and 2002].
Lumped Mechanism
Lumped mechanisms typically classify the primary propagation reactions of the
parent fuel with a limited set of reference kinetic parameters and group the primary
intermediate isomers into a limited number of “lumped” components [e.g., Violi et al.,
39
2002; Agosta et al., 2004]. The smaller species are typically treated in a detailed manner
nearly identical to the detailed mechanism.
The size of a lumped mechanism can vary significantly, but for large hydrocarbon
applications they can still encompass thousands of reactions among hundreds of species
[e.g., Granata et al., 2003; Nehse and Warnatz, 1996].
Reduced Mechanism
A reduced mechanism begins with either a detailed or lumped mechanism. Then
the critical reactions and species are selected by one of several “reduction” methods. The
more useful techniques for automatic reduction are: Quasi-Steady-State Approximation
(QSSA) [Peters and Rogg, 1993], Intrinsic Low-Dimensional Manifolds (ILDM) [Maas
and Pope, 1992], Computational Singular Perturbations (CSP) [Lam and Goussis, 1988],
Directed Relation Graphs [Lu and Law, 2004], and others.
Reduced mechanisms
nominally have tens to hundreds of reactions among tens of species.
Skeletal Mechanism
Skeletal mechanisms are developed from the opposite perspective from the
mechanism categories just described. Instead of starting with an all inclusive detailed
model, they are assembled with just enough of the chemical skeleton to model the
parameters of interest. Normally these mechanisms consist of tens of reactions and tens
of species [e.g., Zheng et al., 2001; 2002a, b].
In skeletal mechanisms, the rate
parameters and thermochemistry are based on “classes” of reactions.
40
A review of such kinetic models and their applications has been made by Griffiths
[1995]. The earliest skeletal kinetic model, based on degenerate-branched-chain and
class chemistry concepts, was developed at the Shell Thornton Research Center by
Halstead et al. [1975, 1977]. This model consisted of 8 generalized reactions and 5
species with the primary interest being to match the ignition delay behavior, while the
phenomenological complexity of hydrocarbon oxidation, such as cool flames, two stage
ignition and NTC behavior were considered to be of secondary importance. This work
formed the basis for later developments and the model was widely used in engine
applications.
Cox and Cole [1985] developed a skeletal chemical kinetic model consisting of 15
reactions and 10 active species. The model was tested against the ignition data using isooctane and PRF90 in a rapid compression machine.
Hu and Keck [1987] further
developed a skeletal chemical kinetic model of 18 reactions and 13 active species.
Keeping a better representation of the chemical reactions similar to Cox and Cole model,
Hu and Keck model treated exothermicity as enthalpy change in each step reaction. The
rate parameters were calibrated using measured explosion limits in a combustion bomb.
The fuels studied were C4-C8 straight chain paraffins and iso-octane. The effects of fuel
structure are reflected in the rate parameter of the RO2 isomerization reaction.
The
model was applied to predict selected data of ignition delay measured in a rapid
compression machine. Cowart et al. [1990] reproduced the overall trend for ignition
delay of specific hydrocarbons of interest with this model. However, significant physical
features such as preignition fuel consumption, cumulative heat release, and species
concentrations are not included in the model [Li et al., 1996]. These deficiencies were
41
addressed by Li et al. [1992, 1995, 1996] and the basic model was further refined by
Zheng et al., [e.g., 2001, 2002a and b] and in the present work.
Global Mechanism
A global mechanism describes the chemistry in terms of a few principal reactants
and products in a small number of functional relations. Typically, global mechanisms
have fewer than ten reactions among fewer than ten species. These types of mechanisms
are extremely attractive for CFD and other heavy computational applications where
“large” mechanisms are computationally expensive.
Global models were first developed to describe high temperature chemistry
[Dryer, 1991]. Later, a 4-reaction model [Müller et al., 1992] and a 5-reaction model
[Schreiber et al. 1994] were developed to describe the full temperature regime. However,
neither of these global models can reflect hydrocarbon oxidation behavior in the Negative
Temperature Coefficient (NTC) regime, since NTC behavior inherently involves
intermediate species (for example, HOOH) that provide branching at 900-1100 K. Hence
these two global models were only used to predict ignition delays, and they are not
suitable for prediction of the full HCCI behavior that occurs with PRF fuels. Bourdon et
al. [2004] proposed an optimized 5-step model for HCCI applications. Zheng et al. [2004]
at Drexel University developed a 7 step model to successfully predict temperature,
pressure, ignition delay, combustion duration, and heat release for PRF20 in an engine
operating in HCCI mode. The model includes five reactions that represent degenerate
chain branching in the low temperature region, including chain propagation, termination
and branching reactions and the reaction of HOOH at the second stage ignition. Two
42
reactions govern the high temperature oxidation, to allow formation and prediction of CO,
CO2, and H2O.
2.5
Low and Intermediate Temperature Regime Fuel Oxidation
Generally, the combustion environment, such as temperature, pressure, and
equivalence ratio effects the location of the boundaries between each regime. At one
atmosphere, the hydrocarbon oxidation process can be divided along the following
approximate boundaries:
(1) low temperature, < 650 K
(2) intermediate temperature, 650-1000 K
(3) high temperature, > 1000 K
In engines, fuel spends a relatively long time in the low and intermediate
temperature regime (<1000 K), where it decomposes significantly and generates many
intermediate species prior to autoignition [Smith et al., 1985; Cernansky et al., 1986;
Green et al., 1987 a and b; Leppard, 1987, 1988; Henig et al., 1989]. As illustrated in the
in-cylinder end gas temperature - pressure trajectories presented in Figure 2-2, engine
autoignition, which is associated with knocking, cold start and misfire, is a low and
intermediate temperature phenomenon.
43
1400
CI Engine Inlet
CI Engine Cold Start Autoignition
SI Engine Inlet
SI Engine Autoignition
Adiabatic Engine Trajectories
Temperature [K]
1200
1000
H+
O2
e
nov
T ur
r
High
800
600
400
Intermediate R + O2 2
Turno
ver
Low
200
10-1
101
100
Pressure [atm]
102
Figure 2–2. Typical SI engine envelope of end gas temperature and pressure
histories leading up to the point of autoignition (Wang, 1999)
Motored engine experiments are good for studying autoignition, because they can
provide the low and intermediate temperature and higher pressure environment
conditions, in which autoignition takes place. Green et al. [1987a and b] studied the
chemical aspects of autoignition of iso-butane and n-butane using a “skip-fired”
technique in a single cylinder research engine. They concluded that low temperature
chemistry plays an important role in end-gas autoignition. Leppard [1988] also studied
the oxidation of iso-butane using a motored engine and developed a reaction mechanism.
Later, Leppard [1989] studied more fuels using the same technique and reported that
44
olefins do not exhibit negative temperature coefficient behavior. (Note: a study at Drexel
University found that large olefins do exhhibit NTC behavior [Prahbu et al., 1996]).
At Drexel University, initial engine experiments were conducted by Henig et al.
[1989] using n-butane, iso-butane and blends using the same skip-fired strategy as Green
et al. [1987 a and b] to investigate the effects of fuel structure on autoignition. Products
sampled from the end gas in fired cycles confirmed the importance of low and
intermediate temperature chemistry prior to autoignition and examined the interaction
between n- and iso-butane. The heat release and chemical species in the second motored
cycles were examined in a later investigation by Addagarla et al. [1989a]. Chemical
pathways were discussed based on the species data. Wilk et al. [1990] modeled the
species data using a detailed chemical kinetic model.
Addagarla et al. [1989b] measured the critical inlet fuel/air conditions of
temperature and pressure which induce autoignition for n-pentane, n-hexane, and the
primary reference fuels under motored engine conditions.
Then, based on gas
composition measurements in the engine prior to ignition, Addagarla et al. [1991] studied
the n-pentane mechanism. Filipe et al. [1992a and b] examined the preignition reactivity
and autoignition behavior of several PRF blends under motored conditions.
Time
resolved concentration profiles of fuels and light intermediate species (C≤4) were
measured. The experimental results indicated that significant amount (up to 40-50%) of
both n-heptane and iso-octane reacted during the cycle.
Li et al. [1994, 1995] conducted experiments in a motored research engine fueled
with neat PRF’s, an 87 octane blend of PRF’s (PRF87), and PRF87 blended with methyl
tert-butyl ether (MTBE), ethyl tert-butyl ether (ETBE), methyl tert-amyl ether (TAME),
45
diisopropyl ether (DIPE), methanol and ethanol. Detailed evolution profiles of reactants,
molecular intermediates, and products were measured prior to autoignition via in-cylinder
sampling combined with gas chromatographic analysis. The results showed that all of the
ethers and alcohols were effective in reducing preignition reactivity and retarding
autoignition, and mechanistic explanations for the behavior were proposed.
Yang [2002] measured species evolution profiles of PRF20 at different
equivalence ratio, additives, such as 1-pentene and toluene, and major EGR components,
such as CO2 and NO. The results were used to elucidate the chemical kinetics controlling
HCCI operation. As noted previously, Zheng et al. [2001, 2002, 2004] has conducted
experimental and computational studies on skeletal mechanisms for HCCI operation.
Generally, in motored engine experiments, chemical species are typically sampled
and analyzed at selected crank angles. When coupling with air flow and fuel mixing, it is
difficult to get detailed data on fuel oxidation under desired temperature and pressure
conditions. Therefore, it is difficult to deduce and develop reaction mechanisms solely
based on experimental data from motored engines.
To obtain detailed speciation and species evolution data, flow reactors are often
used, especially for those conditions where reactions are very fast. The advantages of
flow reactors can be summarized as follows: (1) the reaction temperature and pressure
can be well controlled; (2) the reaction time can be controlled over a broad range (tens of
milliseconds to a few seconds); and (3) gas samples at different locations along the
reactor, which correspond to different reaction times, are easy to withdraw for analysis.
Therefore, more detailed reaction information can be obtained from flow reactors.
46
In the past 20 years, flow reactors were utilized extensively by many researchers
and their results have significantly contributed to our understanding of combustion
chemistry. Dryer and Glassman [1973] applied the flow reactor and studied the CO and
CH4 oxidation at high temperature. Cohen [1977] studied the mechanism of ethane
oxidation at high temperature. Later, Hautman et al. [1981] used a range of flow reactor
data and proposed a multiple-step overall kinetic mechanism for the oxidation of
hydrocarbons. Callahan et al. [1996] performed experiments to study the oxidation of
primary reference fuels over an initial reactor temperature range of 550 - 850 K and with
a constant pressure of 12.5 atm in the Princeton variable pressure flow reactor. Other
experimental efforts to provide experimental data using flow reactors include work done
by Vermeersh et al. [1991] and Bales-Gueret et al. [1992].
At Drexel University, a pressurized flow reactor has been utilized extensively in
investigating hydrocarbon oxidation chemistry. Koert [1990] designed the flow reactor
system and employed it to examine the effect of pressure on the oxidation of propane,
and then he collaborated on a modeling effort to develop a pressure-dependent kinetic
mechanism for propane based on this experimental data [Koert et al., 1996]. In the same
facility, Wood [1994] studied the oxidation of n-pentane and 1-pentene in the low and
intermediate temperature region.
McCormick [1994] studied the C4 hydrocarbon
oxidation and developed an FTIR technique to analyze the samples taken from the reactor.
Later, Prabhu et al. [1996] investigated 1-pentene oxidation and its interaction with nitric
oxide. Wang et al. [1999] employed the pressurized flow reactor to obtain species
information of neopentane to develop a detailed model.
47
The oxidation and ignition characteristics of pure alkanes (n-dodecane and
isocetane),
naphthenes
(methylcyclohexane
and
decalin),
and
aromatics
(α-methylnaphthalene and hexylbenzene) and of their mixtures have been experimentally
studied by Agosta [2002]. The analysis of the interactions controlling the ignition of
binary, ternary and larger mixtures of the compounds listed above has been applied to the
synthesis of a multi-component surrogate for the military aviation fuel JP-8, which is
very similar to the commercial aviation fuel Jet-A. The surrogate has been tailored to
closely match the hydrocarbon distribution in JP-8: a mixture containing 26% n-dodecane,
36% isocetane, 18% α-methylnaphthalene, 14% methylcyclohexane, and 6% decalin,
was shown to accurately reproduce the chemical behavior of JP-8 over different
experimental conditions.
Oxidation of samples of JP-8 and Jet-A were experimentally studied by Lenhert
[2004b]. In his study, a 4-component surrogate of JP-8, with mixture of 43% n-dodecane,
27% iso-cetane, 15% methyl-cyclohexane, and 15% α-methyl-naphthalene was
developed to match the ‘average’ JP-8.
Neat, binary mixtures of the components, and
the full surrogate were oxidized in the PFR and stable intermediate and product species
were identified and quantified using permanent gas analyzers, and gas chromatography
with mass spectrometry (GC/MS).
These detailed studies provided kinetic and
mechanistic information in the low and intermediate temperature ranges (600 – 1000 K)
and at elevated pressures.
48
CHAPTER 3.
EXPERIMENTAL FACILITIES AND GENERAL TEST
METHODOLOGY
The experimental portion of this work was conducted using the pressurized flow
reactor facility and an engine facility in the Frederic O. Hess Engineering Research
Laboratories at Drexel University.
assembled by Koert (1990).
The flow reactor was originally designed and
In recent years, the facility has been enhanced by
incorporating new a gas analysis system, installing a secondary preheater and upgrading
the LabVIEW programmed computer control system. A Cooperative Fuel Research
(CFR) engine was recently set up and used to conduct the associated engine experiments.
Each of these facilities and the experimental methodology employed is described in the
following sections.
3.1
The Pressurized Flow Reactor Facility
The PFR facility is a plug flow reactor designed to investigate the effects of
pressure and temperature on the oxidation of hydrocarbon species, mainly in the low and
intermediate temperature regions at pressures up to 20 atm. The facility was designed
such that chemical processes could be examined in relative freedom from fluid mechanics
and temperature gradient effects. A detailed description of the design and fabrication of
the system can be found in Koert [1990] and Koert and Cernansky [1992]. In the past
years, several modifications to the system have been made, such as the addition of a 3
kW heater [Wang, 1999] and the addition of a liquid fuel delivery system [Wood, 1994].
More recent modifications to the PFR were completed by Agosta [2002] and Lenhert
[2004b]. These modifications included upgrades to the computer control system and
49
LabVIEW code, replacement of the fuel delivery system, and modifications to the
temperature control system, gas sampling systems, sample probe design, and other minor
systems. Independent of these modifications, the basic constructs of the reactor have
remained unchanged.
The detailed description of the facility has been provided
elsewhere by Koert [1990] and Lenhert [2004b] and only an overview of the facility will
be described here
3.1.1
Reactor Flow Systems
The PFR has a temperature range up to 1000 K, which is effective to characterize
the low and intermediate temperature regimes, and it is maintained at nearly adiabatic
conditions so that the heat transfer effects can be ignored. In the PFR, a stream of prevaporized fuel is diluted in a stream of heated nitrogen. Then, as the fuel/N2 mixture
enters the adiabatic quartz reaction duct, it is rapidly mixed with a heated oxidizer stream
consisting of nitrogen and oxygen at a concentration consistent with the level of
reactivity expected or desired. Moreover, the high flow rate inside the reactor and the
cross flow injector establish a turbulent flow regime. The residence time in the reactor is
shorter than the time it takes for radicals and other active species to diffuse radially to the
wall; therefore surface effects can be neglected.
The reactor duct is heated by means of two sets of manually controlled heaters. A
computer controlled probe is moved inside the reactor and extracts samples that are
delivered to the gas analyzers. A schematic of the PFR facility is presented in Figure 3-1.
50
Nitrogen Oxygen
Air
3kW
Heater
Exhaust
Pressure
Exhaust Regulating
Valve
Fuel
NDIR
FTIR & GC/FTIR
Syringe Pump
Pressure
Transducers
10kW
Heater
Mixing
Nozzle
Quartz
Reactor
Sample
Storage
Cart
Gas Sampling
Probe & TC
Computer Controlled
Probe Positioning Table
Probe Cooling
System
Data Control
&
Acquisition System
Figure 3–1. Schematic of the Pressurized Flow Reactor facility
The reactor duct is a 2.25 cm I.D., 40 cm long quartz tube. The temperature is
kept as constant as possible along the reactor length by insulation and by the use of
multiple bead heaters that can be manually adjusted. As noted by Khan [1998], this
multiple heating element system was developed to maintain a relatively flat temperature
profile in the PFR facility. The PFR has been divided in three main sections, inlet, test
and outlet, and the temperature of each of these sections can be monitored and adjusted
by changing the temperature set point on the corresponding bead heater. The current
temperature set points for the bead heaters during the experiments are shown in
Table 3.1.
51
Table 3-1. Bead heaters temperature set points
PFR SECTION
SWITCH
TEMPERATURE
INLET SECTION
1 Æ ON
2 Æ ON
3 Æ ON
700 ºC
700 ºC
700 ºC
TEST SECTION
4 Æ ON
5 Æ ON
6 Æ OFF
550 ºC
550 ºC
—
OUTLET SECTION
9 Æ ON
500 ºC
NITROGEN-FUEL SECTION
7 Æ ON
8 Æ OFF
300 ºC
—
The Nitrogen-Fuel section in Table 3.1 refers to the external line between the
mixing nozzle and the fuel delivery system. It is in this line that the liquid fuel delivered
by the injection pump is vaporized and mixed with the stream of heated nitrogen to
achieve the desired fuel rate and concentration for delivery to the main reactor.
Controlling the temperature along this line is of great importance and its temperature set
points must be between the fuel boiling point and a temperature where significant
decomposition of the fuel occurs. A set point of 300 ºC has been used in this study.
The fuel delivery system is one of the most critical components of the PFR
system. Accurate, repeatable, and stable delivery of liquid fuels is paramount for the
accurate quantification of intermediate oxidation species. A high pressure syringe pump
(ISCO 500D syringe pump) was used in this study. This pump has a 500 ml fuel
reservoir and is capable of delivering up to 200 ml/min at pressures up to 3,750 psi. This
52
pump provides a constant, highly repeatable flow rate with negligible pressure
oscillations. Approximate flow rates for a specified experimental condition, e.g., φ = 0.4,
N2 Dilution 75%, 8 atm, are calculated given the fuels’ density and molecular weight.
The injection system uses a 1/16th inch OD tube, 0.020 or 0.010 inch ID stainless steel
tube. The ID of the injection tube is determined by the syringe pump pressure necessary
to deliver the desired flow rate of the fuel.
3.1.2
Sampling Method and Sample Analysis
Samples of the reacting gases are obtained at different locations along the reactor
via a stainless steel probe whose position is computer controlled. Because of turbulent
flow in the reactor and because of the rapid initial mixing of the fuel and the oxidizer,
radial gradients in the PFR can be neglected and samples taken along the centerline of the
test section characterize the chemistry of the reactions.
The probe motion is
automatically controlled using the ‘Probe Automove’ LabVIEW code developed by
Koert [1990] and modified by Lenhert [2004a].
A pressure drop across the probe orifice and the probe water-cooling system
extracts the reacting gases in such a way that further reactions are rapidly quenched. The
extraction line is also heated to approximately 70 ºC in order to keep the products at a
temperature high enough to avoid condensation of important species such as
formaldehyde. The temperature control of this line is realized via heating tapes and it is
controlled manually.
Continuous samples, extracted from the PFR at constant flow rate, can be
analyzed online using either a Fourier Transform Infrared (FTIR) spectrometer for a
53
quantitative analysis of the products of combustion or a Non-Dispersive Infrared (NDIR)
instrument for CO or CO2 concentrations. The CO concentration is used to map the
overall reactivity of the fuel. Samples can also be stored in a constant temperature
storage unit capable of holding up to 15 gas samples for later GC/MS analysis. The
overall length of the extraction line is approximately 4 m. The volumetric flow rate
inside this line is kept constant at 3 l/min and it is regulated by the rotameter of the
NDIR.
Due to the length of the extraction line, there is a time delay between the
measurement of CO and CO2 for a specific sample in the NDIR and the measurement of
that specific sample temperature by the thermocouple mounted on the tip of the probe.
Applying the equations of ideal fluids without losses, it is possible to calculate the
residence time of the gases extracted from the PFR and flowing into the NDIR. This time
has been estimated to be less than 5 s.
The reactor cooling rate during a typical
controlled cool down experiment is on the order of 3 °C/min at the start and end of
reaction, while it decreases to approximately 2 °C/min in the temperature region close to
the start of NTC – due to the larger amount of heat released. Therefore in the worst case
the temperature change during the 5 s time delay is about 0.05 °C, and this is small
enough that the CO/CO2 and the temperature measurements can be considered to be
simultaneous.
3.1.3
PFR Experimental Methodology
Two types of experiments can be performed in the Pressurized Flow Reactor,
depending on the type of chemical information that needs to be collected. The first
procedure, referred to as Constant Inlet Temperature (CIT) methodology, is particularly
54
suitable for collecting data on the evolution of the intermediate species and final products
of combustion. Samples are collected at various locations along the reactor length, each
position representing a particular residence time and, in turn, a characteristic reaction
time. Therefore, maintaining the reactor at a constant (inlet) temperature, it is possible to
follow the evolution of a species as a function of the reaction time providing useful
information for establishing and evaluating reaction mechanisms.
The second procedure, known as Controlled Cool Down (CCD) or as Constant
Residence Time (CRT) methodology, was used in this study. A CCD experiment is
specifically designed to study the reactivity of a fuel over a wide range of temperatures
while keeping the residence time constant. The data collected during a CCD experiment
are usually represented in a plot of the CO production as function of the temperature,
creating a “reactivity map” of the fuel. A typical fuel reactivity map is presented in
Figure 3-2 where the start and end of reactivity and the NTC region are identified. CO
production is an indicator of the overall reactivity of the fuel at low and intermediate
temperature. At these temperature regimes, the decrease in CO concentration above
700 K is not due to formation of CO2 but due to reduction of reactivity. Reactivity maps
provide useful information on the oxidative behavior of a fuel, including the start and end
of reactivity, and the start and width of the NTC region.
55
1400
Low Temperature
Region
CO Concentration (ppm)
1200
NTC Region
Start of NTC
1000
800
600
400
Start of
Reactivity
End of
Reactivity
200
0
550
600
650
700
750
800
850
Temperature (K)
Figure 3–2. Typical fuel reactivity map
A CCD experiment is performed by heating the reactor up to the maximum
temperature that needs to be investigated, i.e., above the end of reactivity temperature
shown in Fig. 3-2 as determined by the low-to-intermediate temperature turnover.
Depending on the particular fuel studied and on its oxidative behavior, this temperature
usually falls in the range 750 K to 850 K. This preheat stage is accomplished by flowing
air at one atmosphere from the building compressed air system at the same bulk flow rate
as the experimental conditions so that the heat transfer rate remains constant and does not
alter the temperature profile.
Typically, the duration of the preheat stage takes
approximately 6-7 hours for the system to reach nearly isothermal conditions. Near the
end of the preheat stage, the temperature profile of the reactor is measured by taking
sample point temperature measurements along the length of the reactor, typically every
56
2.5 cm. The bead heaters along the length of the reactor are adjusted until a nearly
constant, less than +/- 5 ˚C, profile is achieved.
Once the maximum temperature is reached, the air flow is transferred to two
nitrogen flow streams, the first is the main “oxidizer” nitrogen stream and the second is
the nitrogen stream for fuel vaporization and delivery. Once the nitrogen flow rates are
stabilized, the fuel vaporization line is heated to its operating temperature. Next, the
reactor is pressurized to the desired operating conditions.
At this point, a flow of
approximately 56 l/m is continuously extracted to preheat the transfer lines and warmup
the online analyzers. Once the system is stabilized at the desired temperature, pressure,
and flow rates, fuel is introduced in order to calibrate the fuel flow rate to the desired fuel
concentration.
A J.U.M. Unburned Total Hydrocarbon (UTHC) is used to calibrate the fuel flow
rates. The procedure for calibration of the THC is outlined by Lenhert [2004b]. The
initial fuel flow rate is calculated using the fuel density and desired operating conditions.
Using this flow rate, the concentration was measured and compared to the calibrated
response of the instrument. This flow rate is usually, 10 to 20% higher than desired due
to the compressibility of the fuel and errors in the density. As a result, the concentration
from a second flow rate is measured at a 40% lower flow rate in order to bracket the
desired flow rate.
Next, the concentrations from three intermediate flow rates are
measured. A calibrated flow rate is then interpolated from the linear least squares fit of
the five flow rates and concentrations.
This flow rate is verified by measuring its
concentration and added it to the linear fit. If the new interpolated flow rate changed by
57
more than 0.010 ml/min, then two or three additional flow rates are used near the
interpolated flow rate to achieve an acceptable calibrated flow rate.
After calibration, a portion of the main nitrogen steam is transferred over to
oxygen and the system is allowed to stabilize. Then the 10 kW and 3 kW heaters are
shutdown – with the exclusion of the fuel line heater that is required to vaporize the fuel
prior to the injection – starting the cool down process for the entire system. The reactor
cooling rate is between 2 and 5 °C/min. During the cool down phase, the CO and
unburned hydrocarbons (UHC) were continuously monitored with the Siemens Ultramat
22P NDIR and J.U.M. THC analyzer. During the cooling process, the density of the
gases changes as this is a function of the temperature. Therefore the volumetric flow
changes and, in order to maintain the constant residence time during the entire
experiment, the probe position needs to be continuously adjusted by the computer
controlled positioning table. During the CCD experiments, the pressure of the system is
maintained at the desired pressure within ± 0.05 atm.
3.2
Engine Facility
The HCCI engine experiments in this study were conducted in a newly
developed Cooperative Fuel Research (CFR) engine test facility at Drexel University,
Fig. 3-3.
The engine is a 611.6 cm3, single cylinder, four-stroke, water cooled
Waukesha Motor Corporation Model 48D CFR engine direct coupled to a GE
CD258AT DC motor/ generator dynamometer. Under firing engine conditions, the
dynamometer maintained the speed within ±5 rpm of the set conditions. A flywheel on
58
the other end of the crank shaft dampens and reduces speed variations due to cyclic
variability.
Exhaust
Thermocouple
Inlet
Thermocouple
Pressure
Transducer
Pressure
Transducer
Data Control
&
Acquisition System
Pressure
Transducer
Turbulator
Waukesha CFR Engine
Oil
Thermocouple
Fuel
Injector
Exhaust
Gas Analyzers
CO, CO2 & NOx
Water Thermocouple
GE DC Motor
Dynamometer
Primary air
pressure regulator
Surge
Tank
Heated
Manifold
Air Flow
Controller
Air
Compressor
Figure 3–3. Schematic of Cooperative Fuel Research (CFR) engine facility
The engine is controlled by a General Electric DV-500 electric controller, with a
DV-300 control module. The controller operates on 480 VAC, three phase power,
supplied via a laboratory transformer. The control module converts the AC electrical
supply, to 300 VDC power. The control module is wired to the General Electric, 20
59
horsepower, 33 ampere, and direct current motor/generator dynamometer with a rated
maximum speed of 3000 revolutions per minute.
The engine has a 8.255 cm cylinder bore and a 11.43 cm piston stroke with a
cylinder head, which allows variation of the compression ratio from 4:1 to 18:1. Detailed
engine specifications are shown in Table 3-2.
Table 3-2. Cooperative Fuel Research engine geometry
3.2.1
Bore
82.55 mm
Stroke
114.3 mm
Displacement
611.6 cm3
Compression Ratio
4:1-18:1
Intake Valve Opens
10 deg bTDC
Intake Valve Closes
34 deg aBDC
Exhaust Valve Opens
40 deg bBDC
Exhaust Valve Closes
15 deg aTDC
Intake Manifold
The intake system refers to the basic air and fuel supply to the engine along with
the instruments and controllers to monitor and set the operating parameters. The engine
intake system is connected to the Laboratory compressed air system at a pressure of 5-7
atm, Fig. 3-4.
The compressed air system is filtered and the moisture content is
controlled, which enhances the ability to replicate intake conditions. In addition, the use
of a compressed air supply allows the engine to run under supercharged conditions.
60
The main air supply pressure is reduced to 4 atm via the primary intake regulator.
The regulated air is routed through a Porter Instruments model 114 mass flow meter.
This mass flow meter measures intake flow up to 1000 slpm with an accuracy of ±1 slpm,
and allows a maximum engine speed of 3000 rpm.
Compressed Air Supply
Gas Circulation
Heater
Primary
Pressure
Regulator
Aux. Gas
Supply
Port
&
Mass
Flow
Controller
Experimental
Fuel
Supply
Fuel
Injector
Mass Flow Meter
Temperature
Controlled
Air/Fuel
Mixture
Intake Bypass
Pressure
Transducer
Secondary
Pressure
Regulator
Thermocouple
Intake Manifold
Surge Tank
Figure 3–4. Intake system schematic
An auxiliary gas supply port was incorporated into the air delivery system for the
purposes of simulating the effects of exhaust gas recirculation (EGR). This port operates
at 3 atm and flow is maintained by a Porter Instruments model A202 mass flow controller.
61
This controller can meter up to 100 slpm of air with an accuracy of ±0.1 slpm. This
auxiliary flow merges with the main flow prior to final pressure regulation and delivery
to the engine.
The engine intake pressure is controlled by the secondary intake regulator. This
regulator, in conjunction with the intake pressure transducer, sets and maintains the
desired intake pressure. The maximum intake pressure, limited by the maximum pressure
of the intake transducer, is 2 atm. A 24.3 gallon pressure vessel serving as a surge tank
was secured in the base of the electrical rack to buffer intake air oscillations from the
engine, and allow for accurate mass flow measurement and control.
A Chromalox 5 kW gas circulation heater (Model GCHMTI) was used to preheat
the inlet air stream. This heater is a single zone immersion heater. This heater reduces
the hazards associated with the electrical power distribution, and the immersion heating
element allows for uniform intake heating.
Air/fuel mixing occurs immediately downstream of the intake heating. The test
fuel is injected by a standard 21 lb/hr automotive fuel injector. The fuel injector is
operated by a 12 VDC power supply, controlled by a LabVIEW virtual interface. Post
injection flow passes through a 17 inch long turbulator. The turbulator contains six sets
of oppositely mounted vanes which the intake air and atomized fuel flow through in order
to promote and achieve a homogeneous mixture for delivery to the engine.
Intake pressure and temperature are recorded prior to cylinder head inlet, via an
Omega Engineering 25 psia pressure transducer and 1/16” type-K thermocouple,
respectively. The measurements are stored on the laboratory data acquisition system.
62
The intake piping is wrapped with fiberglass insulation to limit the heat loss from
the intake flow. This is important since intake temperature is critical in maintaining the
vaporization of the fuel. A large temperature drop can cause the fuel to condense in the
intake system, resulting in improper mixture delivery to the engine.
3.2.2
Exhaust Manifold
The engine exhaust system is composed of two sub-systems, the main engine
exhaust and the engine crankcase breather, Fig. 3-5. The main engine exhaust is piped
out of the laboratory through the closest roof vent. The breather, which operates at
atmospheric pressure, was connected to the exhaust vent that serves as the outlet for the
gas analysis equipment.
Atmosphere
Gas Sample
Main Exhaust
Crankcase Breather
Pressure Transducer
Thermocouple
Engine
Figure 3–5. Exhaust system schematic
63
The main engine exhaust has a probe style gas sampling port. This port allows
the gas analysis equipment to withdraw a sample from the center of the exhaust flow,
decreasing the wall effects of the exhaust pipe on the exhaust gases.
The exhaust gas temperature is measured via a 1/16” type-K thermocouple
attached to the cylinder head exhaust port. The exhaust pressure is recorded via a
100 psig Omega Engineering pressure transducer.
The moderate range pressure
transducer was selected due to the potential for restricted exhaust flow in future
experimental setups. The pressure transducer is located two feet away from the engine to
reduce temperature effects on the measurement.
Two lengths of braided flex hose are used in the exhaust piping, to account for the
cylinder head movement, and to reduce the effects of vibration. High temperature,
vibration resistant pipe sealing compound, was used to prevent leakage from threaded
junctions.
3.2.3
Engine monitoring and data acquisition system
A Gurley Precision Instrument Company Model 9125 rotary incremental encoder
was used to produce a trigger signal once per revolution and a clock signal of 3600
periods per revolution. The encoder is mounted on a 2:1 reducing gear to provide a
resolution of 1/5 crank angle degree (CAD).
A Kistler Instruments model 7061B water-cooled pressure transducer was
installed in the cylinder head to monitor the in-cylinder pressure of the engine. The
transducer is coated with high temperature RTV material to protect against thermal shock.
64
The output of the signal is amplified by a Kistler model 5010B amplifier. The charge
amplifier signal is monitored and stored on a PC for the processing. The engine data
source map is shown in Fig. 3-6.
Intake
Assembly
•
•
•
•
•
Waukesha
CFR Engine
Mass Flow Rate
Heater Outlet Temperature
Fuel Flow Rate
Intake Temperature
Intake Manifold Pressure
•
•
•
•
In-Cylinder Pressure
Oil Temperature
Coolant Temperature
Fast Sampling Valve
Exhaust
Assembly
• Exhaust Temperature
• Exhaust Pipe Pressure
• Gas Sampling Analysis
CO, CO2, NOx and UHC
Figure 3–6. Engine data source map
With the inclusion of the shaft encoder programming into the data monitoring
plan, the core data acquisition system was modified by groups of undergraduate students
at Drexel University. The shaft encoder program was developed to run on a common
computer platform. An additional workstation was dedicated to handle the monitoring
and control of the engine controller, and passive data collection. LabVIEW virtual
instruments collected real time data, while serial communication was established with the
engine controller and the General Electric Control Toolbox software.
65
3.2.4
Experiment Methodologies and Approaches
For the HCCI study, the engine was operated at a speed of 800 rpm, intake
manifold pressure 1.0 bar, coolant temperature 80 ˚C and compression ratio at 16:1. The
experiments were run at inlet temperatures of 410, 450 and 500 K, which is above the
boiling points of the fuel mixtures thus eliminating concerns of fuel condensation. The
test fuels were injected into the air stream of the heated inlet manifold well upstream of
the intake valve to assure complete vaporization and mixing. The equivalence ratio was
pre-selected to one of six values between 0.28 and 0.57. n-Heptane (PRF0) iso-octane
(PRF100) and five blends (PRF20, PRF50, PRF63, PRF87 and PRF92) with and without
DTBP addition were tested.
3.3
Closure
This chapter provided a description of the two experimental facilities and
associated experimental techniques and methodologies used in this study.
The details of Pressurized Flow Reactor work examining the effect of DTBP on
the oxidation of SI primary reference fuels are discussed in Chapter 4. Details of the
corresponding engine based studies are provided in Chapter 5. The kinetic model
development work associated with both of these experimental activities is discussed in
Chapter 6.
66
CHAPTER 4. THE EFFECT OF DTBP ON OXIDATION OF SI PRIMARY
REFERENCE FUELS IN A PRESSURIZED FLOW REACTOR *
Reactivity maps of Spark Ignition (SI) primary reference fuels and their blends
with and without DTBP addition have been measured using the controlled cool down
methodology in the Drexel PFR facility. The results from experiments at 8 atm over the
range 600K < T <800 K are presented and discussed in this chapter. These results also
form the database for later modeling studies, which are discussed in Chapter 6.
4.1
Introduction
iso-Octane and n-heptane are the SI engine primary reference fuels (PRFs). The
knock resistance of possible fuels is defined by comparing their knock behavior to that of
mixtures of iso-octane and n-heptane. The octane number of the fuel is assigned based
on the volumetric percentage of iso-octane in the 2 component mixture. n-Heptane,
which is also used to represent a diesel fuel, and iso-octane have quite different oxidation
chemistries. Therefore, n-heptane, iso-octane and their mixtures are natural test fuels to
explore the effect of DTBP on CI and SI fuels.
In the present work, the oxidation of the PRFs for the octane number scale, isooctane (PRF100) and n-heptane (PRF0), and their blends, PRF20, PRF50, PRF63, PRF87
and PRF92, has been studied in a pressurized flow reactor. The objective of these
* The material in this chapter was the basis for Paper No. C03, presented at the 4th Joint
Meeting of the U.S. Sections of the Combustion Institute, Philadelphia, March 2005 [Gong et
al., 2005a]
67
experiments was to confirm the negative temperature coefficient (NTC) phenomena of
pure PRFs and to observe the effects of the additive di-tertiary butyl peroxide (DTBP) on
the oxidation of these fuels and their blends.
During the controlled cool down (CCD) reactivity mapping experiments, the
reactor was first heated to 800 K, which previous work had shown is higher than the end
temperature of NTC regions for all the tested PRFs. After the system had stabilized for
20 minutes, all heaters, except the one for the fuel delivery line, were turned off to let the
reactor slowly cool down at a rate of 2-5ºC/min. All experiments were conducted at
8 atm with the test fuels set to different equivalence ratios (φ), nitrogen dilutions and
residence times based on their respective reactivity behavior at low and intermediate
temperatures, Table 4.1.
Table 4-1. Pressurized flow reactor test conditions
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
Reactant Percentage (V/V liquid, %)
n-Heptane iso-Octane
n-Heptane
100
0
n-Heptane
100
0
n-Heptane + 0.5% DTBP
99.5
0
n-Heptane + 1.0% DTBP
99
0
n-Heptane + 1.5% DTBP
98.5
0
PRF20
80
20
PRF20 + 1.5% DTBP
78.8
19.7
PRF50
50
50
PRF50 + 1.5% DTBP
49.25
49.25
PRF63
37
63
PRF63 + 1.5% DTBP
36.45
62.05
n-Heptane
100
0
PRF87
13
87
PRF87 + 1.0% DTBP
12.87
86.13
PRF92
8
92
PRF92 + 1.0% DTBP
7.92
91.08
PRF92+ 1.5% DTBP
7.88
90.62
iso-Octane
0
100
iso-Octane + 1.0% DTBP
0
99
* [Tanaka et al., 2003]
DTBP
0.0
0.0
0.5
1.0
1.5
0.0
1.5
0.0
1.5
0.0
1.5
0.0
0.0
1.0
0.0
1.0
1.5
0.0
1.0
Oxidizer Comp. (%)
N2
Air
85
15
85
15
85
15
85
15
85
15
85
15
85
15
80
20
80
20
70
30
70
30
70
30
70
30
70
30
65
35
65
35
65
35
62
38
62
38
ON
φ
0.4
0
0.32
0
0.4
0.4
0.4
0.4
20
0.4
0.4
50
0.4
0.5
63
0.5
~0.064 0
0.5
87
0.5
0.6
92
0.6
0.6
0.75 100
0.75 91.2*
Reaction Time
(ms)
100
100
100
100
100
100
100
150
150
200
200
200
200
200
225
225
225
250
250
68
As shown in Table 4-1, iso-octane has the longest residence time (250 ms) and
highest equivalence ratio (φ = 0.75) due to its low reactivity within this temperature range.
A further explanation of the mixture selections is given when the data are presented in
discussion section.
The equivalence ratio is defined as:
φ=
( Fuel / Oxidizer ) Actual
( Fuel / Oxidizer ) Stoichiometric
For each CCD experiment, the residence time was kept constant while reducing
the reaction temperature via natural cooling from 800 to 600 K. The extracted gas
samples were directed to an online CO/CO2 NDIR (non-dispersive infrared) analyzer and
a FID total hydrocarbon (THC) analyzer.
The reactivity map for the experiment consists of a profile of carbon monoxide
(CO) concentration as a function of reaction temperature at constant pressure.
CO
concentration is used to characterize the degree of oxidation. The validation of using CO
as indicator of oxidation is based on: (1) CO is readily produced from hydrocarbon
oxidation in the low temperature regime; and (2) CO is not converted to CO2 at a
significant rate in this temperature range.
DTBP is known to decompose with a half-life of 10 ms at 550 K and less than
0.1 ms at 700 K [Griffiths et al., 1990]. The primary products of decomposition are
methyl radicals and acetone, as shown in Figure 4-1. Acetone is relatively unreactive,
69
and doesn’t oxidize fast enough at temperatures below 900 K to contribute to the
development of autoignition. By contrast, the methyl radicals undergo oxidation on a
microsecond timescale to yield molecular products (e.g., formaldehyde, methanol and
hydrogen peroxide) and to generate heat.
There is no direct evidence that chain
propagation is initiated from this secondary oxidation of methyl radicals and no chain
branching occurs [Griffiths et al., 1990]. The subsequent oxidation of formaldehyde and
the decomposition of hydrogen peroxide occur readily at temperatures in excess of 850 K,
and these reactions may promote chain initiation.
CH3
CH3•
O
CH3
C
O•
CH3
C
CH3
CH3
DTBP
CH3
CH3•
O
CH3
C
O•
CH3
C
CH3
CH3
Figure 4–1. DTBP thermal decomposition (Griffiths et al., 1990)
70
4.2
4.2.1
Results and Discussion
Reactivity of the SI PRFs and Their Blends
Reactivity maps for n-heptane, iso-octane and five of their mixtures (A, R, F, H, J,
M and O of Table 4-1) are shown in Figure 4-2. For all 7 conditions, the maps exhibit
typical negative temperature coefficient behavior.
As expected, n-heptane shows
significantly more reactivity than iso-octane. When iso-octane experiments were run at
the same experimental conditions as n-heptane, no reactivity was observed. Thus, the
iso-octane experiments were run at higher equivalence ratio (φ), lower dilution and longer
reaction time. Even at a much longer time of 250 ms and φ = 0.75, the CO peak for isooctane was only 250 ppm, while n-heptane produced 1250 ppm at 100 ms and φ = 0.40.
The starting temperatures of NTC behavior range from 705 K for n-heptane to 665 K for
iso-octane. In general, the temperature for peak CO concentration is lowered as the ON
of the reactants increases. It can also be seen that reactivity occurs over a narrower
temperature range as the ON increases. n-Heptane has the widest reactivity span, 625 to
775 K, while iso-octane has the narrowest, 630 to 680 K.
Figure 4-2 also shows that for blends with even small amounts of n-heptane, e.g.,
PRF87, the reactivity is much higher than for neat iso-octane. This is due to the faster
low and intermediate temperature reactions of n-heptane.
To examine the effect of iso-octane on the mixtures, two sets of experiments were
conducted. Each set was at the same experimental conditions, e.g. pressure, dilution, and
residence time, except for fuel type and concentration. The first set compared reactivity
of conditions M and L. In these experiments the n-heptane concentration in the reactants
was kept constant.
As shown in Fig. 4-3, the presence of iso-octane narrows the
71
temperature range of low and intermediate temperature reactivity. However, maximum
CO concentration and the temperature of this maximum are essentially unchanged. Data
from a second set of experiments, comparing conditions B and F are shown in Fig. 4-4.
The maximum CO concentration for case B and F are unchanged; however, unlike cases
M and L, the temperature range of reactivity is unchanged.
The iso-octane in these PRF blends act as a radical scavenger over the low and
intermediate temperature range, which can explain the observed narrowing of the range
for the first case that has higher iso-octane concentration.
CO Concentration (ppm)
1400
1200
R:iso-octane
H:PRF50
O:PRF92
F:PRF20
M:PRF87
A:n-heptane
J:PRF63
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–2. Reactivity maps for n-heptane, PRF20, PFR50, PRF63, PRF87,
PRF92 and iso-octane from CCD experiments in a PFR
72
700
M:PRF87
L:n-heptane
CO Concentration (ppm)
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–3. Reactivity maps for PRF87 and n-heptane at a constant nheptane concentration as listed in Table 4-1, cases M and L
1400
B:n-heptane at PRF20 level
F:PRF20
CO Concentration (ppm)
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–4. Reactivity maps for n-heptane, PRF 20 and n-heptane at the
PRF20 level as listed in Table 4-1, cases F and B
A schematic diagram for the branching pathways of low and intermediate
temperature regions is shown in Figure 4-5.
73
-H
RH
R&
+ O2 + M
RO2
+ RH
&
OH
H O& 2 + C = C
+ RH
+ O& H
Q& OOH
+β
R& + ROOH
+ Ether
C = C + R ′′CHO + O& H
H 2 O 2 + R&
+ O2
H 2 O + R&
.
.
OO Q OOH ⎯⎯→ O& H + O Q ' OOH ⎯⎯→ O& H + Rs ' C O + RsCHO
Figure 4–5. Branching pathways for hydrocarbon oxidation at low and
intermediate temperature
As noted previously, the oxidation of hydrocarbons can be separated into three
temperature regimes, the low, intermediate, and high temperature regimes. The low
temperature region is characterized by the reactions of RO2• radicals for smaller
hydrocarbon fuels or QOOH• radicals for larger hydrocarbon fuels and by the formation
of stable oxygenated hydrocarbons [Cernansky et al., 1986].
The intermediate
temperature region is dominated by the reactions of HO2• radicals and the characteristic
stable products are alkenes, stable oxygenated hydrocarbons, and methane. When the
process reaches the high temperature region, the reactions are dominated by OH•, O•,
and H• radicals, and unimolecular decomposition of alkyl radicals, via beta-scission,
becomes important [Wilk et al., 1986].
R• + O2 Ù RO2• and RO2• Ù QOOH• play important roles in low and
intermediate temperature regions. n-Heptane and iso-octane interact through a radical
74
pool of R• and RO2• in the low temperature region and R• in the intermediate
temperature region. Thus, iso-octane affects reactivity over the entire test region. The
observation that the maximum CO concentration is the same for both experiments in Fig.
4-3 and the comparable experiments in Fig. 4-4 can be explained by the fact that the CO
production in the low and intermediate temperature stage is due almost entirely to
reactions involving n-heptane.
The effect of fuel concentration on reactivity is checked in experiments A and B.
In these experiments the n-heptane concentration in the reactants changed, with
experiment B having a lower equivalence ratio (0.32) than experiment A (0.40). As
shown in Figure 4-6, the presence of more n-heptane (increase of equivalence ratio)
increases the overall reactivity, while the temperature ranges for NTC behavior remain
unchanged.
1400
A:n-heptane
B:n-heptane with less concentration
CO Concentration (ppm)
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–6. Reactivity maps for n-heptane at different concentration as
listed in Table 4-1, cases A and B
75
4.2.2
Effects of DTBP on Fuel Oxidation
The effects of the additive DTBP on the oxidation behavior of these PRFs and
their blends were also examined over the low and intermediate temperature regions,
Figures. 4-7 to 4-13. It can be seen that DTBP has a large effect on iso-octane oxidation.
As shown in Figure 4-6, 1.0% DTBP addition to iso-octane by volume increases the peak
CO concentration from 250 to 560 ppm. The temperature span of the reaction region
broadened from 60 to 110 ºC. This result agrees with data from a rapid compression
machine experiment [Tanaka et al., 2003].
700
R:iso-Octane
S:iso-Octane+1.0%DTBP
CO Concentration (ppm)
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–7. Reactivity maps for iso-octane and iso-octane + 1.5% DTBP as
listed in Table 4-1, cases R and S
However, unlike experimental results from Tanaka et al. [2003], no major effects
of DTBP addition on n-heptane and the PRF blends were found even at levels of DTBP
addition up to 1.5%, Figure 4-8 to Figure 4-13.
76
900
O:PRF92
P:PRF92+1.0%DTBP
Q:PRF92+1.5%DTBP
CO Concentration (ppm)
800
700
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–8. Reactivity maps for PRF92 with varying levels of DTBP additive
as listed in Table 4-1, cases O, P and Q
700
M:PRF87
N:PRF87+1.0%DTBP
CO Concentration (ppm)
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–9. Reactivity maps for PRF87 and PRF87 + 1.5% DTBP as listed in
Table 4-1, cases M and N
77
800
J:PRF63
K:PRF63+1.5%DTBP
CO Concentration (ppm)
700
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 4–10. Reactivity maps for PRF63 and PRF63 + 1.5% DTBP as listed
in Table 4-1, cases J and K
1200
H:PRF50
I:PRF50+1.5%DTBP
CO Concentration (ppm)
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
Temperature (K)
Figure 4–11. Reactivity map of PRF50 and PRF50 + 1.5% DTBP as listed in
Table 4-1, cases H and I
78
1200
F:PRF20
G:PRF20+1.5%DTBP
CO Concentration (ppm)
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
Temperature (K)
Figure 4–12. Reactivity maps for PRF20 and PRF20 + 1.5% DTBP as listed
in Table 4-1, cases F and G
A:n-Heptane
D:n-Heptane+1.0%DTBP
1400
C:n-Heptane+0.5%DTBP
E:n-Heptane+1.5%DTBP
CO Concentration (ppm)
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
Temperature (K)
Figure 4–13. Reactivity maps for n-heptane with varying levels of DTBP
additive as listed in Table 4-1, cases A, C, D and E
79
These observations can be explained if DTBP addition acts as a radical scavenger
in a mode similar to n-heptane. However, the effects of DTBP are much larger because
the activation energies for the initiation reactions are lower than those for n-heptane and
therefore it more readily produces the radical scavengers. The effect of DTBP on isooctane seems to show evidence of a direct, free radical chain initiation of hydrocarbon
oxidation, and an interaction with iso-octane is possible via the free radicals generated
from DTBP, thereby contributing to the initial heat release rate. Methyl radical, the
primary product of decomposition of DTBP, undergoes oxidation on a small enough
timescale to yield the molecular products formaldehyde, methanol and hydrogen peroxide.
With iso-octane, DTBP should not just raise the local temperature by exothermic
decomposition, but it also should have a direct chemical impact. For PRF blends, due to
the presence of n-heptane and the radical pool between n-heptane and iso-octane, the
chemical impact is not as obvious as with neat iso-octane.
4.3
Closure
In this chapter, the experimental results of the oxidation of the SI PRFs and their
blends in a pressurized flow reactor have been reported, and the effects of the additive
DTBP on these fuels were also reported and discussed.
All of the PRF components and blends exhibit typical negative temperature
coefficient behavior, with n-heptane showing significantly more reactivity than isooctane, as expected. In PRF blends, iso-octane acts as a radical scavenger and the
reactivity at low and intermediate temperatures is due almost entirely to reactions of nheptane.
80
DTBP addition was only effective in modifying the reactivity of iso-octane; no
changes were observed in the behavior of the n-heptane or the PRF blends tested even
with higher DTBP addition.
With DTBP addition to neat iso-octane, there is evidence of a radical chain
initiation of the hydrocarbon oxidation process. Thus, DTBP’s effect appears to be
chemical rather than just thermal.
81
CHAPTER 5.
EFFECTS OF DTBP ON THE COMBUSTION OF SI PRIMARY
REFERENCE FUELS IN AN HCCI ENGINE*
In this chapter, the effects of DTBP on spark ignition (SI) primary reference fuels
(PRFs, n-heptane and iso-octane) and their blends (PRF20, PRF50, PRF63, PRF87 and
PRF92) were investigated during HCCI engine operation. The results from Chapter 4 are
useful in analyzing the observed effects.
5.1
Introduction
In HCCI engines, different fuel combustion characteristics might be
required/desired in different HCCI operating ranges, e.g., for high load it is desirable to
use a fuel with a low cetane rating to delay the ignition to near TDC, while for low load a
fuel with a high cetane rating may be desirable. Therefore, understanding the combustion
characteristics of typical fuels is extremely important to realizing successful HCCI
operations. This need has motivated this part of the study.
A key to practical implementation of the HCCI concept is developing methods to
control combustion timing. Control methods must be designed to adjust the heat release
process to occur at the appropriate time in the engine cycle. One option for ignition
control is to use small amounts of ignition-enhancing additives to tailor the ignition
properties to the desired load condition.
* The material in this chapter was the basis for SAE Paper No. 2005-01-3740, to be
presented at the Powertrain & Fluid Systems Conference & Exhibition, San Antonio,
Texas, October 24-27, 2005 [Gong et al., 2005b].
82
The effects of DTBP on the combustion characteristics of the spark ignition
primary reference fuels, n-heptane and iso-octane, and their blends were studied for an
engine operating in the HCCI mode. Our goal was to see how much DTBP can really
affect engine performance.
A 611.6 cm3, single cylinder, Waukesha Motor Corporation Model 48D,
Cooperative Fuels Research (CFR) engine directly coupled to a GE CD258AT DC motor
dynamometer was used in the present study. Detailed engine specifications and a general
description of the overall facility are provided in Chapter 3.
The compression ratio was fixed at 16:1, the inlet manifold pressure was 1.0 bar
and the engine was operated at a constant speed of 800 rpm. The experiments were run at
inlet temperatures of 410, 450 and 500 K, which is above the boiling points of the fuel
mixtures thus eliminating concerns of fuel condensation. The test fuels were injected into
the air stream of the heated inlet manifold well upstream of the intake valve to assure
complete vaporization and mixing. Table 5.1 summarizes the test conditions. The
equivalence ratio was pre-selected to one of six values between 0.28 and 0.57. n-Heptane
(PRF0) iso-octane (PRF100) and five blends (PRF20, PRF50, PRF63, PRF87 and PRF92)
were tested.
83
Table 5-1. Engine test conditions
Reactant Percentage (V/V liquid, %)
n-Heptane iso-Octane
n-Heptane
100
0
n-Heptane + 1.5% DTBP
98.5
0
PRF20
80
20
PRF20 + 1.5% DTBP
78.8
19.7
PRF50
50
50
PRF50 + 1.5% DTBP
49.25
49.25
PRF63
37
63
PRF63 + 1.5% DTBP
36.45
62.05
PRF87
13
87
PRF87 + 1.5% DTBP
12.5
86
PRF92
8
92
PRF92+ 1.5% DTBP
7.88
90.62
iso-Octane
0
100
iso-Octane + 0.5% DTBP
0
99
iso-Octane + 1.5% DTBP
0
98.5
iso-Octane + 2.5% DTBP
0
97.5
: Tin = 410 K
5.2
5.2.1
Equivalence Ratio
DTBP 0.28 0.35 0.39 0.42 0.49
0.0
1.5
0.0
1.5
0.0
1.5
0.0
1.5
0.0
∆
∆
∆
∆
∆
1.5
∆
∆
∆
∆
∆
0.0
∆
∆
∆
∆
∆
1.5
∆
∆
∆
∆
∆
0.0
×
×
×
×
×
1.0
×
×
×
×
×
1.5
×
×
×
×
×
2.5
×
×
×
×
×
∆: Tin = 410 and 450 K
0.57
∆
∆
∆
∆
×
×
×
×
×: Tin = 410, 450 and 500 K
Experimental Results
Operating Range Definition
HCCI operation is limited by misfire at low loads and knocking at high loads,
therefore any study of HCCI must establish criteria determining the stable operating
range. At low load, due to reduced average combustion temperature and slow fuel
oxidation, engine cycle to cycle variation increases.
Hence, engine cycle to cycle
variation is used to evaluate the lower limit for HCCI stable operation. Cycle to cycle
variations of the combustion process in an engine can be monitored by the fluctuations in
both maximum cylinder pressure and the indicated mean effective pressure (IMEP). In
this study, IMEP fluctuation was used as a measure of the cycle-to-cycle variations and
84
was expressed as the coefficient of variation (COVIMEP).
The COVIMEP for 50
consecutive engine cycles was calculated as the standard deviation ( σ IMEP ) divided by the
mean value (IMEP) in percent [Koopmans and Denbratt, 2002].
There are two
definitions for IMEP. Net IMEP refers to the entire cycle while gross IMEP refers to the
compression and expansion strokes only. Net IMEP is used here; thus,
COV IMEP =
σ IMEP
IMEP
× 100%
As the load increases, the HCCI combustion rates also increase and intensify,
which gradually cause unacceptable noise, potential engine damage, and unacceptable
levels of NOx emissions. Therefore, the upper stability limit of HCCI combustion can be
defined with respect to the maximum rate of pressure rise in the cylinder.
A COVIMEP of 3% -5% is generally used as the stability limit by most researchers.
However, in this study, we are interested in showing that the addition of DTBP to SI
PRFs has the ability to change the COV in a consistent manner, thus a 10% COVIMEP
level was used to define our lower stability limit.
A maximum pressure rise of
1.0 MPa/CAD (dP/dφmax = 1.0 MPa/CAD) was chosen as the upper bound. These
definitions were used for all fuels and test conditions. Both COVIMEP and dP/dφmax were
determined from the recorded cylinder pressure histories.
85
5.2.2
The Effect of Fuel on In-Cylinder Pressure
Figure 5-1 shows some typical pressure traces for HCCI operation at φ = 0.42 and
Tin = 410 K with the seven different test fuels, with and without DTBP addition.
70
PRF63
Pressure (Bar)
60
50
40
PRF87
PRF50
PRF92
PRF20
30
PRF100
PRF0
20
Motored
10
0
300
315
330 345 360 375
Crank Angle (CAD)
390
405
(a)
70
PRF92
PRF87
Pressure (Bar)
60
50
40
30
20
PRF63
PRF100
PRF50
PRF20
PRF0
10
0
300
Motored
315
330 345 360 375
Crank Angle (CAD)
390
405
(b)
Figure 5–1. Typical pressure traces for HCCI operation with the different test
fuels at: (a) φ = 0.42 and Tin = 410 K; (b) φ = 0.42, Tin = 410 K and 1.5% DTBP
86
In general, at the same equivalence ratio the lower the octane number of a fuel the
shorter its ignition delay time. In our experiments, n-heptane had the earliest ignition
timing and iso-octane had the latest, as expected. The two highest octane number fuels
tested, PRF92 and PRF100, did not show hot ignition under these experimental
conditions. 1.5% DTBP addition promoted oxidation and ignition for all fuels, but
particularly so for the higher octane fuels, PRF87, PRF92 and iso-octane. Hot ignitions
were observed for all test fuels with the addition of DTBP, as shown in Figure 5-1 (b).
Neat n-heptane, PRF20, PRF50 and PRF63 exhibit a two stage ignition behavior,
indicated by the early pressure rise preceding the main combustion event. As shown in
Figure 5-2 to 5-5, both pure PRF fuels and PRF fuels with DTBP addition show typical
two stage ignition at different equivalence ratios.
70
Pressure (Bar)
60
PRF0+1.5%DTBP
PRF0
Motored
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–2. Two stage ignition of n-heptane and n-heptane + 1.5% DTBP
at φ = 0.39 and Tin = 410 K
87
60
Pressure (Bar)
50
PRF20+1.5%DTBP
PRF20
Motored
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–3. Two stage ignition of PRF20 and PRF20 + 1.5% DTBP
at φ = 0.28 and Tin = 410 K
70
Pressure (Bar)
60
PRF50+1.5%DTBP
PRF50
Motored
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–4. Two stage ignition of PRF50 and PRF50 + 1.5% DTBP
at φ = 0.42 and Tin = 410 K
88
60
Pressure (Bar)
50
PRF63+1.5%DTBP
PRF63
Motored
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–5. Two stage ignition of PRF63 and PRF63 + 1.5% DTBP
at φ = 0.35 and Tin = 410 K
Pure PRF87 does not clearly exhibit two stage ignition behavior at inlet
temperature of 410 K, neither at relatively high equivalence ratio (i.e., Figure 5-6, φ =
0.57) nor at low equivalence ratio (i.e., Figure 5-7, φ = 0.39). However, with the addition
of 1.5% DTBP, PRF87 exhibits two stage ignition, Figure 5-6 and 5-7. Similar results
are observed at inlet temperature of 450 K too, Figure 5-8 and 5-9.
89
70
Pressure (Bar)
60
PRF87+1.5%DTBP
PRF87
Motored
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–6. Two stage ignition of PRF87 + 1.5% DTBP and single
stage ignition of PRF 87 at φ = 0.57 and Tin = 410 K
60
Pressure (Bar)
50
PRF87+1.5%DTBP
PRF87
Motored
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–7. Two stage ignition of PRF87 + 1.5% DTBP at φ = 0.39
and Tin = 410 K
90
PRF87+1.5%DTBP
PRF87
Motored
60
Pressure (Bar)
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–8. Two stage ignition of PRF87 + 1.5% DTBP and single
stage ignition of PRF87 at φ = 0.35 and Tin = 450 K
70
Pressure (Bar)
60
PRF87+1.5%DTBP
PRF87
Motored
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–9. Two stage ignition of PRF87 + 1.5% DTBP and single
stage ignition of PRF87 at φ = 0.57 and Tin = 450 K
91
Similar to PRF87, PRF92 does not exhibit any two stage ignition behavior.
However, unlike PRF87, there is no clear indication of two stage ignition with the
addition of DTBP to PRF92 at inlet temperature of 410 K and 450 K, Fig. 5-10 and 5-11.
60
Pressure (Bar)
50
PRF92+1.5%DTBP
PRF92
Motored
40
30
20
10
0
315
330
345
360
375
390
405
Crank Angle (CAD)
Figure 5–10. Single stage ignition of PRF92 + 1.5% DTBP at φ = 0.42 and
Tin = 410 K
70
Pressure (Bar)
60
PRF92+1.5%DTBP
PRF92
Motored
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–11. Single stage ignition of PRF92 + 1.5% DTBP and PRF92 at
φ = 0.49 and Tin = 450 K
92
The pressure traces and heat release curves for pure iso-octane and iso-octane
with 3 different DTBP levels are shown in Fig. 5-12. The thermodynamic model for
calculating heat release from measured pressure data was based on the work of Ferguson
et al. [1987], Li et al. [1995], and Zheng et al., [2001]. It is a one-zone model in which
the boundary layer’s effect is considered.
Both pressure and heat release curves indicate that iso-octane only exhibits single
stage ignition since there are no obvious pressure and heat release changes presence of a
first stage ignition, even with 2.5% DTBP addition. However, the addition of DTBP
35
350
30
300
25
Motored
PRF100
PRF100+0.5%DTBP
PRF100+1.5%DTBP
PRF100+2.5%DTBP
M t d
20
15
250
200
150
10
100
5
50
0
0
340
345
350 355 360
Crank Angle (CAD)
365
R ate of H eat R elease
(J/C A D )
Pressure (B ar)
significantly advanced the ignition timing and shortened the combustion duration.
370
Figure 5–12. The effect of DTBP concentration on iso-octane autoignition
at φ = 0.57 and Tin = 450 K
93
5.2.3
The Effect of Equivalence Ratio on In-Cylinder Pressure
Figures 5-13 to 5-19 show the effect of equivalence ratio on pressure traces of all
7 test fuels, with and without DTBP addition at Tin = 410 K. As expected, the peak
cylinder pressure increases and advances with increasing equivalence ratio, due to the
larger energy release and higher cylinder temperature and the resulting accelerated
chemical reaction.
70
Pressure (Bar)
60
50
PRF0
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
PRF0+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–13. The effect of equivalence ratio on PRF0 autoignition
at Tin = 410 K
94
PRF20
70
Pressure (Bar)
60
50
40
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
PRF20+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–14. The effect of equivalence ratio on PRF20 autoignition
at Tin = 410 K
As shown in Figures 5-13, 5-14, 5-15 and 5-16, for low octane number fuels,
HCCI operation can be realized at relatively low equivalence ratios, such as φ = 0.28 and
0.35. For the high octane number fuels, the in-cylinder charge is hard to ignite at lean
conditions, as shown in Figures 5-17, 5-18 and 5-19.
95
PRF50
70
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
Pressure (Bar)
60
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
PRF50+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motered
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–15. The effect of equivalence ratio on PRF50 autoignition
at Tin = 410 K
96
70
Pressure (Bar)
60
50
40
PRF63
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
70
Pressure (Bar)
60
50
40
PRF63+1.5%DTBP
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–16. The effect of equivalence ratio on PRF63 autoignition
at Tin = 410 K
97
PRF87
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
PRF87+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–17. The effect of equivalence ratio on PRF87 autoignition
at Tin = 410 K
98
PRF92
60
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
Pressure (Bar)
50
40
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
PRF92+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–18. The effect of equivalence ratio on PRF92 autoignition
at Tin = 410 K
99
PRF100
40
35
Pressure (Bar)
30
25
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
20
15
10
5
0
315
330
345
360
375
390
Crank Angle (CAD)
80
70
Pressure (Bar)
60
50
40
PRF100+1.5%DTBP
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–19. The effect of equivalence ratio on PRF100 autoignition
at Tin = 410 K
100
The pressure traces for PFR87, PRF92 and PRF100 at inlet temperature of 450 K
for different equivalence ratios are also shown in Figures 5-20 to 5-22. With the increase
of inlet temperature, HCCI operation extends to lower equivalence ratios.
PRF87
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
*
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
70
Pressure (Bar)
60
50
40
PRF87+1.5%DTBP
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–20. The effect of equivalence ratio on PRF87 autoignition
at Tin = 450 K
101
80
PRF92
70
Pressure (Bar)
60
50
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
40
30
20
10
0
315
330
345
360
375
390
375
390
Crank Angle (CAD)
PRF92+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
Crank Angle (CAD)
Figure 5–21. The effect of equivalence ratio on PRF92 autoignition
at Tin = 450 K
102
70
PRF100
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
Pressure (Bar)
60
50
40
30
20
10
0
330
345
360
375
390
Crank Angle (CAD)
PRF100+1.5%DTBP
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–22. The effect of equivalence ratio on PRF100 autoignition
at Tin = 450 K
103
The pressure traces for PRF100 at inlet temperature of 500 K are shown in
Fig. 5-23.
PRF100
70
Pressure (Bar)
60
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
70
Pressure (Bar)
60
50
40
PRF100+1.5%DTBP
Phi=-0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
375
390
Crank Angle (CAD)
Figure 5–23. The effect of equivalence ratio on PRF100 autoignition
at Tin = 500 K
104
5.2.4
The Effect of DTBP Concentration on iso-Octane
Figure 5-24 shows the effect of DTBP concentration level on the ignition and
combustion behavior of iso-octane at an inlet temperature of 410 K for several
equivalence ratios. With the increase of DTBP concentration in the mixture, stable
operation was expanded to lower equivalence ratios for iso-octane. Similar effects were
also seen for the other inlet temperatures, Figure 5-25 and 5-26.
PRF100
40
30
25
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
60
50
Pressure (Bar)
35
Pressure (Bar)
PRF100+0.5%DTBP
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
20
15
10
40
30
20
10
5
0
315
0
330
345
360
375
390
315
330
Crank Angle (CAD)
345
(a)
80
40
60
50
Pressure (Bar)
Pressure (Bar)
50
30
40
20
10
10
0
330
405
345
360
Crank Angle (CAD)
(c)
375
390
390
405
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
315
390
PRF100+2.5%DTBP
70
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
60
375
(b)
PRF100+1.5%DTBP
70
360
Crank Angle (CAD)
0
315
330
345
360
375
Crank Angle (CAD)
(d)
Figure 5–24. The effect of DTBP addition on iso-octane autoignition
at Tin = 410 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP
105
70
PRF100
50
40
60
Pressure (Bar)
Pressure (Bar)
60
PRF100+0.5%DTBP
70
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
50
40
30
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
20
10
0
0
330
345
360
375
390
315
330
Crank Angle (CAD)
(a)
50
40
60
Pressure (Bar)
Pressure (Bar)
70
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
60
360
375
390
(b)
PRF100+1.5%DTBP
70
345
Crank Angle (CAD)
30
20
10
50
40
PRF100+2.5%DTBP
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
0
315
330
345
360
Crank Angle (CAD)
(c)
375
390
315
330
345
360
Crank Angle (CAD)
(d)
Figure 5–25. The effect of DTBP addition on iso-octane autoignition
at Tin = 450 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP
375
390
106
PRF100
70
50
40
60
Pressure (Bar)
Pressure (Bar)
60
PRF100+0.5%DTBP
70
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
50
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
0
315
330
345
360
375
390
315
330
345
Crank Angle (CAD)
(a)
70
375
390
70
60
50
Pressure (Bar)
Pressure (Bar)
40
390
PRF100+2.5%DTBP
PRF100+1.5%DTBP
50
375
(b)
Phi=-0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
60
360
Crank Angle (CAD)
30
20
10
40
Phi=0.57
Phi=0.49
Phi=0.42
Phi=0.39
Phi=0.35
Phi=0.28
Motored
30
20
10
0
315
330
345
360
Crank Angle (CAD)
(c)
375
390
0
315
330
345
360
Crank Angle (CAD)
(d)
Figure 5–26. The effect of DTBP addition on iso-octane autoignition
at Tin = 500 K: (a) PRF100; (b) PRF100 + 0.5%DTBP;
(c) PRF100 + 1.5%DTBP; (d) PRF100 + 2.5%DTBP
107
5.2.5
The Effect of DTBP on Ignition Timing
Table 5-2 lists the ignition timings for all test conditions, where NI indicates an
absence of hot ignition and COV indicates a COVIMEP exceeding 10%. It shows that the
addition of DBTP always reduced the ignition delay time, with the maximum of 12.8
CAD for PRF87 at inlet temperature of 410 K and equivalence ratio of 0.42. However,
the magnitude and φ dependence of the reduction varied.
Table 5-2. DTBP effect on ignition timing
T in (K) Fuel
n-Heptane
n-Heptane + 1.5% DTBP
PRF20
PRF20 + 1.5% DTBP
410
PRF50
PRF50 + 1.5% DTBP
PRF63
PRF63 + 1.5% DTBP
PRF87
PRF87 + 1.5% DTBP
PRF87
450
PRF87 + 1.5% DTBP
PRF92
410
PRF92+ 1.5% DTBP
PRF92
450
PRF92+ 1.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
410
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
450
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
500
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
0.28
342.0
336.0
343.2
335.8
345.6
344.2
351.8
348.2
NI
COV
NI
COV
NI
NI
NI
COV
NI
NI
NI
NI
NI
NI
NI
COV
NI
COV
COV
351.0
Ignition timing at different
0.35
0.39
0.42
338.0
337.0
336.0
333.0
332.0
332.0
339.6
337.6
335.4
331.0
329.0
328.1
344.8
345.4
343.2
341.6
341.0
340.0
351.4
349.9
346.4
345.6
343.2
341.6
NI
NI
363.8
COV
351.4
351.0
COV
359.2
358.4
348.8
348.4
346.2
NI
NI
NI
COV
355.0
353.8
NI
NI
COV
351.4
349.4
349.6
NI
NI
NI
NI
NI
NI
NI
COV
359.0
COV
355.8
354.8
NI
NI
NI
NI
COV
COV
COV
356.0
354.6
COV
351.8
350.6
COV
COV
COV
COV
351.6
351.4
COV
351.2
350.8
350.0
348.2
347.2
φ’s
0.49
332.0
326.0
334.2
324.8
341.0
336.4
344.8
339.2
354.4
347.8
349.6
344.0
NI
351.4
353.2
346.2
NI
NI
357.8
353.6
NI
354.6
351.8
349.4
351.6
348.4
347.4
344.0
(CAD)
0.57
353.8
347.4
348.8
343.8
357.2
351.4
353.4
346.0
NI
362.2
351.0
351.0
357.8
351.8
349.2
346.8
349.8
345.2
345.2
341.8
108
With the increase of octane number, autoignition becomes much more difficult
and the stable operating range becomes narrower.
iso-Octane did not undergo hot
ignition for any tested φ at Tin = 410 K, and the autoignition of pure iso-octane occurred
only with φ = 0.57 and Tin = 450 K. Similar instances were also seen in cases of PRF92
and PRF87. Table 5-2 also shows that stable HCCI operation can be reached by a small
addition of DTBP for the high octane number fuels at low inlet temperatures and low
equivalence ratios. For example, for iso-octane at Tin = 410 K, stable operation can be
realized at φ = 0.39 and above with 2.5% volume addition of DTBP.
The effect of DTBP on the ignition timing also increases as its concentration in
Ignition Timing Reduction (CAD)
the fuel mixture increases, as shown in Figures 5-27 and 5-28.
12.0
0.5%DTBP
1.5%DTBP
2.5%DTBP
10.0
8.0
6.0
4.0
2.0
0.0
0.35
0.40
0.45
0.50
0.55
0.60
Equivalence Ratio
Figure 5–27. The effect of DTBP addition on ignition timing reduction
for neat iso-octane at selected φ’s and Tin = 500 K
109
PRF0
PRF20
PRF50
PRF63
Ignition Timing (CAD)
355
PRF0+1.5% DTBP
PRF20+1.5% DTBP
PRF50+1.5% DTBP
PRF63+1.5% DTBP
350
345
340
335
330
325
320
0.25
0.30
0.35
0.40
0.45
0.50
Equivalence Ratio
Figure 5–28. The effect of DTBP addition on ignition timing at selected φ’s
for PRF0, PRF20, PRF50 and PRF63at Tin = 410 K
5.2.6
Effect of DTBP on IMEP and Cycle to Cycle Variations
Figure 5-29 shows IMEP as a function of equivalence ratio for all the fuels being
considered.
The data show that IMEP is not always an increasing function of
equivalence ratio at the test conditions, mostly due to the early ignition timing.
Significant differences of IMEP are observed for different fuels. High octane number
fuels, such as iso-octane and PRF92 have higher IMEP than the lower octane number
fuels. As might be expected at compression ratio of 16 and speed of 800 rpm, the low
octane number fuels are not practical for HCCI operation since the ignition timings
precede TDC. The IMEP for all tested fuels was smaller with the addition of DTBP due
to advanced ignition timing.
110
2.5
PRF0
PRF20
PRF50
PRF63
PRF0+1.5% DTBP
PRF20+1.5% DTBP
PRF50+1.5% DTBP
PRF63+1.5% DTBP
IMEP (Bar)
2
1.5
1
0.5
0
0.25
0.3
0.35
0.4
0.45
0.5
Equivalence Ratio
Figure 5–29. The effect of equivalence ratio on IMEP for PRF0, PRF20,
PRF50 and PRF63 at Tin = 410 K
Examples of the effects of DTBP on cycle to cycle pressure variations are shown
in Figures 5-30 and 5-31. Figure 5-30 shows the pressure traces for eight consecutive
cycles for PRF100 at inlet temperature of 450 K and equivalence ratio of 0.49. It clearly
indicates major variations between each cycle. It also shows poor combustion in terms of
pressure rise (misfire and partial burn, PMax < 30 bar in all cases). With the addition of
1.5% DTBP, Figure 5-31, the pressure trace stability has been improved significantly,
and hot ignition occurred with consisting firing and burn of charge (PMax > 60 bar in all
cases). Similar behavior was also observed with the other test fuels; Figures 5-32 and 533 show the results for PRF92 at inlet temperature of 450 K and equivalence ratio of 0.49,
and Figures 5-34 and 5-35 show the results for PRF87 at inlet temperature of 410 K and
equivalence ratio of 0.42.
111
30
1
25
20
1
15
10
5
340
360
380
20
400
10
5
Crank Angle (CAD)
15
2
10
5
1060
1080
1100
30
20
15
3
10
5
1780
1800
6
10
5
3940
3960
3980
4000
Crank Angle (CAD)
30
3
25
6
15
0
3920
1120
1820
7
25
Pressure (Bar)
Pressure (Bar)
3280
20
Crank Angle (CAD)
20
10
5
0
4640
1840
7
15
Crank Angle (CAD)
4660
4680
4700
4720
Crank Angle (CAD)
30
30
4
20
15
4
10
5
2500
2520
2540
Crank Angle (CAD)
8
25
Pressure (Bar)
25
Pressure (Bar)
3260
25
Pressure (Bar)
Pressure (Bar)
2
20
0
2480
3240
30
25
0
1760
3220
Crank Angle (CAD)
30
0
1040
5
15
0
3200
0
320
5
25
Pressure (Bar)
Pressure (Bar)
30
2560
20
15
10
5
0
5360
5380
5400
5420
Crank Angle (CAD)
Figure 5–30. Pressure variation for eight consecutive cycles for PRF100
at φ = 0.49 and Tin = 450 K
5440
112
70
70
1
50
40
30
20
10
0
320
340
360
380
50
40
30
20
10
0
3200
400
Crank Angle (CAD)
2
Pressure (Bar)
Pressure (Bar)
40
30
20
10
1060
1080
1100
6
40
30
20
10
0
3920
1120
3940
3960
3980
4000
Cr ank Angle (CAD)
70
70
3
50
40
30
20
10
1780
1800
1820
7
60
Pressure (Bar)
60
Pressure (Bar)
3280
50
Cr ank Angle (CAD)
50
40
30
20
10
0
4640
1840
Crank Angle (CAD)
4660
4680
4700
4720
Crank Angle (CAD)
70
70
4
50
40
30
20
10
2500
2520
2540
Cr ank Angle (CAD)
8
60
Pressure (Bar)
60
Pressure (Bar)
3260
60
50
0
2480
3240
70
60
0
1760
3220
Cr ank Angle (CAD)
70
0
1040
5
60
Pressure (Bar)
Pressure (Bar)
60
2560
50
40
30
6
20
10
0
5360
5380
5400
5420
Cr ank Angle (CAD)
Figure 5–31. Pressure variation for eight consecutive cycles for
PRF100 + 1.5% DTBP at φ = 0.49 and Tin = 450 K
5440
113
80
60
50
40
30
20
10
0
320
5
70
Pressure (Bar)
Pressure (Bar)
80
1
70
60
50
40
30
20
10
0
340
360
380
3200
400
Crank Angle (CAD)
80
50
40
30
20
6
60
50
40
30
20
0
0
1060
1080
1100
3920
1120
80
3980
4000
50
40
30
20
10
7
70
Pressure (Bar)
Pressure (Bar)
60
60
50
40
30
20
10
0
0
1780
1800
1820
1840
4640
4660
4680
4700
80
80
4
8
70
Pressure (Bar)
70
4720
Crank Angle (CAD)
Crank Angle (CAD)
Pressure (Bar)
3960
80
3
70
3940
Crank Angle (CAD)
Crank Angle (CAD)
60
50
40
30
20
60
50
40
30
\
20
10
10
0
0
2480
3280
10
10
1760
3260
70
Pressure (Bar)
Pressure (Bar)
60
1040
3240
80
2
70
3220
Crank Angle (CAD)
2500
2520
2540
Crank Angle (CAD)
2560
5360
5380
5400
5420
Crank Angle (CAD)
Figure 5–32. Pressure variation for eight consecutive cycles for
PRF92 at φ = 0.49 and Tin = 450 K
5440
114
70
50
40
30
20
10
0
50
40
30
20
10
0
320
340
360
380
3200
400
Crank Angle (CAD)
70
3260
3280
50
40
30
20
10
6
60
Pressure (Bar)
Pressure (Bar)
3240
70
0
50
40
30
20
10
0
1040
1060
1080
1100
1120
3920
Crank Angle (CAD)
70
3940
3960
3980
4000
Crank Angle (CAD)
70
3
50
40
30
20
10
7
60
Pressure (Bar)
60
Pressure (Bar)
3220
Crank Angle (CAD)
2
60
0
50
40
30
20
10
0
1760
1780
1800
1820
1840
4640
Crank Angle (CAD)
4660
4680
4700
4
8
60
Pressure (Bar)
60
4720
Crank Angle (CAD)
70
70
Pressure (Bar)
5
60
Pressure (Bar)
Pressure (Bar)
70
1
60
50
40
30
20
10
0
50
40
30
20
10
0
2480
2500
2520
2540
Crank Angle (CAD)
2560
5360
5380
5400
5420
Crank Angle (CAD)
Figure 5–33. Pressure variation for eight consecutive cycles for
PRF92 + 1.5% DTBP at φ = 0.49 and Tin = 450 K
5440
115
60
1
50
Pressure (Bar)
Pressure (Bar)
60
40
30
20
10
40
30
20
10
0
0
320
340
360
380
3200
400
Crank Angle (CAD)
40
30
20
10
1040
1080
1120
1160
1200
30
20
10
3920
3960
3980
60
3
50
3940
4000
Crank Angle (CAD)
40
30
20
10
7
50
Pressure (Bar)
Pressure (Bar)
6
0
1000
60
0
40
30
20
10
0
1780
1800
1820
1840
4560
4600
4640
4680
4720
4760
60
60
4
8
50
Pressure (Bar)
50
4800
Crank Angle (CAD)
Crank Angle (CAD)
40
30
20
10
0
2480
3280
40
Crank Angle (CAD)
1760
3260
50
0
960
3240
60
2
50
3220
Crank Angle (CAD)
Pressure (Bar)
Pressure (Bar)
60
Pressure (Bar)
5
50
40
30
20
10
0
2500
2520
2540
Crank Angle (CAD)
2560
5360
5380
5400
5420
Crank Angle (CAD)
Figure 5–34. Pressure variation for eight consecutive cycles for
PRF87 at φ = 0.42 and Tin = 410 K
5440
116
70
50
40
30
20
10
0
50
40
30
20
10
0
320
340
360
380
3200
400
Crank Angle (CAD)
70
3260
3280
50
40
30
20
10
6
60
Pressure (Bar)
Pressure (Bar)
3240
70
0
50
40
30
20
10
0
1040
1060
1080
1100
1120
3920
Crank Angle (CAD)
70
3940
3960
3980
4000
Crank Angle (CAD)
70
3
50
40
30
20
10
7
60
Pressure (Bar)
60
Pressure (Bar)
3220
Crank Angle (CAD)
2
60
0
50
40
30
20
10
0
1760
1780
1800
1820
1840
4640
Crank Angle (CAD)
4660
4680
4700
50
40
30
20
10
8
60
Pressure (Bar)
4
60
4720
Crank Angle (CAD)
70
70
Pressure (Bar)
5
60
Pressure (Bar)
Pressure (Bar)
70
1
60
50
40
30
20
10
0
0
2480
2500
2520
2540
Crank Angle (CAD)
2560
5360
5380
5400
5420
Crank Angle (CAD)
Figure 5–35. Pressure variation for eight consecutive cycles for
PRF87 + 1.5%DTBP at φ = 0.42 and Tin = 410 K
5440
117
Figure 5-34 clearly indicates major variations between each cycle for PRF87 at
Tin = 410 K. It also shows poor combustion in terms of earlier ignition (i.e., cycle 2 and 7)
and partial burn (all the cases). With the addition of 1.5% DTBP, Figure 5-35, the
pressure stability has been improved significantly.
Comparisons of the peak pressure variations for these same PRF100, PRF92 and
PRF87 cases are shown in Figures 5-36, 5-37 and 5-38. The addition of DTBP reduces
the peak pressure variation significantly and, as noted previously, actually enables hot
ignition in the PRF100 cases.
Table 5-3 lists the COVIMEP and IMEP for all tested conditions. In order to have a
better understanding of the effect of DTBP on COVIMEP itself, all COVIMEP values are
included while NI indicates no hot ignition. As we know, the decrease of the average
combustion temperature of HCCI operation at low load leads to an increase in cycle to
cycle variations. Thus, the overall COVIMEP at these conditions is relatively large, due to
a combination of low IMEP along with the early ignition timing for the low RON fuels.
In fact, because of the early ignition timing with the low octane number fuels, the
addition of DTBP didn’t result in much improvement with respect to cycle to cycle
variations. However, for higher octane number fuels, the addition of DTBP typically
reduced the cyclic variability and significantly improved the stability.
118
PRF100
30
P re s s u re (B a r)
25
20
15
10
5
0
0
720
1440
2160
2880
3600
4320
5040
5760
6480
4320
5040
5760
6480
Crank Angle (CAD)
PRF100+1.5%DTBP
70
P ressure (B ar)
60
50
40
30
20
10
0
0
720
1440
2160
2880
3600
Crank Angle (CAD)
Figure 5–36. Comparison of peak pressure variation for eight consecutive cycles
for PRF100 and PRF100 + 1.5%DTBP at φ = 0.49 and Tin = 450 K
119
PRF92
80
P r e s s u r e (B a r )
70
60
50
40
30
20
10
0
0
720
1440
2160
2880
3600
4320
5040
5760
6480
4320
5040
5760
6480
Crank Angle (CAD)
PRF92+1.5%DTBP
70
P re s s u re ( B a r)
60
50
40
30
20
10
0
0
720
1440
2160
2880
3600
Crank Angle (CAD)
Figure 5–37. Comparison of peak pressure variation for eight consecutive cycles
for PRF92 and PRF92 + 1.5%DTBP at φ = 0.49 and Tin = 450 K
120
PRF87
60
P re s s u re ( B a r)
50
40
30
20
10
0
0
720
1440
2160
2880
3600
4320
5040
5760
6480
4320
5040
5760
6480
Crank Angle (CAD)
PRF87+1.5%DTBP
70
P re s s u re ( B a r)
60
50
40
30
20
10
0
0
720
1440
2160
2880
3600
Crank Angle (CAD)
Figure 5–38. Comparison of peak pressure variation for eight consecutive cycles
for PRF87 and PRF87 + 1.5%DTBP at φ = 0.42 and Tin = 410 K
121
Table 5-3. DTBP effect on COVIMEP
COV IMEP (%) / IMEP (Bar) at different Equivalence Ratios
Tin Fuel
n-Heptane
n-Heptane + 1.5% DTBP
PRF20
410 PRF20 + 1.5% DTBP
PRF50
PRF50 + 1.5% DTBP
PRF63
PRF63 + 1.5% DTBP
PRF87
410
PRF87 + 1.5% DTBP
PRF87
450
PRF87 + 1.5% DTBP
PRF92
410
PRF92+ 1.5% DTBP
PRF92
450
PRF92+ 1.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
410
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
450
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
iso-Octane
iso-Octane + 0.5% DTBP
500
iso-Octane + 1.5% DTBP
iso-Octane + 2.5% DTBP
0.28
0.35
0.39
0.42
0.49
0.57
5.93/1.18 7.07/1.1 4.53/1.15 6.38/1.12 7.16/1.00
9.40/0.61 7.30/0.70 8.83/0.72 8.27/0.63 7.45/0.54
9.18/0.98 5.17/1.02 7.00/1.01 5.90/1.16 9.54/1.04
8.25/0.79 9.41/0.75
9.90/1.04 8.73/1.47
9.10/0.89 7.42/1.13
9.20/1.80 9.40/2.05
9.00/1.50 3.94/1.81
NI
NI
20.21/1.67 9.60/2.14
NI
11.5/2.4
10.4/1.66 8.75/1.97
NI
NI
NI
11.0/2.11
NI
NI
16.0/1.81 7.90/2.18
NI
NI
NI
NI
NI
NI
NI
18.5/2.56
NI
NI
NI
NI
NI
13.2/2.06
21.3/1.45 14.0/1.97
NI
17.4/2.07
18.0/1.59 14.7/2.00
14.7/1.39 13.5/1.95
10.1/1.76 9.50/2.12
9.10/0.68
7.82/1.5
9.35/1.16
3.61/2.2
3.77/1.78
NI
5.31/2.33
9.50/2.77
8.65/1.94
NI
7.84/2.17
NI
8.60/2.18
NI
NI
25.9/2.05
9.81/2.51
NI
16.1/2.26
9.64/2.41
9.41/2.23
12.1/2.30
9.82/2.30
9.32/2.17
9.23/2.32
8.02/0.81
7.17/1.49
7.51/1.16
3.92/2.22
4.07/1.80
4.27/3.14
4.70/2.64
9.6/3.05
5.87/2.20
NI
6.40/2.46
10.8/3.15
8.54/2.42
NI
NI
6.35/2.76
9.58/2.72
NI
13.4/2.66
8.72/2.5
9.65/2.58
10.6/2.60
9.02/2.57
8.90/2.28
9.66/2.35
9.40/0.65
8.16/1.42
8.30/1.27
5.19/2.2
5.9/1.80
4.63/3.17
4.86/2.54
8.80/3.01
5.21/2.23
NI
7.25/2.82
8.0/3.31
5.00/2.55
NI
NI
7.91/3.13
6.72/3.16
NI
9.50/3.08
8.50/3.04
8.30/3.00
8.31/2.76
9.05/3.04
8.97/2.58
8.00/2.51
5.03/3.48
5.37/2.86
9.0/3.40
6.05/2.36
6.08/3.40
4.25/3.05
9.00/3.41
4.90/2.84
NI
2.50/3.95
6.51/3.73
7.32/3.48
8.6/3.8
9.13/3.38
5.90/3.16
8.04/3.00
8.86/3.30
9.50/3.10
11.9/2.66
9.50/2.54
122
5.2.7
Observation of a Unique Phasing Phenomenon
An interesting phenomenon was observed during start up of the iso-octane
experiment at φ = 0.57 and Tin = 450 K. As shown in Figure 5-39 (a) - (d), the cycle to
cycle cylinder pressure varies in a unique, reproducible pattern.
In this case, 200
continuous cycles were recorded. Cycle 1 is defined as the first cycle that pressure
changes were observed after the fuel was injected into the inlet port. At the beginning of
the cycles (cycle 1 - cycle 17), very small deviations were observed in the pressure traces,
which can be associated with the gradually changing in-cylinder gas properties as the fuel
is introduced.
(a)
70
Pressure (BAR)
21
20
20
15
19
10
18
5
0
320
23
17
340
360
380
(b)
60
22
25
Pressure (Bar)
30
400
50
20
10
340
50
128
12
9
40
30
60
127
126
125
124
20
10
0
320
340
360
380
400
Crank Angle (CAD)
420
70
(c)
Pressure (BAR)
Pressure (BAR)
60
83
86
87
30
Crank Angle (CAD)
70
84
40
0
320
420
85
360
380
Crank Angle (CAD)
400
420
50
40
30
(d)
146
145
144
143
20
10
0
320
340
360
380
400
Crank Angle (CAD)
Figure 5–39. Examples of cylinder pressure “phasing” during iso-octane
start up at Tin = 450 K and φ = 0.57
420
123
As shown in Figure 5-39 (a), a multiple-cycle ignition sequence began, starting
with cycle 18. Due to low in-cylinder temperature, fuel oxidation was slow and only a
small amount of heat was produced. The temperature increase and residuals left from
cycle 18 resulted in more heat release for cycle 19. However, due to the low average
cylinder temperature, hot ignition was not initiated until cycle 22. If cycles 18 to 22 are
called “phase 1”, then, cycle 23 began a new “phase 1” after the consumption in cycle 22
of the residuals from previous cycles. This type of “phasing” continues, with each new
sequence producing higher average in-cylinder temperature and pressure, as shown in
Figure 5-39 (b) and (c), until stable combustion is reached after almost 150 cycles,
Figure 5-39 (d).
This unique pattern of ignition and engine stabilization was not observed for isooctane with addition of 0.5% DTBP under these same experimental conditions. With the
DTBP addition, the combustion stabilized rapidly within a couple of cycles. However, a
similar unique pattern was observed at iso-octane + 1.5% DTBP at a lower equivalence
ratio (φ = 0.49) and at a lower inlet temperature of 410 K, which also represents a
limiting case for ignition (see Table 5-3).
5.3
Discussion
As noted, autoignition plays a critical role in HCCI engines and autoignition is in
turn dominated by the fuel’s low and intermediate temperature chemistry. The structure
of the low and intermediate temperature kinetic mechanism is based on degenerate chain
124
branching which is illustrated in Fig. 4-5 and described in detail in Appendix A. For
convenience, highlights are provided here.
For instance, the low temperature region (< 650 K) is characterized by the
reactions of RO2• radicals for smaller hydrocarbons or QOOH• radicals for larger
hydrocarbons,
and
by
the
formation
of
stable
oxygenated
hydrocarbons.
R• + O2 Ù RO2• is the primary mechanism necessary for simulating Negative
Temperature Coefficient (NTC) behavior, and the isomerization reaction of
RO2• Ù QOOH• determines the extent of the preignition reaction. The intermediate
temperature region (650 – 900+ K) is dominated by the reactions of HO2• radicals.
Another key reaction is HOOH + M = 2OH• + M. This reaction controls the transition
from the NTC region to the second stage ignition.
Most observed preignition and
autoignition behavior, including single and multiple cool flames, can be explained in
terms of this generalized mechanism.
As mentioned in Chapter 2, n-heptane and iso-octane have quite different
oxidation chemistries. n-Heptane auto-ignition occurs in two stages, while iso-octane
auto-ignition happens in a single stage at higher temperature. The earlier ignition timing
of n-heptane and low octane number fuels is due to the faster low and intermediate
temperature reactions of n-heptane. In other words, if the fuel is all iso-octane (PRF100),
the amount of heat released by exothermic reactions is very small in the low temperature
region. n-Heptane and iso-octane interact through a radical pool of R• and RO2• in the
low temperature region and R•, OH•, and HO2• in the intermediate temperature region.
The iso-octane in PRFs acts as a radical scavenger over the entire low and intermediate
125
temperature range, which can explain the observed narrower operating range for higher
octane number fuels.
Methyl radical, the primary product of decomposition of DTBP, undergoes rapid
oxidation even at temperatures below 900 K to yield the molecular products
formaldehyde, methanol and hydrogen peroxide. The interactions of these products and
fuel may lead to greater initial exothermicity, thus having a thermal effect on ignition.
However, from studies described in Chapter 4, it is evident that DTBP has the ability to
have a chemical effect on ignition. The fuel parameter that determines which effect will
dominate a given scenario is octane number.
From relevant work on the addition of formaldehyde and methanol to SI Primary
reference fuels, it was shown that they have the ability to decrease the reactivity of SI
primary reference fuels leading to an increase in ignition delay [Kuwahara et al., 2004].
Formaldehyde acts as an OH• radical scavenger in the low temperature regime via the
reaction CH2O + OH• = HCO• + H2O. Since DTBP decomposes to form formaldehyde,
if there were a dominant chemical effect, then DTBP should reduce the overall reactivity
and cause a delay in ignition timing. In the present study DTBP shows strong evidence
of being an ignition promoter. This suggests that for lower octane number fuels, the
thermal effect of the DTBP seems to be what is driving the reduction in ignition delay
time.
Furthermore, the hydrogen peroxide (H2O2), formed during the secondary
oxidation of methyl radicals, would be expected to be a very effective accelerant because
of its decomposition to form two reactive OH• radicals. However, the bond dissociation
energy of H2O2 is high (52 kcal/mole, 217 kJ/mole) corresponding to an activation
126
temperature of 2615 K. This means that H2O2 doesn’t decompose in the compression
stroke until the temperature reaches around 900 K. This may explain why stable
combustion could not be reached for some cases of iso-octane at inlet temperature of
410 K.
Results from Chapter 4 indicated that the effect of DTBP on iso-octane is by
direct chain initiation of hydrocarbon oxidation via the free radicals generated from
DTBP. In the higher ON blends and with DTBP addition to iso-octane, the radicals from
either DTBP or n-heptane reactions increase reactivity. However, in the lower ON PRF
blends, the presence of additional radicals from the decomposition of DTBP does not
have a significant impact on its overall reactivity and the effect is primarily thermal.
Final resolution of the issue of DTBPs mode of operation awaits the definitive set
of experiments, i.e., intake temperature sweeps for selected ON PRF blends with and
without DTBP in an engine to map the autoignition behavior. However, at this time
evaluation of the literature suggests the modes stated above.
5.4
Closure
In this chapter, combustion characteristics of the SI primary reference fuels and
their blends with and without the addition of DTBP in a CFR engine have been reported.
As expected, low octane number fuels have shorter ignition delay times and wider
operating ranges, with n-heptane having the earliest ignition timing and iso-octane the
latest. Also, at the tested compression ratio and engine speed, the IMEP that can be
127
obtained for low RON fuels is small; lower compression ratio or higher engine speed is
required for these fuels to obtain higher IMEP.
Experimental results show that ignition delay time, cycle to cycle variation, and
stable operating range were all improved with the addition of less than 2.5% DTBP by
volume. For example, the addition of DTBP had the following effects: ignition delay
time reduction by at least 3 CAD for all tested fuels; COVIMEP improvement to <10% (a
37.5% reduction) for PRF92 at inlet temperature of 450 K and equivalence ratio of 0.49;
and extension of stable HCCI operations for relatively high RON fuels to a broader
equivalence ratio range and to lower inlet temperatures (e.g., 2.5% DTBP by volume in
iso-octane, extended stable operation to an equivalence ratio of 0.39 at inlet temperature
410 K). These results indicated that DTBP can successfully alter the ignition properties
of PRFs and their blends, thus improving their HCCI combustion characteristics
significantly.
With addition of DTBP:
(1) The HCCI operating region can be extended. In other words, HCCI operation can
be realized over a fairly wide equivalence ratio and low inlet temperature for
relatively high RON fuels, like iso-octane.
(2) Ignition timing is advanced, and the ignition delay time decreases with DTBP
concentration increase.
(3) Cycle to cycle variation is improved for higher RON fuels, while low RON fuels
did not exhibit the same reduction, mainly due to the reduced IMEP from
advanced ignition timing.
128
(4) Possible explanations for the mode of action of DTBP have been proposed. For
high ON PRF blends the effect of DTBP is primarily chemical, while for low ON
PRF blends the effect of DTBP is primarily thermal. More detailed experiments
need to be run in order to be clear on the effect of DTBP on lower ON fuels.
A reproducible multi-cycle ignition phenomenon was observed for iso-octane at
Tin = 450 K and φ = 0.57 and for iso-octane + 1.5% DTBP at an inlet temperature of
410 K and equivalence ratio of 0.57, where the cycle to cycle cylinder pressure varies in a
unique, reproducible pattern during the startup process. In both cases, the experimental
conditions represent limiting cases for ignition. This “phasing” behavior is consistent
with partial oxidation and the carry over of partial oxidation products enhancing the
preignition chemistry in the next cycle.
129
CHAPTER 6.
DEVELOPMENT OF A SKELETAL KINETIC MODEL FOR
PREDICTION OF PREIGNITION REACTIVITY OF PRFS*
Understanding the ignition and oxidation chemistry of typical fuels is extremely
important in homogenous charge compression ignition (HCCI) engine operation. A
model that correctly simulates fuel oxidation at HCCI conditions would be a useful
design tool. This need has motivated the current effort to update an existing skeletal
kinetic model for simulation of the autoignition behavior of SI primary reference fuels
and their mixtures. In this chapter, the effort to reformulate these skeletal models to be
compatible with the standard CHEMKIN package is also reported.
6.1
Introduction
As described in Chapter 2, there are five kinds of chemical kinetic models,
detailed, lumped, reduced, skeletal and global [Zheng et al., 2004]. Detailed models [e.g.,
Curran et al., 1998, 2002] try to include all of the important elementary reactions and
individual species using the best available rate parameters and thermochemical data.
However, there are uncertainties in the selection of reactions and rate parameters and
detailed models are often developed for a single hydrocarbon and only validated over a
rather limited range of conditions. Nonetheless, detailed models remain the ultimate goal.
As a practical matter, however, until computers and algorithms get more efficient
there is a place for smaller mechanisms. Lumped mechanisms have evolved as a method
* The material in this chapter was the basis for Paper No. 18587, presented at the 37th
ACS Middle Atlantic Regional Meeting (MARM), Rutgers University, NJ, May 2225, 2005 [Gong et al., 2005c].
130
of reducing the overall size and complexity of mechanisms. The size of a lumped
mechanism can vary significantly, but usually encompass thousands of reactions among
hundreds of species. Detailed models can be culled to produce a third type of model, the
reduced model. These models contain the most critical elements of the full mechanism.
A fourth form of model is the skeletal model that consists of a sequence of composite
kinetic steps representing the reaction progress. These kinetic steps can be elementary,
generic, or global reactions. Rate parameters and thermochemistry are based on the best
information but represent “classes” of reactions. Global models describe the chemistry in
terms of a few of the principal reactants and products in one or more overall functional
relations.
Studies to develop reliable chemical kinetic models for autoignition have been
conducted in our laboratory for several years. In order to interpret our data we have
developed models that range from detailed chemical kinetic schemes, to skeletal
mechanisms, to a simple seven-step global scheme.
Our previous skeletal reaction model has 29 reactions and 20 active species [Li et
al., 1992, 1996]. The characteristics of this model are that small species oxidation was
considered and a formation path for CO was provided. This model, predicted the ignition
delay and the pre-ignition heat release for these fuels to within 15%. The model was
modified to reflect the oxidation chemistries of butanes [Wang et al., 1996 a and b]. The
results indicate that this reduced model can be applied to predict the preignition reactivity
of butanes.
The model was further developed and successfully used for predicting HCCI preignition behavior including temperature, pressure, ignition delay and heat release for
131
PRF20 and PRF50 [Zheng et al., 2001]. More recently, this model was extended to
incorporate low, intermediate, and high temperature chemistry [Zheng et al. 2002]. This
skeletal model (69 reactions and 45 species in this case) was shown to be a useful tool to
study HCCI engine operation.
However, all these models were based on our in-house programs and tested
against selected engine conditions.
For flexibility, portability and to use existing
sensitivity analysis tools, it is desirable to reformulate these models to be compatible
with the standard simulation packages, such as CHEMKIN.
The data used in this effort were generated in the Drexel Pressurized Flow
Reactor (PFR) facility, as described earlier in Chapters 3 and 4. Results of five test fuels
chosen from Chapter 4, namely, n-heptane, iso-octane and PRF20, PRF63 and PRF92
were examined in this effort; detailed experimental conditions are listed in Table 6-1.
Table 6-1. Pressurized flow reactor test conditions
A
B
C
D
E
Reactant Percentage (V/V liquid, %) Oxidizer Comp. (%)
Air
n-Heptane iso-Octane
N2
n-Heptane
100
0
85
15
PRF20
80
20
85
15
PRF63
37
63
70
30
PRF92
8
92
65
35
iso-Octane
0
100
62
38
ON
φ
0.4
0
0.4
20
0.5
63
0.6
92
0.75 100
Reaction Time
(ms)
100
100
200
225
250
132
6.2
Skeletal Modeling Methods
Skeletal kinetic models, based on degenerate-branched-chain and class chemistry
concepts, were developed in the 1970s for prediction of autoignition delay time [Halstead
et al., 1975], and this work formed the basis for later developments [Cox and Cole, 1985;
Hu and Keck, 1987; Li et al., 1996]. These skeletal models follow a chemical framework
suggested by Benson [1981, 1982], which is essentially represented by R1-R17 and
species 1-14 in the model of Li et al. [1996] as shown in Table 6-2. The low temperature
and negative temperature coefficient behavior is represented by R1-R8, and transition to
the second stage, hot ignition is controlled by R9.
In the model of Li et al. [1996] the oxidation of smaller allyl radicals (Rs•) was
added to increase heat release without forcing complete consumption of the fuel;
Rs• + O2 Ù RsO2•,
(R19)
RsO2• Ù C=C + HO2•,
(R20)
RsO2• + RH (or RCHO) => RsOOH + R• (or RCO•), and
(R22)
RsOOH => RsO• + OH•.
(R23)
133
Table 6-2. Skeletal chemical kinetics model of Li et al. [1996]
A.
20 Active Species
1. RH
6. OOQOOH•
11. OQ’OOH•
16. RsO2•
B.
3. R •
8. OH•
13. C=C
18. RsO•
2. O2
7. OQO•
12. RCHO
17. RsOOH
4. RO2 •
9. HO2•
14. RCO•
19. RO•
5. QOOH•
10. HOOH
15. Rs•
20. ROOH
-E/RT
29 Reactions (units: mole, s, kcal) Arrhenius parameters of rate constants k=Ae
+
Reaction
∆H°300
1. RH+O2 Ù R•+HO2•
2. R•+O2 Ù RO2•
3. RO2• Ù QOOH•
n-heptane
iso-octane
4. QOOH•+O2 <=> QOOHOO•
5. OOQOOH• => OQ’OOH•+OH•
6. OH•+RH => H2O+R•
7. OQ’OOH• => OQ’O•+OH•
8. HO2•+HO2• => HOOH+O2
9. HOOH+M => 2OH•+M
10. OQ’O• => 2RCHO+RCO•
n-heptane
OQ’O• => 2RCHO+Rs•
iso-octane
11. QOOH• =>
C=C+RCHO+OH•
12. RO2•+RCHO =>
ROOH+RCO•
13. HO2•+RCHO =>
HOOH+RCO•
14. C=C+HO2• =>Epox+OH•
15. HO2•+RH Ù R•+HOOH
16. RO2•+RH Ù ROOH+R•
17. RCHO+OH• =>RCO•+H2O
n-heptane
iso-octane
18. RCO•+M => Rs•+CO+M
19. Rs•+O2 Ù RsO2•
20. RsO2• => C=C+OH•
21. RCHO+RsO2•
=>RsOOH+RCO•
22. RH+RsO2• Ù RsOOH+R•
23. RsOOH => RsO•+OH•
24. RsO•+O2 => Rs’O+HO2•
25. C=C+OH• => 2OXY+OH•
46.4
–30.1
26. ROOH = RO•+OH•
27. RO• => Rs•+RCHO
28. RO2• => C=C+HO2•
29. RO2• => ether+OH•
n-heptane
iso-octane
k-
k
Equilibrium
46.0
0.0
11.9
11.0
11.5
11.3
13.3
15.6
12.3
16.88
19.0
22.4
0.0
17.0
3.0
40.0
0.0
46.0
–17.5
14.0
15.0
18.5
-3.0
14.0
14.4
15.0
31.0
-0.6
11.45
8.6
-0.6
11.7
8.64
10.95
11.7
11.2
10.0
16.0
16.0
13.22
13.57
16.78
12.0
11.75
0.0
0.0
15.0
0.0
28.9
11.53
8.6
11.28
15.6
10.6
12.72
–0.23
8.0
8.0
-31.5
-31.5
10.7
-31.0
17.5
46.0
-27.4
0.9
0.0
-1.9
8.0
11.24
-27.4
0.9
1.1
-1.4
E
E+
log
+
A
13.5
12.0
7.50
7.50
-30.1
–26.6
-23.5
43.6
-38.5
51.4
Log
A
1.5
-1.4
8.0
8.0
-27.4
-0.6
8.0
43.6
-26.5
–75.5
1.18
8.0
43.6
-10.0
4.0
15.6
13.3
9.85
16.0
43.0
2.14
1.04
43.0
15.0
23.0
-25.0
-25.0
9.48
8.78
18.0
18.0
E-
log
A
12.0
13.4
0.0
27.4
11.0
11.0
13.4
11.0
11.0
27.4
10.8
10.1
8.0
8.0
13.4
27.4
10.1
8.0
134
This improvement allowed the prediction of ignition delay times and the
preignition heat release for these fuels to within 15%. This model also considered the
chemical path for CO production, which allowed proper prediction of CO concentration.
RCO• + M => Rs• + CO,
(R18)
This extended skeletal model has been successfully applied to PRF87 and PRF63
to predict the ignition delay and the pre-ignition heat release for these fuels to within 15%.
The model was further extended to incorporate low, intermediate, and high
temperature chemistry by Zheng et al. [2002]. This skeletal model (69 reactions and 45
species in this case) was shown to be a useful tool to study HCCI engine operation. The
reactions that are important in low and intermediate temperature regimes are shown in
Table 6-3.
It is essential that the correct numbers of fuel C and H atoms are carried through
to the final products of combustion. As written in Table 6-2 the pre-ignition reaction
OQ’O• => 2RCHO + Rs•
does not conserve atoms; RCHO and Rs• are considered to represent a class of surrogate
species. Measurements show that HCHO and C3H7CHO are primary oxygenates and that
C2H3 and C3H5 are primary small hydrocarbons. Therefore, for use in the final model of
Zheng et al. [2002], this reaction was rewritten as
135
OQ’O• => HCHO + C3H7CHO + mC2H3 + nC3H5
(R10)
which conserves atoms.
The values of m and n in the reaction depend on the number of carbons and
hydrogens per molecule of fuel. Specifically, for mixtures of the primary SI reference
fuels, m and n can be related to the pump octane number (PON) as follows:
m = 1 - PON/100 and n = PON/100.
136
Table 6-3. Skeletal model for low temperature, NTC and intermediate temperature
regions by Zheng et al. [2002]
Reaction
∆H
0
300
+
LogA
E
LogA
E
+
-
LogA
E
-
1. RH+O2<=>R•+HO2•
46.4
1.5
46.0
13.5
46.0
12.0
0.0
2. R•+O2<=>RO2•
-31.0
-1.4
-27.4
12.0
0.0
13.4
27.4
7.5
7.5
0.9
0.63
8.0
8.0
11.9
11.63
19.0
19.0
11.0
11.0
11.0
11.0
7.5
0.0
11.24
11.0
22.4
11.0
11.0
-31.0
-1.9
-27.4
11.5
0.0
13.4
27.4
-26.6
11.3
17.0
3. RO2•<=>QOOH•
n-heptane
20 PRF
iso-octane
4. QOOH•+O2 <=>OOQOOH•
5. OOQOOH•=>
OQ′OOH•+OH•
6. OH•+RH=>H2O+R•
-23.8
13.3
3.0
7. OQ′OOH•=>OQ′O•+OH•
43.6
15.6
40.0
8. HO2•+ HO2•=>HOOH+O2
-38.48
12.3
0.0
9. HOOH+M=> 2OH•+M
51.23
17.08
45.5
**
14.0
15.0
**
14.4
31.0
-0.6
11.45
8.6
-0.6
11.7
8.64
-20.28
10.95
10.0
-25.29
10.95
10.0
10. OQ′O•=>HCHO
+R*CHO+mC2H3+nC3H5
11. QOOH•=>R*CHO
+OH•+mC3H6+nC4H8
12. RO2•+R*CHO=>
ROOH+R*CO•
13. HO2•+R*CHO=>
HOOH+R*CO•
14. C3H6+HO2•=>
C3H6O+OH•
15. C4H8+HO2•=>
C4H8O+OH•
16. HO2•+RH<=>R•+HOOH
17. RO2•+RH<=>ROOH+R•
7.92
0.9
8.0
11.7
16.0
10.8
8.0
7.92
1.1
8.0
11.2
16.0
10.1
8.0
13.29
0.0
16.78
15.0
13.4
27.4
10.1
8.0
18. R*CHO+OH•=>
R*CO•+H2O
19. R*CO•+M=>R*+CO•+M
12.09
20. R*•+O2<=>R*O2•
-31.0
21. R*O2•=>C3H6+HO2•
22. R*CHO+R*O2•=>
R*OOH+R*CO•
23. RH+ R*O2<=>R*OOH+R•
24. R*OOH=>R*O•+OH•
25. R*O•+ O2=>R*′O+HO2•
26. C7H14+OH•=>R*CHO+
C2H5CHO+OH•
27. C8H16+OH•=>2C3H7CHO
+OH•
28. ROOH=> RO•+OH•
29. RO•=>m R*•+nC4H9+
R*CHO
30. RO2•=>mC7H14+nC8H16+
HO2•
31. RO2•=>ether+ OH•
Note: R*=C3H7 and R*’=C3H6.
-32.3
12.0
0.0
14.92
11.75
28.9
-0.6
11.53
8.6
11.28
16.0
7.92
-1.4
1.18
-27.4
8.0
43.6
15.6
43.0
-27.12
10.6
2.14
-79.03
12.72
-1.04
-59.4
12.75
-1.04
43.6
15.6
43.0
**
13.3
15.0
**
9.85
23.0
-25.0
9.34
18.0
137
6.3
Current Model Development
As noted, the models of Li et al. [1996] and Zheng et al. [2002] were used with
our in-house kinetic programs, and the current model development focused on
reformulating these models to be compatible with the CHEMKIN software package.
In order to have each reaction conserve atoms, it was necessary to add 15 species
to the existing model. Most of these additions replace a single generic species with
several specific species. For example, C=C, representing the alkenes in the Li et al.
[1996] model, is now replaced by 4 different molecules C”C-3, C”C-4 and C”C-7 and
C”C-8. A complete list of the final active species is provided in Table 6-4.
Table 6-4. Active species of current model
1. RH
7. OOQOOH•
13. HO2•
19. RsCO•
25. Rs•
31. Rs'CHO
2. R•
8. Q'OOH
14. CO
20. RsO•
26. Epox
32. Rs'CO•
3. RO2•
9. HOOH
15. CO2
21. RsO2•
27. RO•
33. Rs"O•
4. ROOH
10. N2
16. C"C-3
22. RsOOH
28. ETHE
34. RO•
5. O2
11. OH•
17. Rs•
23. Rs"O•
29. C"C-4
35. Rs" •
6. QOOH•
12. H2O
18. RsCHO
24. C"C-8
30. C"C-7
As noted, Table 6-4 lists the active species in the current model, and Tables 6-5
and 6-6 present the associated skeletal mechanism and the recommended fuel specific
rate parameters. In the current model only the rate parameters of three reactions (R3 and
R7 and R27) are adjusted to account for variation in the fuel.
138
Table 6-5. Current skeletal model
Arrhenius parameters of rate constants k=Ae-E/RT (units: mole, s, kcal)
k+
Reaction
1. RH+O2 Ù R•+HO2•
2. R•+O2 Ù RO2•
3. RO2• Ù QOOH•
n-heptane
PRF20
PRF63
PRF92
iso-octane
4. QOOH•+O2 <=> QOOHOO•
5. OOQOOH• => OQ’OOH•+ OH•
6. OH•+RH => H2O+R•
7. OQ'OOH => RsCHO+Rs'CO•+OH•
n-heptane
PRF20
OQ'OOH => RsCHO+RsCO•+OH•
PRF63
PRF92
iso-octane
8. HO2•+HO2• => HOOH+O2
9. HOOH+M => 2OH•+M
10. QOOH• => C"C-4 (C"C-3)+RsCHO+OH•
11. RO2•+RsCHO => ROOH+RsCO•
12. HO2•+RsCHO => HOOH+RsCO•
13. C"C-4 (C"C-3)+HO2• => Epox+OH•
14. HO2•+RH => R•+HOOH
15 RO2•.+RH => ROOH+R•
16. RsCHO+OH• => RsCO•+H2O
17. RsCO• (Rs'CO•)+M => Rs• (Rs'•)+CO+M
18. Rs•+O2 Ù RsO2•
19. RsO2• => C"C-3+HO2•
20. RsCHO+RsO•=> RsOOH+RsCO
21. RH+RsO2•=> RsOOH+R•
22. RsOOH => RsO•+OH•
23. RsO•+O2 => Rs"O•+HO2•
24. C"C-8+OH•+O2 => 2RsCHO+OH•
25. ROOH => RO•+OH•
26. RO• => Rs'•+RsCHO
27. RO2• => (C"C-7)+HO2•
n-heptane
PRF20
RO2•=> C"C-8 +HO2•
PRF63
PRF92
iso-octane
28. RO2• => ETHE+OH•
k-
A+
3.16E+13
1.00E+12
E+
46.0
0.0
A1.0E+12
2.51E+13
E0.0
27.4
9.80E+11
6.10E+11
1.58E+11
6.00E+10
5.74E+10
3.16E+11
2.00E+11
2.05E+13
18.8
18.9
19.2
20
20.5
0.0
17.0
3.0
1.10E+11
1.10E+11
1.10E+11
1.10E+11
1.10E+11
2.51E+13
11.0
11.0
11.0
11.0
11.0
27.4
9.85E+16
8.10E+16
47.8
47.2
1.62E+16
4.75E+15
6.55E+15
2.00E+12
7.60E+16
2.52E+14
2.82E+11
5.01E+11
8.91E+10
5.01E+11
1.58E+11
1.37E+13
6.05E+16
1.00E+12
2.20E+11
3.39E+11
1.90E+11
3.98E+15
3.98E+10
5.25E+12
3.98E+15
2.00E+13
44.5
42.2
40.8
0.0
46.0
31.0
8.6
8.64
10.0
16.0
16.0
6.31E+10
1.26E+10
8.0
8.0
15.0
0.0
28.9
8.6
16.0
43.0
2.14
-1.04
43.0
15.0
2.50E+13
27.4
1.26E+10
8.0
3.30E+10
2.58E+10
22.20
22.20
1.88E+10
2.15E+10
3.18E+10
3.01E+09
22.2
23.0
23.0
18.0
139
Table 6-6. Key fuel specific reaction parameters in current skeletal model
n-heptane
Reaction
PRF20
PRF63
PRF92
iso-octane
A
E
A
E
A
E
A
E
A
E
9.80E+11
18800
6.10E+11
18900
1.58E+11
19200
6.00E+10
20000
5.74E+10
20500
3
1.10E+11
11000
1.10E+11
11000
1.10E+11
11000
1.10E+11
11000
1.10E+11
11000
7
9.85E+16
47800
8.10E+16
47200
1.62E+16
44500
4.75E+15
42200
6.55E+15
40800
3.30E+10
22200
2.58E+10
22200
1.88E+10
22200
2.15E+10
23000
3.18E+10
23000
27
k = A T**b exp (-E/RT)
(A units mole-cm-sec-K, E units cal/mole)
Most chemical reaction rate parameters in the current model (i.e., the preexponential factor, A, and the activation energy, E) are from published data [Li et al.,
1996; Zheng et al., 2002], except for reactions R3, R7 and R27. The thermodynamic
properties of different species were chosen from detailed models of Curran et al. [1998,
2002]. For each species in this skeletal model, we found a species in the detailed model
that is closest to it. For example, RH thermodynamic properties are chosen from nC7H16
for n-heptane and PRF20, iC8H18 are used for PRF63, PRF92 and iso-octane. For small
molecules, such as ETHE, thermo properties were chosen based on the molecular
structure and the carbon number necessary to preserve.
Following the initial H atom abstraction from the fuel ((R1) in Table 6-5),
molecular oxygen addition to the alkyl radical takes place:
R• + O2 Ù RO2•.
(R2)
140
The reverse of the oxygen addition to the alkyl radical reaction becomes more
important at higher temperature resulting in more stable C = C and HO2• being formed
such that branching is retarded. This is the primary mechanism necessary for simulating
Negative Temperature Coefficient (NTC) behavior.
RO2• Ù QOOH•
(R3)
For larger hydrocarbons, the QOOH•, isomerization product of RO2•, is also very
important for reproducing the temperature dependence of the NTC behavior.
The
chemical kinetic parameters of this reaction were adjusted slightly for the test fuels to
reflect the different octane numbers. Specially, activation energies of 18000, 18900,
19200, 20000 and 20500 cal were chosen for the five different fuels in order of increasing
octane number. The values of the forward pre-exponential constant A were adjusted to
improve agreement with the CO mole fraction. As shown in Table 6-6, A was varied
from 9.80E+11 to 5.74E+10. Such a variation is reasonable because of differences in the
pool of C1 to C3 species that scavenge the active radicals, OH• and HO2•, and affect the
pre-ignition behavior.
In previous skeletal models, carbonylhydroperoxide OQ’OOH• first decomposes
to OQ’O• and OH, then OQ’O• continues to decompose to form oxygenated radical
species:
OQ’O• => RsCHO + Rs’CO•
141
Since OQ’O• appears in these skeletal models only as an intermediary product, without
inducing any branching, we eliminated the species OQ’O• and associated reactions to
allow OQ’OOH• to form oxygenated radical species and OH• directly. This direct
decomposition of carbonylhydroperoxide also has been reported in detailed mechanisms
of n-heptane and iso-octane [Curran et al., 1998, 2002]. In both of these mechanisms,
reaction parameters were chosen as A = 1.50E+16, b = 0.00, E = 4.160E+04. In the
current model, the chemical kinetic parameters of this reaction were adjusted slightly for
reference fuels and blends with different octane numbers.
Also, Rs and Rs’ here were chosen as C3H7 and C2H5 to ensure that the correct
numbers of fuel C and H atoms are carried through to the final products of combustion.
Measurements show that HCHO and C3H7CHO are the primary oxygenates. Therefore,
for use in the final model the reaction (R7) is written as
OQ’OOH• => RsCHO + Rs’CO• + OH• (R7) (for n-heptane and PRF20)
and
OQ’OOH• => RsCHO + RsCO•+ OH• (R7) (for PRF63, PRF92 and iso-octane)
This
newly
introduced
reaction
(R7),
the
direct
decomposition
of
carbonylhydroperoxide, plays a very important role in shifting the NTC regions. The
activation energies were chosen as 47800, 47200, 44500, 42200 and 40800 cal for
n-heptane, PRF20, PRF63, PRF92 and iso-octane, respectively. This in turn causes the
temperature for peak CO mole fraction to shift from 696 K for n-heptane to 662 K for
iso-octane, and brings the predictions in good agreement with the experimental data.
142
The intermediate temperature region (650 - 800 K) is dominated by the reactions
of HO2• radicals RO2• => C=C + HO2•. This reaction and the following reaction HOOH
+ M = 2OH• + M control the transition from NTC to the intermediate temperature region.
Rate parameters of R27 were slightly changed for different fuels as shown in Table 6-6.
6.4
Experimental Results
Reactivity maps for n-heptane, iso-octane and three blends at the conditions listed in
Table 6-1 are shown in Figure 6-1.
1400
n-heptane
PRF92
1200
PRF20
iso-octane
PRF63
[CO] (ppm)
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 6–1. Reactivity maps for n-heptane, PRF20, PRF63, PRF92 and isooctane from CCD experiments in a PFR
For all 5 conditions, the maps exhibit typical negative temperature coefficient
behavior. As expected, n-heptane shows significantly more reactivity than iso-octane.
The starting temperatures of NTC range from 705 K for n-heptane to 665 K for iso-
143
octane. In general, the temperature for peak CO concentration is lowered as the ON of
the reactants increases.
It can also be seen that reactivity occurs over a narrower
temperature range as the ON increases. n-Heptane has the widest reactivity span, 625 to
775 K, while the iso-octane has the narrowest, 630 to 680 K. Figure 6-1 also shows that
for blends with even small amounts of n-heptane, e.g., PRF92, the reactivity is much
higher than for neat iso-octane.
This is due to the faster low and intermediate
temperature reactions of n-heptane.
6.5
Model Validation
The plug flow application from Chemkin 3.7.1 was used to perform the
calculations. For all experiments, an adiabatic condition was assumed along the length of
the flow reactor.
To model the CCD experiments, a series of calculations were
preformed for inlet temperatures from 600 – 800 K at 5 °C increments. The species,
including CO and CO2 were generated with a resolution of 0.5 cm. Calculations were
carried out in the geometry shown in Figure 6-2.
D = 2.2cm
L = 40cm
Figure 6–2. The plug flow reactor geometry for CHEMKIN calculations
144
Comparisons of the experimental data with detailed models [Curran et. al., 1998,
2002] and skeletal model predictions are provided in Figures 6-3 to Figure 6-7. In
general, the skeletal model successfully predicted the reactivity behaviors in the
600-800 K regions for all five fuels.
The detailed model only agrees with the
experimental data relatively well for low octane number fuels, i.e., n-heptane, PRF20 and
PRF63.
For PRF92 and iso-octane, the detailed model predicts much higher CO
concentrations than what is observed experimentally and predicted by the skeletal model.
Compared to the detailed model, this modified skeletal model reduced the CPU time by
almost 3 orders of magnitude.
Experiment
Detailed Model
Skeletal Model
1600
1400
[CO] (ppm)
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 6–3. Comparison of n-heptane reactivity measured experimentally
and predicated using detailed and skeletal models
145
Experiment
Detailed Model
Skeletal Model
1600
1400
[CO] (ppm)
1200
1000
800
600
400
200
0
600
625
650
675
700
725
Temperature (K)
750
775
800
Figure 6–4. Comparison of PRF20 reactivity measured experimentally
and predicated using detailed and skeletal models
Experiment
Detailed Model
Skeletal Model
1000
900
[CO] (ppm)
800
700
600
500
400
300
200
100
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 6–5. Comparison of PRF63 reactivity measured experimentally and
predicated using detailed and skeletal models
146
Experiment
Detailed Model
Skeletal Model
1800
1600
[CO] (ppm)
1400
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 6–6. Comparison of PRF92 reactivity measured experimentally and
predicated using detailed and skeletal models
Experiment
Detailed Model
Skeletal Model
2000
1800
1600
[CO] (ppm)
1400
1200
1000
800
600
400
200
0
600
625
650
675
700
725
750
775
800
Temperature (K)
Figure 6–7. Comparison of iso-octane reactivity measured experimentally
and predicated using detailed and skeletal models
147
6.6
Closure
A skeletal chemical kinetic model for the SI reference fuels (PRFs) and their
blends has been developed and tested against data from a Pressurized Flow Reactor. The
model was developed as an extension of our previous preignition model by modifying
several reactions to incorporate recent advances in our understanding of the relevant
chemistry.
The model was also reformulated to be compatible with the standard
CHEMKIN simulation package. Key features of the model include provision for element
conservation, choosing thermodynamic properties for relevant species, and adoption of
the CHEMKIN package. The current model consists of 28 reactions and 35 species.
n-Heptane, iso-octane and three of their mixtures corresponding to PRF20, PRF63
and PRF92 were examined. The reaction rate parameters for the modified model were
selected initially as those used in our previous work or based upon similar reactions in the
case of the new reactions. The rate parameters were then “tuned” using the PRF data
from the flow reactor. These “tuned” reaction rate parameters included only three fuelsensitive reaction rates, which were correlated to the octane number of the specific
hydrocarbon mixture.
The model was able to satisfactorily reproduce the negative
temperature coefficient region and general reactivity behavior observed in the PFR, as
well as the measured CO species evolution profiles. Compared to the detailed model, this
modified skeletal model reduced the CPU time by almost 3 orders of magnitude.
148
CHAPTER 7.
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
This study investigated the oxidation chemistry of gasoline primary reference
fuels and their mixtures, including the effects of Di-tertiary Butyl Peroxide (DTBP)
addition. The overall objective was to improve our understanding of the hydrocarbon
oxidation process, particularly at low and intermediate temperatures, to elucidate the
mode of action of DTBP and to provide insight on HCCI combustion control using fuel
additives. The effort involved both experimental and numerical modeling work.
The work carried out for this dissertation consisted of three major efforts: (1)
characterization of the low and intermediate temperature behavior of the selected fuels in
a pressurized flow reactor; (2) characterization of the ignition and combustion processes
for these fuels in an HCCI engine; and (3) development of a skeletal model for these fuels
based on the PFR data.
The results and conclusions, as well as recommendations for further work are
summarized in this chapter.
7.1
Results and Conclusions
1. Characterization of low and intermediate temperature behavior in a
pressurized flow reactor
The oxidation of the primary reference fuels for the octane number scale, iso-
octane (PRF100) and n-heptane (PRF0), and their blends, PRF20, PRF50, PRF63, PRF87
and PRF92, has been studied in a pressurized flow reactor. Experiments were run at a
pressure of 8 atmospheres over the temperature range 600 to 800 K, for equivalence
149
ratios between 0.4 and 0.75. The effects of the additive DTBP on the oxidation of these
fuels were also examined. Samples were extracted and analyzed using standard online
CO/CO2 and Total Hydrocarbon analyzers.
All of the PRF components and blends exhibit typical negative temperature
coefficient behavior, with n-heptane showing significantly more reactivity than isooctane, as expected. In PRF blends, iso-octane acts as a radical scavenger and only
contributes a small amount of exothermicity, such that the energy released at low and
intermediate temperatures is due almost entirely to reactions of n-heptane.
DTBP addition was only effective in modifying the reactivity of iso-octane; no
changes were observed in the behavior of the n-heptane or the PRF blends tested even
with higher DTBP addition.
With DTBP addition to neat iso-ocatane, there is evidence of a radical chain
initiation of the hydrocarbon oxidation process. Thus, DTBP’s effect appears to be
chemical rather than just thermal.
2. Characterization of ignition and combustion processes in an HCCI engine
The experimental results of the SI primary reference fuels and their blends in a
CFR engine have been reported, and the effects of the additive DTBP on these fuels were
also reported and discussed. As expected, low octane number fuels have shorter ignition
delay times and wider operating ranges, with n-heptane having the earliest ignition timing
and iso-octane the latest. Also, at the tested compression ratio and engine speed, the
150
IMEP that can be obtained for low RON fuels is small; lower compression ratio or higher
engine speed is required for these fuels to obtain higher IMEP.
Experimental results show that ignition delay time, cycle to cycle variation, and
stable operating range were all improved with the addition of less than 2.5% DTBP by
volume. For example, the addition of DTBP had the following effects: ignition delay
time reduction by at least 3 CAD for all tested fuels; COVIMEP improvement to <10% (a
37.5% reduction) for PRF92 at inlet temperature of 450 K and equivalence ratio of 0.49;
and extension of stable HCCI operations for relatively high RON fuels to a broader
equivalence ratio range and to lower inlet temperatures (e.g., 2.5% DTBP by volume in
iso-octane, extended stable operation to an equivalence ratio of 0.39 at inlet temperature
410 K). These results indicated that DTBP can successfully alter the ignition properties
of PRFs and their blends, thus improving their HCCI combustion characteristics
significantly.
With addition of DTBP:
(1) The HCCI operating region can be extended. In other words, HCCI operation can
be realized over a fairly wide equivalence ratio and low inlet temperature for
relatively high RON fuels, like iso-octane.
(2) Ignition timing is advanced, and the ignition delay time decreases with DTBP
concentration increase.
(3) Cycle to cycle variation is improved for higher RON fuels, while low RON fuels
did not exhibit the same reduction, mainly due to the reduced IMEP from advanced
ignition timing.
151
(4) Possible explanations for the mode of action of DTBP have been proposed. For
high ON PRF blends the effect of DTBP is primarily chemical, while for low ON
PRF blends the effect of DTBP is primarily thermal. More detailed experiments
need to be run in order to be clear on the effect of DTBP on lower ON fuels.
An interesting reproducible multi-cycle ignition phenomenon was observed for
iso-octane at Tin = 450 K and φ = 0.57 and for iso-octane + 1.5% DTBP at an inlet
temperature of 410 K and equivalence ratio of 0.49, where the cycle to cycle cylinder
pressure varies in a unique, reproducible pattern during the startup process. In both cases,
the experimental conditions represent limiting cases for ignition.
3. Development of a skeletal model based on the PFR data.
A skeletal chemical kinetic model for spark ignition primary reference fuels
(PRFs) and their blends has been developed and tested against data from a Pressurized
Flow Reactor. The model was developed as an extension of our previous preignition
model by modifying several reactions to incorporate recent advances in our
understanding of the relevant chemistry.
The model was also reformulated to be
compatible with the standard CHEMKIN simulation package. Key features of the model
include provision for element conservation, selection of species thermodynamic
properties and use with CHEMKIN package. The current model consists of 28 reactions
and 35 species. n-Heptane, iso-octane and three of their mixtures corresponding to
PRF20, PRF63 and PRF92 were examined.
The reaction rate parameters for the
modified model were selected initially as those used in our previous work or based upon
similar reactions in the case of the new reactions. The rate parameters were then “tuned”
152
using the PRF data from the flow reactor.
These “tuned” reaction rate parameters
included only three fuel-sensitive reaction rates, which were correlated to the octane
number of the specific hydrocarbon mixture.
The model was able to satisfactorily
reproduce the negative temperature coefficient region and general reactivity behavior
observed in the PFR, as well as the measured CO species evolution profiles. Compared
to the detailed model, this modified skeletal model reduced the CPU time by almost 3
orders of magnitude.
7.2
Recommendations for Future Work
Although substantial work on DTBP effects on oxidation of SI PRFs has been
completed and reported in this study, additional work is necessary in both the
experimental and modeling areas to further improve the understanding of the chemistry
of the additives and their effects. This section recommends the following extensions.
1. Further improvement of our model
The model modification in this study has been tested against data from a
pressurized flow reactor operating in the low and intermediate temperature regimes.
Although the results agree well with the experimental data, further efforts are necessary
to extend this model to a high temperature regime with the application to HCCI operation.
This extension can be based on the high temperature skeletal model of Zheng et al.
[2002].
Model development is also necessary to incorporate the decomposition
153
mechanism of DTBP to further support the experimental results and to help understand
the chemistry associated with PRFs and DTBP.
2. In-cylinder critical species measurement in HCCI conditions
It is desirable to obtain species evolution information from both the engine and
PRF experiments. Stable species that survive the sampling process can be measured with
existing GC methods. Obtaining critical radical species information, such as OH, HO2
and RO2 profiles are difficult and will require the use of advanced optical diagnostics.
Knowledge of such stable intermediated and radical species can provide detailed insight
into the reaction pathways leading to hot ignition and the data necessary for developing
and validating chemical kinetic models, or even suggesting alternative low temperature
chemistry. Species measurement will also be helpful to determine the effect of DTBP.
3. Study of DTBP effects on more fuels
As mentioned, more detailed experiments need to be run in order to be more clear
on the effect of DTBP on lower ON fuels. Also, only the oxidation of iso-octane, nheptane and their mixtures with and without DTBP were examined in this. To gain more
insight into the preignition chemistry of hydrocarbon fuels and DTBP effects, more fuels
including alkanes, alkenes, aromatics and their blends should be tested under the
controlled conditions. Examining the effect of DTBP on a real fuel such as gasoline
operating in HCCI conditions is appropriate. At this moment, chemical mechanisms are
available for primary reference fuels, but not for complex mixtures such as gasoline.
154
However, it would be helpful to explore the effect of DTBP on the combustion phasing of
regular gasoline.
155
LIST OF REFERENCES
Aceves, S.M., Smith, J.R., Westbrook, C.K. and Pitz, W.J. (1999), “Compression Ratio
Effect on Methane HCCI Combustion,” ASME Journal of Engineering for Gas
Turbines and Power 121, 569-574.
Aceves, S.M., Flowers, D.L., Westbrook, C.K., Smith, J.R. and Pitz, W.J. (2000), “A
Multi-Zone Model for Prediction of HCCI Combustion and Emissions,” SAE Paper
No. 2000-01-0327.
Aceves, S.M., Flowers, D.L., Martinez-Frias, J., Smith, J.R., Westbrook, C.K., Pitz, W.J.,
Dibble, R., Wright, J.F., Akinyemi, W.C. and Hessel, R.P. (2001a), “A Sequential
Fluid-Mechanic Chemical-Kinetic Model of Propane HCCI Combustion,” SAE
Paper No. 2001-01-1027.
Aceves, S.M., Martinez-Frias, J., Flowers, D.L., Smith, J.R., Dibble, R., Wright, J.F. and
Hessel, R.P. (2001b), “A Decoupled Model of Detailed Fluid Mechanics Followed
by Detailed Chemical Kinetics for Prediction of iso-Octane HCCI Combustion,”
SAE Paper No. 2001-01-3612.
Aceves, S.M., Flowers, D., Martinez-Frias, J., Espinosa-Loza, F. and Pitz, W.J. (2003),
“Fuel and Additive Characterization for HCCI Combustion,” SAE Paper No. 200301-1814.
Addagarla, S., Filipe, D.J., Miller, D.L., Cernansky, N.P. and Green, R.M. (1991), “The
Effects of Speed and Manifold Pressure on Autoignition in a Motored Engine,”
SAE Paper No. 910566. SAE Trans. 100(4), 221-230.
Addagarla, S., Henig, Y., Wilk, R.D., Miller, D.L. and Cernansky, N.P. (1989a), “Effect
of Fuel-Air Mixture Stressing on Preignition Heat Release in a Knock Research
Engine,” SAE Paper No. 892082, SAE Trans. 98(4), 905-921.
Addagarla, S., Miller, D.L., Cernansky, N.P. and Green, R.M. (1989b), “The
Investigation of the Effects of Inlet Conditions on Autoignition in a Motored Knock
Research Engine,” Paper No. WSSCI 89-72, Presented at the Fall Meeting of the
Western States Section/The Combustion Institute, October 23-24, Livermore, CA.
Agosta, A. (2002), “Development of a Chemical Surrogate for JP-8 Aviation Fuel Using
a Pressurized Flow Reactor,” M.S. Thesis, Drexel University, Philadelphia, PA.
Agosta, A., Cernansky, N.P., Miller, D.L., Faravelli, T. and Ranzi, E. (2004), “Reference
Components of Jet Fuel: Kinetic Modeling and Experimental Results,” Exp. Therm.
Fluid. Sci. 28(7), 701-708.
Ai-Rubaie, M.A.R., Griffiths, J.F. and Sheppard, C.G.W. (1991), “Some Observation on
the Effectiveness of Additives for Reducing the Ignition Delay Period of Diesel
Fuels,” SAE Paper No. 912333.
156
Anderson, M.J. and Wilson, G.C. (1952), “Effectiveness of Amyl Nitrate in a Full-Scale
Diesel Engine,” SAE Trans. 6(2), 230-238.
Aoyama, T., Hattori, Y. and Mizuta, J. (1996), “An Experimental Study on PremixedCharge Compression Ignition Gasoline Engine,” SAE Paper No. 960081.
Bales-Gueret, C., Cathonnet, M., Boettner, J., and Gaillard, F. (1992), “Experimental
Study and Kinetic Modeling of Higher Hydrocarbons Oxidation in a Jet-Stirred
Flow Reactor,” Energy and Fuels, 6, 189.
Bartok, W. and Sarofim, A.F. (1991), “Fossil Fuel Combustion: A Sourcebook,” John
Wiley and Sons, Inc., New York.
Benson, S.W. (1981), “The Kinetics and Thermochemistry of Chemical Oxidation with
Application to Combustion and Flames,” Prog. Energy Combust. Sci. 7, 125-134.
Benson, S.W. (1982), “Cool Flames and Oxidation: Mechanism Thermochemistry and
Kinetics,” Oxidation Communication 2, N. 3-4, 169-188.
Bogen, J.S. and Wilson, G.C. (1944), “Ignition Accelerators for Compression-Ignition
Engine Fuel,” Petroleum Refiner, 23(7), 118-128.
Borman, G.L. (1980), “Modeling Flame Propagation and Heat Transfer in Engines-An
Introductory Overview, Combustion Modeling in Reciprocating Engines,” Edited
by Mattavi, J. N. and Amann, C. A., Plenum, New York, NY.
Bourdon, A., Rymer, G. and Wanker, R. (2004), “Optimization of 5-Step Kinetic Scheme
for HCCI Applications,” SAE Paper No. 2004-01-0559.
Brien, E.B. (1956), “Fuel Oil Upgraded to Diesel by Additive,” The Petroleum Engineer,
26(2), C37-C40.
Callahan, C.V., Held, T.J., Dryer, F.L., Minetti, R., Ribaucour, M., Sochet, L.R.,
Faravelli, T., Gaffuri, P. and Ranzi, E. (1996), “Experimental Data and Kinetic
Modeling of Primary Reference Fuel Mixtures,” Proc. Combust. Inst., 26, 739-746.
Cernansky, N.P., Green, R.M., Pitz, W.J. and Westbrook, C.K. (1986), “Chemistry of
Oxidation Preceding End-Gas Autoignition,” Comb. Sci. Tech. 50, 3-25.
Cheng, W., Hamrin, D., Heywood, J., Hochgreb, S., Min, K. and Norris, M. (1993), “An
Overview of Hydrocarbon Emissions Mechanisms in Spark Ignition Engine,” SAE
Paper No. 932708.
Christensen, M., Johansson, B. and Einewall, P. (1997), “Homogeneous Charge
Compression Ignition (HCCI) Using Isooctane, Ethanol and Natural Gas – a
Comparison with Spark Ignition Operation,” SAE Paper No. 972874.
157
Christensen, M., Hultqvist, A. and Johansson, B. (1999), “Demonstrating the Multi Fuel
Capability of a Homogeneous Charge Compression Ignition Engine with Variable
Compression Ratio,” SAE Paper No. 1999-01-3679.
Christensen, M., Hultqvist, A. and Johansson, B. (2001), “The Effect of Piston Topland
Geometry on Emissions of Unburned Hydrocarbons from a Homogenous Charge
Compression Ignition (HCCI) Engine,” SAE Paper No. 2001-01-1893.
Clothier, P.Q., Moise, A. and Pritchard, H.O. (1990), “Effect of Free-Radical Release on
Diesel Ignition Delay under Simulated Cold-Starting Conditions,” Combust. Flame,
81, 242-250.
Cohen, R. S. (1977), “The High Temperature Oxidation and Pyrolysis of Ethane,” Ph.D.
Thesis, Princeton University.
Cowart, J.S., Keck, J.C., Heywood, J.B., Westbrook, C.K. and Pitz, W.J. (1990), “Engine
Knock Predictions Using a Fully-Detailed and a Reduced Chemical Kinetic
Mechanism,” Proc. Combust. Inst., 23, 1055-1062.
Cox, R.A. and Cole, J.A. (1985), “Chemical Aspects of the Autoignition of
Hydrocarbon/Mixtures,” Combust. Flame, 60, 109-123.
Curran, H.J., Gaffuri, P, Pitz, W.J., Westbrook, C.K. and Leppard, W.R. (1995),
“Autoignition Chemistry of the Hexane Isomers: An Experimental and Kinetic
Modeling Study,” SAE Paper 952406, SAE Trans. 104(4), 1184-1195.
Curran, H.J., Gaffuri, P., Pitz, W.J. and Westbrook, C.K. (1998), “A Comprehensive
Modeling Study of n-Heptane Oxidation,” Combust. Flame, 114, 149-177.
Curran, H. J., Gaffuri, P, Pitz, W. J., and Westbrook, C. K. (2002), “A Comprehensive
Modeling Study of iso-Octane Oxidation,” Combust. Flame, 129(3), 253-280.
Dec, J. (2002), “A Computational Study of the Effects of Low Fuel Loading and EGR on
Heat release Rates and Combustion Limits in HCCI Engines,” SAE Paper No.
2002-01-1309
Dechaux, J.C. (1973), “The Negative Temperature Coefficient in the Oxidation of
Hydrocarbons,” Oxidation and Combustion Reviews, 6, 75.
Downs, D., Walsh, A.D. and Wheeler, R.W., (1951), “Reactions that Lead to Knock in
the Spark Ignition Engine,” Proc. Royal Soc. Lond. A 243, 299-358.
Dryer, F. L. and Glassman, I. (1973), “High Temperature Oxidation of CO and CH4,”
Proc. Combust. Inst., 14, 987-1003.
Dryer, F.L. and Glassman, I. (1978): “Combustion Chemistry of Chain Hydrocarbons,”
Progress in Astronautics and Aeronautics, 62, 255-261.
158
Dryer, F.L. (1991), “The Phenomenology of Modeling Combustion Chemistry,” Fossil
Fuel Combustion-A Source Book, 121-213, W. Bartok and A.F. Sarofim, eds.,
Wiley Interscience, New York.
Dunskus, T. and Westwater, J.W. (1961), “The Effect of Trace Additives on the Heat
Transfer to Boilding Isopropanol,” Chem. Engng. Prog. Symp. Ser., 57(32), 173181
Duret, P. and Venturi, S. (1996), “Automotive Calibration of the IAPAC Fluid
Dynamically Controlled Two-Stroke Combustion Process,” SAE Paper No. 960363
Easley, W. L., Agarwal, A. and Lavioe, G. A. (2001), “Modeling of HCCI Combustion
and Emissions Using Detailed Chemistry,” SAE Paper No. 2001-01-1029.
Epping, K., Aceves, S., Bechtold, R. and Dec, J. (2002), “The Potential of HCCI
Combustion for High Efficiency and Low Emissions,” SAE Paper No. 2002-011923.
Ferguson, C.R., Green, R.M. and Lucht, R.P. (1987), “Unburned Gas Temperature in an
Internal Combustion Engine. II: Heat Release Computations,” Combust. Sci. Tech.
55, 63-81.
Filipe, D.J. (1992a), “Reactivity Studies of Primary Reference Fuel Blends in a Motored
Engine,” M.S. Thesis, Drexel University, Philadelphia, PA
Filipe, D.J., Li, H., Miller, D.L. and Cernansky, N.P. (1992b), “The Reactivity Behavior
of n-Heptane and Isooctane Blends in a Motored Knock Research Engine,” SAE
Paper No. 920807.
Flowers, D. L., Aceves, S.M., Smith, J.R., Torres, J., Girard, J. and Dibble, R. (2000),
"HCCI in A CFR Engine: Experiments and Detailed Kinetic Modeling,” SAE Paper
No. 2000-01-0328.
Flynn P.F., Hunter, G.L., Zur Loye, A.C., Akinyemi, W.C., Durrett, R.P., Moore, G.A.,
Muntean, G.G., Peters, L.L., Pierz, P.M., Wagner, J.A., Wright, J.F. and Yeager,
J.M. (1999), “Premixed Charge Compression Ignition Engine with Optimal
Combustion Control,” International Patent WO9942718, World Intellectual
Property Organization.
Flynn, P.F., Hunter, G.L., Durrett, R.P., Farrell, L.A. and Akinyemi, W.C. (2000),
“Minimum Engine Flame Temperature Impacts on Diesel and Spark-Ignition
Engine NOx Production,” SAE Paper 2000-01-1177.
Gluckstein, M.E. and Walcutt, C (1964), “End-Gas Temperature Pressure Histories and
Their Relation to Knock,” SAE Transactions, 69.
159
Gong, X. and Choi, K.J. (2003), “Mixing of Wall-Injected Liquid Jets in Swirling
Flows,” Proceedings of the 3rd Joint Meeting of the U.S. Sections of the
Combustion Institute, March 2003, Chicago.
Gong, X., Zheng, J., Lenhert, D., Miller, D.L. and Cernansky, N.P. (2005a), “The Effect
of DTBP Additive on the Oxidation of Gasoline Primary Reference Fuels in a
Pressurized Flow Reactor,” Paper No. C03, Proceedings of the 4th Joint Meeting of
the U.S. Sections of the Combustion Institute, Philadelphia, PA, March 2005.
Gong, X., Johnson, R.O., Miller, D.L. and Cernansky, N.P. (2005b), “Effects of DTBP
on the HCCI Combustion Characteristics of SI Primary Reference Fuels,” SAE
Paper No. 2005-01-3740, Powertrain & Fluid Systems Conference & Exhibition,
San Antonio, Texas, October 24-27.
Gong, X., Zheng, J., Miller, D.L. and Cernansky, N.P. (2005c), “Experimental and
Computational Study on Preignition Chemistry of SI Primary Reference Fuels in a
Pressurized Flow Reactor,” Paper No. 18587, 37th ACS Middle Atlantic Regional
Meeting (MARM), Rutgers University, NJ, May 22-25.
Gong, X., Choi, K.J. and Cernansky, N.P. (2005d), “Atomization of Liquid Jets in
Swirling Flow in Lean Direct Wall Injection Mode,” Journal of Propulsion and
Power, in press.
Granata, S., Faravelli, T. and Ranzi, E. (2003), “A Wide Range Kinetic Modeling Study
of the Pyrolysis and Combustion of Naphthenes,” Combust. Flame, 132(3), 533-544.
Gray III, A. W. and Ryan III, T. W. (1997), “Homogeneous Charge Compression Ignition
(HCCI) of Diesel Fuel,” SAE Paper No.971676.
Green, R.M., Cernansky, N.P., Pitz, W.J. and Westbrook, C.K. (1987a), “The Role of
Low Temperature Chemistry in the Autoignition of n-Butane,” SAE Paper No.
872109, SAE Trans., 96, Fuels and Lubricants, 575-591.
Green, R.M., Parker, C.D., Pitz, W.J. and Westbrook, C.K. (1987b), “The Autoignition of
Isobutane in a Knocking Spark Ignition Engine,” SAE Paper No. 870169, SAE
Trans., 96, Reciprocating Engines-Spark Ignition and Diesel, 212-230.
Griffiths, J.F. and Phillips, C.H. (1990), “Experimental and Numerical Study of the
Combustion of Ditertiary Butyl Peroxide in the Presence of Oxygen at Low
Pressures in a Mechanically Stirred Closed Vessel,” Combust. Flame, 81, 304-316.
Griffiths, J.F. (1995), “Reduced Kinetic Models and Their Application to Practical
Combustion Systems,” Prog. Energy Combust. Sci., 21, 25-107.
Griffiths, J.F., Halford-Maw, P.A. and Mohamed, C. (1997), “Spontaneous Ignition
Delays as a Diagnostic of the Propensity of Alkanes to Cause Engine Knock,”
Combust. Flame, 111(4), 327-337.
160
Halstead, M.P., Kirsch, L.J., Prothero, A. and Quinn, C.P. (1975), “A Mathematical
Model for Hydrocarbon Autoignition at High Pressure,” Proc. Roy. Soc., A346,
515-538.
Halstead, M.P., Kirsch, L.J. and Quinn, C.P. (1977), “The Autoignition of Hydrocarbon
Fuels at High Temperatures and Pressures-Fitting of a Mathematical Model,”
Combust. Sci. and Tech., 30, 45-60.
Hautman, D.J., Dryer, F.C., Schug, K.D. and Glassman, I. (1981), “A Multiple Step
Overall Kinetic Mechanism for the Oxidation of Hydrocarbons,” Combust. Sci. and
Tech., 25, 219.
Henig, Y., Addagarla, S., Miller, D.L., Wilk, R.D. and Cernansky, N.P. (1989):
“Autoignition of n-Butane/Iso-butane Blends in a Knock Research Engine,” SAE
Paper No. 890157, SAE Trans., 98(40), 136-148.
Heywood, J.B. (1988), “Internal Combustion Engines Fundamentals,” McGraw Hill,
New York, NY.
Heywood, J.B. and Sher, E. (1999), “The Two-Stroke Cycle Engine,” Taylor and Francis,
Inc., Philadelphia, PA.
http://www.dieselnet.com/standards/us/light.html#tier2
Hu, H. and Keck, J. (1987), “Autoignition of Adiabatically Compressed Combustible Gas
Mixtures,” SAE Paper No. 872110, SAE Trans., 96, Fuels and Lubricants, 592-604.
Hurn, R.W. and Hughes, K.J. (1956), “Ignition Accelerator and Auto-Ignition
Environment,” Ind. and Eng. Chem., 48 (10), 1904-1908.
Inomata, T., Griffiths, J.F. and Pappin, A.J. (1990), “The Role of Additive as Sensitizer
for the Spontaneous Ignition of Hydrocarbons,” Proc. Combust. Inst., 23, 17591766.
Ishibashi, Y., and Asai, M. (1996), “Improving the Exhaust Emissions of Two-Stroke
Engines by Applying the Activated Radical Concept,” SAE Paper No. 960742.
Kaahaaina, N.B., Simon, A.J., Caton, P.A. and Edwards, C.F. (2001), “Use of Dynamic
Valving to Achieve Residual-Affected Combustion,” SAE Paper No. 2001-01-0549.
Kamel, M.M. (1984), “Effect of Fuel Ignition Improvers on Diesel Engine Performance,”
SAE Paper No. 840109
Kelly-Zion, P.L. and Dec, J.E. (2000), “A Computational Study of Effect of Fuel-Type
on Ignition Time in HCCI Engines,” Proc. Combust. Inst. 28, 1187-1194.
161
Khan, A.R. (1998), “Oxidation of Industry Standard Fuels and Blends of Primary
Reference Fuels in the Low Temperature and Negative Temperature Coefficient
Regions,” M.S. Thesis, Drexel University, Philadelphia, PA.
Kimura, S., Aoki, O., Ogawa, H. and Muranaka, S. (1999), “New Combustion for UltraClean and High-Efficiency Small DI Diesel Engines,” SAE Paper No. 1999-013681.
Kimura, S., Aoki, O., Kitahara, Y. and Aiyoshizawa, E. (2001), “Ultra-clean Combustion
Technology Combining a Low- temperature and Premixed Combustion Concept for
Meeting Future Emission Standards,” SAE Paper No. 2001-01-1831.
Kirsch, L.J., Rosenfield, J.L. and Sumers, R. (1981), “Studies of Fuel Injection into
Rapid Compression Machine,” Combust. Flame, 43, 11-21
Koert, D.N. (1990), “Effects of Pressure on Hydrocarbon Oxidation Chemistry,” Ph.D.
Thesis, Drexel University, Philadelphia, PA.
Koert, D.N. and N.P. Cernansky. (1992), “A Flow Reactor for the Study of
Homogeneous Gas-Phase Oxidation of Hydrocarbons at Pressures up to 20 atm (2
MPa),” Meas. Sci. Tech., 3(6), 607-613.
Koert, D.N., Pitz, W.J., Bozzelli, J.W., and Cernansky, N.P. (1996), “Chemical Kinetic
Modeling of High Pressure Propane Oxidation and Comparison to Experimental
Results,” Proc. Combust. Inst., 26, 633-640.
Koopmans, L. and Denbratt, I. (2002), “Cycle to Cycle Variations and Their Influence on
Cycle Resolved Unburned Hydrocarbons and Gas Temperature from a Camless
Gasoline Compression Ignition Engine,” SAE paper No. 2002-01-0110.
Kraft, M., Maigaard, P., Mauss, F., Christensen, M. and Johansson, B. (2000),
“Investigation of Combustion Emissions in a HCCI Engine: Measurements and A
New Computational Model,” Proc. Combust. Inst., 28, 1195-1201.
Lam, S.H. and Goussis, D.A. (1988), “Understanding Complex Chemical Kinetics with
Computational Singular Perturbation,” Proc. Combust. Inst., 22, 931-941.
Lenhert, D. B. (2004a), “The Oxidation of a Gasoline Fuel Surrogate in the Negative
Temperature Coefficient Region,” M.S. Thesis, Drexel University, Philadelphia, PA.
Lenhert, D.B. (2004b), “The Oxidation of JP-8 and its Surrogates in the Low and
Intermediate Temperature Regime,” Ph.D. Thesis, Drexel University, Philadelphia,
PA.
Leppard, W.R. (1987), “The Autoignition Chemistry of n-Butane: An Experimental
Study,” SAE paper 872150, SAE Trans., 91, Fuels and Lubricants, 934-957.
162
Leppard, W.R. (1988), “The Autoignition Chemistry of Isobutane: A Motored Engine
Study,” SAE paper 881606, SAE Trans., 92, Fuels and Lubricants, 658-680.
Leppard, W.R. (1989), “A Comparison of Olefin and Paraffin Autoignition Chemistries:
A Motored Engine Study,” SAE paper No. 892081, SAE Trans. 98(4), 879-904.
Leppard, W.R. (1992), "The Autoignition Chemistry of Primary Reference Fuels,
Olefin/Paraffin Binary Mixtures, and Non-Linear Octane Blending," SAE Paper No.
922325, SAE Trans., 101(4), 1683-1705.
Li, H., Miller, D.L. and Cernansky, N.P. (1992), "A Study on the Application of a
Reduced Chemical Reaction Model to Motored Engines for Heat Release
Prediction," SAE Paper No. 922328, SAE Trans., 101(3), 2021-2034.
Li, H., Prabhu, S.K., Miller, D.L. and Cernansky, N,P. (1994), “Autoignition Chemistry
Studies on Primary Reference Fuels in a Motored Engine,” SAE paper No. 942062,
SAE Tran., 103(4), 1948-1968.
Li, H. (1995), “Autoignition Chemistry Studies of Primary Reference Fuels and Their
Mixtures with Oxygenates in a Research Engine,” Ph.D. Thesis, Drexel University,
Philadelphia, PA.
Li, H., Miller, D. L. and Cernansky, N. P. (1996), “Development of a Reduced Chemical
Kinetic Model For Prediction of Preignition Reactivity and Autoignition of Primary
Reference Fuels,” SAE Paper No. 960498.
Li, T.M. and Simmons, R.F. (1986), “The Action of Ignition Improvers in Diesel Fuels,”
Proc. Combust. Inst., 21, 455-463.
Lida, N. (1994), “Combustion Analysis of Methanol-Fueled Active Thermo-Atmosphere
Combustion (ATAC) Engine Using a Spectroscopic Observation,” SAE Paper No.
940684.
Lida, N. (1997), “Alternative Fuels and Homogeneous Charge Compression Ignition
Combustion Technology,” SAE Paper No. 972071.
Lignola, P.G. and Reverchon, E. (1987), “Cool Flames,” Prog. Energy Combust. Sci., 13,
75.
Lovell, W.G. (1948), “Knocking Characteristics of Hydrocarbons,” Ind. and Eng. Chem.,
40(12), 2388-2438.
Lu, T. and Law, C.K. (2005), “A Directed Relation Graph Method for Mechanism
Reduction,” Proc. Combust. Inst., 30, 1333-1341.
Martinez-Frias, J., Aceves, S.M., Flowers, D., Smith, J.R. and Dibble, R. (2000), “HCCI
Engine Control by Thermal Management,” SAE Paper 2000-01-2869.
163
Maas, U. and Pope, S.B. (1992), “Simplifying Chemical Kinetics: Intrinsic LowDimensional Manifolds in Composition Space,” Combust. Flame, 88(3-4), 239-264.
McCormick, T.W. (1994), “C4 Oxidation Analysis Using FTIR Spectroscopy,” M.S.
Thesis, Drexel University, Philadelphia, PA.
McGreath, C.G. (1971), “The Effect of Fuel Additives on the Exhaust Emissions From
Diesel Engines,” Combust. Flame, 17, 359-366.
Miller, J. and Bowman, C. (1989), “Mechanism and Modeling of Nitrogen Chemistry in
Combustion”, Prog. Energy Combust. Sci., 15, 287-338.
Müller, C.U., Peters, N. and Liňán, A. (1992), “Global Kinetics for n-Heptane Ignition at
High Pressures,” Proc. Combust. Inst., 24, 777-784.
Najt, P.M. and Foster, D.E. (1983), “Compression-ignited Homogeneous Charge
Combustion,” SAE Paper No. 830264.
Nehse, M. and Warnatz, J. (1996), “Kinetic Modeling of the Oxidation of Large Alphatic
Hydrocarbons,” Proc. Combust. Inst., 26, 773-780.
Noguchi, M., Tanaka, Y., Tanaka, T. and Takeuchi, Y. (1979), “A Study on Gasoline
Engine Combustion by Observation of Intermediate Reactive Products during
Combustion,” SAE Paper No. 790840.
Ogink, R. (2004), “Computer Modeling of HCCI Combustion,” Ph.D. Thesis, Chalmers
University of Technology, Goteborg, Sweden.
Oguma, H., Ichikura, T. and Lida, N. (1997), “Adaptability of Gasoline, Methanol,
Methane Fuels to Lean Burn in an ATAC Engine,” Trans. of the J. Soc. of Mech.
Engineers (B), 613 (63), 248-255.
Olsson, J.O., Tunestal, P. and Johansson, B. (2001), “Closed-Loop Control of an HCCI
Engine,” SAE paper 2001-01-1031.
Onishi, S., Hong Jo, S., Shoda, K., Do Jo, P. and Kato, S. (1979), “Active ThermoAtmosphere Combustion (ATAC) – A New Combustion Process for Internal
Combustion Engine,” SAE Paper No. 790501.
Peng, Z., Zhao, H. and Ladommatos, N. (2003), “Effect of Air/Fuel Ratios and EGR
Rates on HCCI Combustion of n-Heptane, a Diesel Type Fuel,” SAE Paper No.
2003-01-0747.
Peters, N. and Rogg, B. (1993), “Reduced Kinetic Mechanisms for Applications in
Combustion Systems,” Springer-Verlag, Berlin.
Pilling, M.J. (1997), “Comprehensive Chemical Kinetics,” Elsevier.
164
Pishinger, F., Reuter, U. and Scheid, E. (1988), “Self-Ignition of Diesel Sprays and its
Dependence on Fuel Properties and Injection Parameters,” SAME J. of Engng. for
Gas Turbine and Power, 110, 399-404.
Prabhu, S.K., Bhat, R.K., Miller, D.L. and Cernansky, N.P. (1996), “1-Pentene Oxidation
and Its Interaction with Nitric Oxide in the Low and Negative Temperature
Coefficient Regions,” Combust. Flame, 104, 377.
Ranzi, E., Gaffuri, P, Faravelli, T. and P. Dagaut. (1995), “A Wide-Range Modeling
Study of n-Heptane Oxidation,” Combust. Flame, 103(1-2), 91-106.
Ranzi, E., Faravelli, T., Gaffuri, P, A. Sogaro, A. D'Anna and A. Ciajolo. (1997), “A
Wide-Range Modeling Study of iso-Octane Oxidation,” Combust. Flame, 108(1-2),
24-42.
Rassweiler, G.M. and Withrow, L. (1933), “Spectrographic Detection of Formaldehyde
in an Engine Prior to Knock,” Ind. and Eng. Chem., 25(12), 1350-1366.
Robbins, W.E., Audette, R.R. and Reynolds, N.E. (1951), “Performance and Stability of
Some Diesel Fuel Ignition Quality Improvers,” SAE Trans., 5(3), 404-417.
Ryan III, T.W. and Callahan, T.J. (1996), “Homogeneous Charge Compression Ignition
of Diesel Fuel,” SAE Paper No. 961160.
Salooja, K.C. (1962), “The Role of Combustion Promoters,” J. Inst. Petroleum, 48 (460),
119-129
Satcunanathan, S. and EI-Nesr, M.S. (1972), “Effect of Additive on the Cetane Ratings of
Diesel Fuels as Related to the Ignition Delays of Liquid Fuel Droplets Impinging on
a Hot Surface,” Proc. IMechE., 186(28), 317-321.
Schommers, J., Duvinage, F., Stotz, M., Peters, A., Ellwanger, S., Koyanagi, K. and
Gildein, H. (2000), “Potental of Common Rail Injection System for Passenger Car
DI Diesel Engines,” SAE Paper No. 2000-01-0944.
Semenov, N.N. (1958), “Some Problems in Chemical Kinetics and Reactivity,” Vol. 2,
Chapter XII, Princeton University Press, Princeton, NJ.
Sharke, P. (2000), “Otto or Not, Here it Comes,” Mechanical Engineering, 122(6), 62-66.
Shimazaki, Z., Akagawa, H. and Tsujimura, K. (1999), “An Experimental Study of
Premixed Lean Diesel Combustion,” SAE Paper No. 1999-01-0181.
Schreiber, M., Sakak, S.A., Lingens, A. and Griffiths, F.J. (1994), “A Reduced
Thermokinetic Model for the Autoignition of Fuels with Variable Octane Ratings,”
Proc. Combust. Inst. 25, 933-940.
165
Smith, J.R., Green, R.M., Westbrook C.K. and Pitz, W.J. (1985), “Experimental and
Modeling Studies of Engine Knock,” Proc. Combust. Inst., 20, 91-100.
Tanaka, S., Ayala, F., Keck, J.C. and Heywood, J.B. (2003), “Two-stage Ignition in
HCCI Combustion and HCCI Control by Fuels and Additives,” Combust. Flame,
132, 219-239.
Theobald, M.A. and Henry, R. (1994), “Control of Engine Load via Electromagnetic
Valve Actuators,” SAE Paper No. 940816.
Thring, R.H. (1989), “Homogeneous Charge Compression Ignition (HCCI) Engines,”
SAE Paper No. 892068.
Turns, S.R. (1999), “An Introduction to Combustion-Concepts and Applications,” Second
Edition, McGraw Hill.
Vermeersch, M.L., Held, T.J., Stein, Y. and Dryer, F.L. (1991), “Autoignition Chemistry
Studies of n-Butane in a Variable Pressure Flow Reactor,” SAE Paper No. 912326.
Valdmanis, E. and Wulfhorst, D.E. (1970), “The Effect of Emulsified Fuels and Water
Induction on Diesel Combustion,” SAE Paper No. 700736.
Violi, A., Yan, S. Eddings, E.G. and Sarofim, A.F. (2002), “Experimental Formulation
and Kinetic Model for JP-8 Surrogate Mixtures,” Comb. Sci. and Tech., 174 (11-12),
399-417.
Walsh, A.D. (1963), “The Knock Ratings of Fuels,” Proc. Combust. Inst., 9, 1046-1055.
Wang, S., Miller, D.L. and Cernansky, N.P. (1996a), “Prediction of Preignition
Reactivity for n-Butane and iso-Butane Blends using a Reduced Chemical Kinetic
Model,” SAE Paper No. 961154.
Wang, S., Miller, D.L. and Cernansky, N.P. (1996b), “A Reduced Chemical Kinetic
Model for Autoignition of the Butanes,” SAE Paper No. 962106.
Wang, S., Miller, D.L., Cernansky, N.P., Curran, H.J., Pitz, W.J. and Westbrook, C.K.
(1999), “A Flow Reactor Study of Neopentane Oxidation at 8 Atmospheres:
Experiments and Modeling,” Combust. Flame, 118, 415-430.
Wang, S. (1999), “Experimental and Modeling Study of Preignition Chemistry of
Hydrocarbons,” Ph.D. Thesis, Drexel University, Philadelphia, PA.
Westbrook, C.K. and Pitz, W.J. (1984), “A Comprehensive Chemical Kinetic Mechanism
for Oxidation and Pyrolysis of Propane and Propene,” Combust. Sci. and Tech. 37,
117-131.
Westbrook, C.K. and Pitz, W.J. (1987), “Detailed Kinetic Modeling of Autoignition
Chemistry,” SAE Paper No. 872107, SAE Trans. 96.
166
Westbrook, C.K., Pitz, W.J. and Leppard, W.R. (1991), “The Autoignition Chemistry of
Paraffinic Fuels and Pro-Knock and Anti-Knock Additives: A Detailed Chemical
Kinetic Study,” SAE Paper No. 912314, SAE Trans., 100(4), 605-622.
Wilk, R.D. (1986a), “Pre-ignition Oxidation Characteristics of Hydrocarbon Fuels,” Ph.D.
Thesis, Drexel University, Philadelphia, PA.
Wilk, R.D., Cernansky, N.P. and Cohen, R.S. (1986b), “The Oxidation of Propane at
Low and Transition Temperatures,” Combust. Sci. and Tech., 49, 41-78.
Wilk, R.D., Pitz, W.J., Westbrook, C.K., Addagarla, S., Miller, D.L., Cernansky, N.P.
and Green, R.M. (1990), “Combustion of n-Butane and iso-Butane in an Internal
Combustion Engine: A Comparison of Experimental and Modeling Results,” Proc.
Combust. Inst., 23, 1047-1056.
Wood, C. H. (1994), “The Oxidation of n-Pentane and 1-Pentene in the Low Temperature
and Negative Temperature Coefficient Regions,” M. S. Thesis, Drexel University,
Philadelphia, PA.
Yamaguchi, J. (1997), “Honda Readies Activated Radical Combustion Two-stroke
Engine for Production Motorcycle,” Automotive Engineering 105(1), 90-92.
Yang, W. (2002), “The Chemistry Controlling Post Combustion Hydrocarbon Oxidation
and Homogenous Charge Compression Ignition,” Ph. D. Thesis, Drexel University,
Philadelphia, PA.
Yao, M., Zheng, Z., Zhang, B., and Chen, Z. (2004), “The Effect of PRF Fuel Octane
Number on HCCI Operation,” SAE Paper No. 2004-01-2992.
Zheng, J., Yang, W., Miller, D.L. and Cernansky, N.P. (2001), “Prediction of Pre-ignition
Reactivity and Ignition Delay for Using a Reduced Chemical Kinetic Model,” SAE
Paper No. 2001-01-1025, SAE Trans. 110(3), 999-1006.
Zheng, J., Miller, D.L. and Cernansky, N.P. (2002a), “A Skeletal Chemical Kinetic
Model for the HCCI Combustion Process,” SAE Paper No. 2002-01-0423, SAE
Trans. 111(3), 898-912.
Zheng, J., Miller, D.L. and Cernansky, N.P. (2002b), “Development of a Skeletal Kinetic
Model for Prediction of Preignition Reactivity of Hydrocarbons ,” Poster No. WIP
18 – 1388, 29th Intl. Symp. on Combust., Chicago, IL, July 2004
Zheng, J., Miller, D.L. and Cernansky, N.P. (2004), “A Global Reaction Model for the
HCCI Combustion Process,” SAE Paper No. 2004-01-2950.7
167
APPENDIX A: HYDROCARBON OXIDATION AND AUTOIGNITION
CHEMISTRY
Since the autoignition phenomena is strongly dependent on the oxidation
chemistry, understanding the chemical processes that cause autoignition is critical for
solving the problem associated with HCCI engines. A brief review of the HCCI related
hydrocarbon oxidation and autoignition chemistry is given in this Appendix.
1. Introduction
Studies of autoignition and hydrocarbon oxidation began in the early 1900’s when
knock was identified as a limitation on engine output and fuel efficiency. Over the years,
sampling, measurement and analysis techniques have improved and so has our
understanding of the fundamentals of autoignition and hydrocarbon oxidation.
Historically, some low and intermediate temperature hydrocarbon oxidation
phenomena were encountered accidentally. In 1882, Perkin first observed cool flames
[Lignola and Reverchon, 1987]. As early as 1920, researchers noticed the differences in
the autoignition characteristics of different pure hydrocarbons. A negative temperature
coefficient behavior was found when Pease was studying oxidation of propane in a flow
reactor in 1929 [Dechaux, 1973). The first evidence of preflame reactions was presented
by Rassweiler and Withrow [1933]. In 1948, Lovell published an extensive review and
tabulation of the autoignition characteristics of over 325 hydrocarbons [Lovell, 1948].
Notably, Lovell related the chemical structure of hydrocarbons to the tendency of fuel
autoignition. However, Lovell did not forward a kinetic or mechanistic explanation for
168
the observed phenomena. It was not until Walsh [1963] proposed a mechanistic link
between autoignition tendency and fuel structure that there was a reasonable explanation
for the wide differences in knock behavior. Walsh suggested that the isomerization of the
RO2• radical (where R is the original fuel molecule, minus one hydrogen atom) plays a
critical role in the oxidation of hydrocarbons, since the isomerized radical can lead to a
series of chain branching reactions. Thus, an approach to understanding the autoignition
behavior of a fuel is to investigate the mechanism of the fuel decomposition and
oxidation prior to the point of autoignition.
Since Walsh, extensive studies have been conducted on the oxidation of
hydrocarbons, greatly increasing the understanding of the combustion process. In general,
the combustion process may be described as a series of complex chain branching,
carrying, and terminating reactions involving stable and radical species. It is commonly
accepted that the hydrocarbon oxidation process may be separated into three distinct
temperature regimes. Corresponding to each of these temperature regimes is a dominant
branching agent [Wilk, 1986; Koert, 1990; Dryer, 1991], namely alkylperoxy radicals in
the low temperature region, hydroperoxy radical in the intermediate temperature region,
and hydroxyl, and atomic oxygen and hydrogen radicals in the high temperature region.
Generally, the combustion environment, such as temperature, pressure, and
equivalence ratio effects the location of the boundaries between each regime. At one
atmosphere, the hydrocarbon oxidation process can be divided along the following
approximate boundaries:
(1) low temperature, < 650 K
(2) intermediate temperature, 650-1000 K
169
(3) high temperature, > 1000 K
Since many of the reactions in each regime are pressure dependent, the
temperature of each regime will shift as the pressure of the combustion process increases.
The temperature regime where the autoignition process occurs has been experimentally
measured by several researchers [e.g., Gluckstein and Walcutt, 1964; Smith et al., 1985;
Griffiths et al., 1997] and although disputed by some researchers, it is generally accepted
that the fuel autoignites in the intermediate temperature regime. Since the fuel spends
considerable time in the low temperature regime, it becomes critical to understand the
oxidation process in both the low and intermediate temperature regimes at elevated
pressures in order to understand the autoignition phenomena.
2. Mechanisms of Hydrocarbon Oxidation
Mechanisms of hydrocarbon oxidation are evolutionary products and they change
with time as new insights are developed. Several general mechanisms of hydrocarbon
oxidation at low and intermediate temperatures have been outlined and developed.
Semenov [1958] first introduced the concept of degenerate branched chain
reactions. The general process is that fuel and oxygen first form a pool of relatively
unreactive intermediate species. The intermediates subsequently react along one of two
paths to form either stable molecules, which lead to non chain branching, or highly
reactive free radicals, which lead to chain branching. The relative importance of either
path is influenced strongly by the reaction conditions. Essentially, the oxidation process
can be modeled by a sequence of elementary chemical reactions in which radicals are
170
created, propagated, or destroyed.
These reactions can be grouped into several
fundamental classifications [Pilling, 1997]:
a. Primary Initiation: formation of radicals from parent fuel molecule;
b. Secondary Initiation: radicals formed from other “stable” intermediates;
c. Chain Propagation: reaction where the number of radicals remain unchanged;
d. Chain Branching: reaction where the number of radicals increases;
e. Termination: removal of radicals from the reactive pool.
These general classifications are applicable to any hydrocarbon class, e.g. alkane,
alkene, naphthene, or aromatics.
The primary initiation can occur by thermal
decomposition of the fuel or by chemical reaction with another species. Many factors
affect the relative ratio between these two processes, including but not limited to
temperature, pressure, and chemical structure. Once the initial radical pool has been
established, the radicals can interact with other stable species or radicals. If the reaction
increases the numbers of radicals, then the reaction is referred to as a chain branching
pathway. For example, the overall temperature dependence and the exothermicity of the
reaction process could lead to the very complex kinetic behavior associated with the
chemical induction period, cool flames, and the negative temperature coefficient behavior
[Bartok and Sarofim, 1991]. This degenerate branching mechanism set the basis for later
mechanism development.
Benson introduced a general oxidation mechanism for low molecular weight
alkanes in the low temperature regime [Benson, 1981]. According to Dryer [1991], this
mechanism can be written as follows:
171
RH+ O2+ M = R• + HO2• + M
(A-1)
R• + O2• = RO2•
(A-2)
R• + O2 (+M) = olefin + HO2• (+M)
(A-3)
RH +RO2• = ROOH + R•
(A-4)
RO2• = R'CHO + R"O
(A-5)
RH + HO2• = HOOH + R•
(A-6)
ROOH = RO• + OH•
(A-7)
OH• + RH = H2O + R•
(A-8)
R'CHO + O2 = R'CO• + HO2•
(A-9)
RO2• Æ destruction
(A-10)
HO2• Æ destruction
(A-11)
and extended to higher temperature with
HOOH + M = OH• + OH•
(A-12)
and for larger alkanes
RO2• = QOOH•
(A-13)
Olefin + HO2• = epoxide + OH•
(A-14)
QOOH• + O2 = R'CHO + ketone + 2OH•
(A-15)
172
Where RH and R'CHO represent the fuel and aldehydes, respectively, and
reactions (A-1) to (A-11) describe the low temperature oxidation mechanism for C2's and
C3's alkanes, along with other reactions having to be added as temperature increases (e.g.,
reaction A-12) and as the initial fuel hydrocarbon molecule larger than 3 carbon atoms
(e.g., reactions A-13 to A-15). Some of these reactions are not elementary processes (e.g.,
reaction A-15) but represent the result of several elementary reactions.
A brief description of this mechanism is as follows. In general, alkanes are
essentially unreactive below 400 K unless either chemical or photochemical initiators are
active. Above 420 K oxidation is initiated by the removal of a hydrogen by molecular
oxygen (A-1). However, this step, called an abstraction reaction, is highly endothermic,
roughly 45-55 kcal/mol depending on the bond energy of the abstracted H atom.
Therefore, as the rate is characterized by activation energy proportional to the
endothermicity, it is very slow. Due to the variations in the bond energies, the abstraction
process is very selective as to which hydrogen is removed and depending on the
abstraction site, a different alkyl radical R• will be formed [Westbrook et al., 1991;
Leppard, 1992]. In the low temperature regime, the next step is addition of the oxygen
molecule to the alkyl radical R• forming alkylperoxide radicals, RO2• (A-2). RO2•
subsequently produces the chain branching agent ROOH (A-4), which decomposes to
form two radicals OH• and RO• (A-7). The reactions (A-4) and (A-7) represent small
molecule chain branching. For larger hydrocarbon molecules (> C3), reaction (A-13), an
important isomerization reaction, will occur and chain branching follows. Reaction (A15) represents the overall result of this branching, the consumption of the parent fuel
molecule is accomplished by reactions (A-4), (A-6) and (A-8). Due to the high reactivity
173
of the hydroxyl radical OH•, the fuel is consumed primarily by the attack of radicals such
as OH• via (A-8).
As temperature increases, reaction (A-2) becomes effectively
reversible, and another oxidation path of R• radical, (A-3), becomes important. Since
(A-3) produces alkenes and HO2•, relatively stable species at these temperatures, it has
an inhibiting effect on the overall reaction rate.
The mechanism shift explains the
decrease of overall reaction rate with the increase of temperature (due to effectively
reversible reaction (A-2) and non-chain branching reaction (A-3)), known as negative
temperature coefficient (NTC) behavior. For many hydrocarbons, there is such a NTC
temperature range, which is usually between 600 K to 800 K. As the temperature is
further increased into the intermediate temperature regime, the decomposition of
hydrogen peroxide becomes the dominant chain branching path (A-12) and the reaction
again accelerates.
174
APPENDIX B: ATOMIZATION OF LIQUID JETS IN SWIRLING FLOWS
USING A LABORATORY GAS TURBINE COMBUSTOR
OPERATING IN LEAN DIRECT WALL INJECTION MODEL*
Similar to SI and CI engine systems, meeting the environment and energy
challenges for gas turbine power plants requires new combustion concepts. In this
appendix, an initial study on a new ultra-low-emissions gas turbine combustor concept is
reported. As the first stage toward understanding the combustion phenomena in a lean
direct wall injection (LDWI) mode, the hydrodynamic behavior of wall-injected liquid
jets in a confined cold swirling air flows was investigated.
Three vane-type swirlers (with vane angle α = 30°, 45° and 60°) were employed
in this study to generate swirling airflows in a circular channel. Liquid jets injected from
a simple round orifice were used to characterize the initial breakup and subsequent jet
atomization in the swirling airflows. With water as the test liquid, the parameters that
affect the atomization phenomena, such as jet diameter, momentum rate ratio of air to jet,
liquid jet inclination angle and swirler configurations, which are directly related to swirl
number, were experimentally investigated. The atomization phenomena were observed
and recorded by photographs and videos. Results indicated that there are optimal jet
inclination angles for different swirlers for uniform distribution of droplets
(e.g.,
optimum jet inclination angles were found as 32°, 35° and 42° under swirler vane angles
α = 30°, 45° and 60°). Relations between the jet diameter and the optimum momentum
* The material in this appendix was the basis for Paper No.E10, presented at the 3rd
Joint Meeting of the U.S. Sections of the Combustion Institute, Chicago, IL, March 2003
[Gong et al., 2003] and for a Journal of Propulsion and Power paper by Gong et al.
[2005d] which is in press.
175
rate ratio of air to liquid were studied. Also, a dimensionless analysis was conducted to
correlate the optimum atomization and the above parameters.
B.1. Nomenclature
d
dv
dh
D
Dr
FN
L
•
m
Mg
•
M jet
•
M air
R
Rv
Rh
Rmom
SN
Vair
Vjet
x
X
α
θ
ρ
Φ
∆P
=
=
=
=
=
=
=
Test Chamber and Swirler Diameter, mm
Vane Diameter, mm
Inner Hub Diameter, mm
Injector Inner Diameter, mm
Reference Diameter, 1mm
Flow Number, mm2
Length of Injector, mm
= Mass Flow Rate, ρvA, g/s
= generic functions
= Liquid Jet Momentum Rate, ρAVjet, N
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
Air Momentum Rate, ρAVair 2, N
Radius of Test Chamber and Swirler, mm
Vane Radius, mm
Inner Hub Radius, mm
Momentum Rate Ratio of Air to Liquid Jet
Swirl Number
Mean Air Velocity, m/s
Mean Liquid Jet Velocity, m/s
Distance From the Swirler, mm
Test Section Length, 254mm
Swirler Van Angle, °
Jet Inclination Angle, °
Density, kg/m3
Equivalence Ratio
Differential Pressure, Pa
B.2. Introduction
The environmental and energy challenges for gas turbines require new
combustion concepts. The design of a low emission, high thermal efficiency gas turbine
combustor consists of a balance between providing enough time and sufficiently high
176
temperatures to complete combustion and keeping time short and temperatures low
enough to minimize NOx1.
Concepts that have experimentally demonstrated low
emission include the lean-premixed-pre-vaporized (LPP), the rich-burn/quick-mix/leanburn (RQL), the lean-direct injection (LDI) and catalytic combustion1, 2. Of these, LPP
has received the most attention.
Lean-direct injection, in which the fuel is injected directly into the flame zone,
has been under consideration as an alternative to LPP, because it does not have a
potential for autoignition or flashback, which could be the main disadvantage of LPP.
The technique described in this study is called Lean Direct Wall Injection (LDWI) and
can be described as when fuel jet is injected from combustor wall directly, without
premixed and pre-vaporized, into swirling flow of the main combustor3.
Atomization, vaporization and mixing of a liquid into a gaseous medium such as
air are generally achieved through the atomization of the liquid into fine droplets or
ligaments from the initial bulk liquid and subsequent evaporation of the liquid. The
vaporized liquid can then be mixed into the gaseous medium. Ideally, the time and space
required for complete vaporization should be minimized. This requires the production of
the smallest possible droplets in the gaseous medium. Liquid jet atomization is thus a
critical process for LDWI, since the fuel is not premixed and pre-vaporized and the
combustion efficiency and NOx emission of this concept depend heavily on the fuel
distribution. It will be necessary to produce uniform and rapid atomization of the fuel jet
in LDWI in order to form a uniform gaseous phase fuel and air mixture in the practical
application. In LDI, traditional central injection results in a rich zone in the central core
flow and makes it difficult to obtain uniform fuel-air ratio in a cross-section of
177
combustors
4, 5
. Choi et al.6 compared wall injection with central injection using a
visualization technique. It was found that atomization of spray injected coaxially was
inferior to the case of wall injection.
Experimental results also showed that NOx
emissions were reduced significantly by using wall injection 7.
Swirling flows have been commonly used and studied for decades due to the
stabilization of high-intensity combustion processes that it provides by means of forming
recirculation zones and reducing combustion lengths8.
Ahmed and Nejad9
experimentally investigated isothermal swirling flow in a dump combustor and showed
that the size and the strength of the core recirculation region were dependent on the
swirler design and strength.
Sheen et al.10 and Young et al.11 experimentally and
computationally studied the velocity field and recirculation zones in confined geometry
and found that the characteristics of the flow structures are dependent on two
dimensionless parameters, the Reynolds number Re and the swirl number SN, and that
the length of the recirculation zone varies with flow conditions.
Most of the studies on the breakup and atomization of liquid jets were conducted
in a cross flow (or transverse flow). The behavior of a liquid jet injected transversely into
a high velocity cross flow has been examined in both supersonic and subsonic flows
largely through experiment
12-15
. Inamura and Nagai
12
and Baranovsky and Schetz
13
have shown that the fuel distribution is very sensitive to the jet operating conditions such
as liquid-to-air momentum ratio and injection angle, and may be controlled through
various parameters, such as nozzle diameters and shapes, injectant flow rate and
freestream pattern. It was also found that the liquid fuel jet disintegrates into small
particles because of the shear force between the fuel jet and the air flow. In general, the
178
jet breaks up into liquid clumps and the liquid clumps then disintegrate into finer
particles13. The phenomena associated with angled injection into subsonic crossflows
have been studied by Fuller et al.16. It was found that the column fracture is governed by
non-aerodynamic breakup, such as a turbulent liquid jet in a quiescent gas for Tb > 1 and
by aerodynamic breakup for Tb < 1 (where Tb is a breakup regime parameter).
The behavior of a liquid jet in the confined swirling flow of gas turbines is
significantly different from that in cross flow. Results from angled injection in the cross
flow regime are not directly applicable to the swirling flow regime. The phenomena
associated with angled injections into swirling flow and the detailed mechanism of
breakup of the liquid fuel jet and atomization phenomena in confined swirling airflow has
not been revealed up to this point.
In general, the atomization processes in LDWI are complicated, and they are
nearly impossible to accurately predict with current computational techniques. Therefore,
combustor designers must rely on empirical correlations and extensive databases
covering a very wide range of operating conditions and geometrical configurations. It
was the purpose of this investigation to examine the effects of atomization factors on the
breakup and atomization processes of liquid jets in swirling flow.
B.3. Experimental Apparatus and Instruments
Figure B-1 shows the detailed schematic of the test facility. Air was supplied by a
3.5 kW centrifugal blower. Uniform air flow was produced through a flow straightener
which was located upstream of the swirler. The air velocity was adjusted by a bypass
179
valve installed before the straightener and measured by a pitot static tube with a Dwyer
air velocity kit (Model series 400) 254 mm upstream of the swirler. The velocity kit has
a range of 0 - 97.5 m/s with an accuracy of ±2%.
The transparent test section was
constructed of an acrylic tubing, 254 mm long with 76 mm internal diameter and 6 mm
wall. Figure B-2 shows the details of the transparent test section.
Lens and Mirror Assembly
N2
Argon Laser
Water
Laser Beam (sheet)
Tank
P
CCD Camera
Pressure control panel
Straightener
Transparent Channel
Air from Blower
Manometer
Spray Collector
Swirler
Beam Dump
Figure B-1 Schematic of model gas turbine combustor facility
25.4mm
2Rv
D = 76 mm
θ
2Rh
x X = 254 mm
Figure B-2 Test section detail
180
A twenty-gallon stainless steel vessel pressurized with nitrogen was used to
provide water.
Water was injected into the test section by using five hypodermic
injectors with diameters of 1.19, 0.84, 0.60, 0.515 and 0.344 mm, Table B- 1.
Table B-1 Injector configurations
Diameter, mm
0.344
0.515
0.60
0.84
1.19
L/D
49.4
44.7
41.7
25.0
21.8
FN, mm2
0.078
0.208
0.254
0.548
1.085
Each injector maintained similar exit conditions with lengths chosen to ensure
fully developed turbulent flows. Water mass flow rate was controlled by a pressure panel
and measured by an A&D ET-300B electronic balance, which has a capability of 310g ×
0.01g. Liquid jet velocity and resulting momentum rate was calculated based on this
measured mass flow rate. Flow Number (FN, in mm2) for each injector was calculated
from the measured mass flow rate and pressure differential between jet and ambient air.
FN is defined as fuel flow rate in kg/s divided by the square root of the product of fuel
differential-pressure in Pa and fuel density in kg/m3.
.
FN = 10 6
m[kg / s ]
ρ [kg / m ] ∆P[ pa ]
3
[ mm 2 ]
(B- 1)
181
In order to visualize the fast motion of jet breakup and atomization in the test
section, a green beam of 514.5 nm wavelength from an Argon laser (Coherent Innova 70)
was spread into a thin sheet by mirrors and lens. Live images were captured by an SVHS camcorder (Panasonic Model Ag-450U) of which the charged coupled devices
(CCD) have 360,000 pixels. This CCD gives a spatial resolution of approximate 300
micron at the experimental optical settings.
Three vane-type swirlers with thin vanes of different constant chord and
angle were designed in accordance with the dimensions calculated in a computational
study17 to produce swirling flow with a recirculation zone. Swirl numbers calculated
based on the following equation are 0.49, 0.86 and 1.48, corresponding to the swirler
vane angles 30°, 45° and 60°, respectively, Table B- 2, where Rh is hub radius, Rv is Vane
radius.
2 1 − ( Rh / Rv ) 3
] tan α
SN = [
3 1 − ( Rh / Rv ) 2
(B- 2)
Table B-2 Swirler configurations
Vane angle α, °
30
45
60
Rh, mm
18.3
18.3
18.3
Rv, mm
27.9
26.15
26.15
SN
0.49
0.86
1.48
182
Table B-3 summarizes the test conditions. Experiments were run at air velocities
ranging from 6.6 m/s to 20.8 m/s, and liquid injection velocities ranging from 7.8 m/s to
26.2 m/s. The resulting air-liquid momentum rate ratios varied from 5.94 to 66.6.
Table B-3 Experimental conditions
D, mm
0.344
0.515
0.60
0.84
1.19
α = 30°
13.99 - 19.79
12.87 - 17.63
12.87 - 17.63
11.58 - 15.68
7.79 - 10.11
Vjet, m/s
α = 45°
16.41 - 26.18
15.34 - 20.52
11.82 - 18.16
10.40 - 14.60
8.32 - 11.62
α = 60°
17.84 - 24.24
14.15 - 19.31
12.95 - 17.05
10.40 - 14.25
8.51 - 10.89
Since the jet diameter becomes an important parameter affecting the atomization
process, in this study the momentum rate (ρAV2, in units of N) is used instead of
momentum flux (ρV2, in units of N/m2), which is used conventionally in most other
studies. The momentum rate is defined as the transfer of momentum per unit time. The
velocity of the free air stream before entering the swirler was used to calculate the air
momentum rate.
B.4. Results and Discussion
The purpose of these experiments was to assess the impacts of the key
parameters on jet atomization in confined swirling flow. As such, criteria for defining the
atomization quality had to be established. As mentioned, it is very important in LDWI to
183
produce uniform and rapid atomization of the fuel in order to reduce NOx emission.
Therefore, in the present study, optimum atomization was defined as an atomization
where, by quick visual, relatively uniform distribution of droplets could be observed
within 50.8 mm downstream of the swirler. Flow visualization and image observation
were applied to analyze the instantaneous motion of jet breakup and atomization
phenomena in r-θ plane of the circular tube.
Figure B-3 shows instantaneous photographs of atomization phenomena of the
same injector with diameter D = 0.515 mm at different inclination angles at three
•
•
different planes with SN = 0.86 (α = 45°), M air = 0.889 N and M jet = 0.086 N. The
photographs show the typical effects of the liquid jet inclination angles at the r-θ plane on
the atomization phenomena. The jet atomization phenomena were very sensitive to the
liquid jet inclination angle in the r-θ plane. Misalignment of an injector can cause an
unbalanced impingement of liquid particles onto the inner wall of the test section, Figure
B-3 A and C. In order to find the best injection angle, tests were done for a relatively
wide ranges until the θ = 35° was identified as the optimum inclination angle. It was also
found that within ±1° of this angle, the distribution of droplets remained relatively
uniform.
Experiments were conducted for all five injector sizes of interest to check if the
injector diameter has an influence on the optimum angle. All experiments show similar
results, namely, there is an optimum inclination angle for each specified swirler
regardless of injector size. With the swirler vane angle α = 45°, for each diameter of
injector, a relatively uniform distribution of droplet can be reached within 25.4 mm
184
downstream of injection at the jet inclination angle θ = 35°, Figure B-3 B. With smaller
inclination angle (i.e., θ = 30°) and larger angle (i.e., θ = 40°) in Figure B-3 A and C,
non-uniform droplet distributions were observed 25.4 mm downstream of the injection.
Similar results were obtained for different swirlers. The optimum inclination angles were
found as θ = 42° and 32° for α = 60° and α = 30°, respectively.
A: θ = 30˚
B: θ = 35˚
C: θ = 40˚
x/X=0.10
x/X=0.15
x/X=0.20
Figure B-3 The effect of injection angle on atomization at three different axial
•
•
locations with SN = 0.86, M air = 0.889 N and M jet = 0.086 N.
Column B of Figure B-3 also shows the evolution of disintegration processes of a
liquid jet with respect to distance from the injection plane under constant air-to-liquid
185
momentum rate ratios (Rmom). Primary breakup and peeling-off of ligaments from the
surface of the jet took place right after the injection (x/X = 0.10). Near the liquid injector
exit, the liquid jet was bent in the tangential direction and downstream. When the jet
underwent more downstream bending and disturbance on the surface, it was transformed
into a liquid column with larger surface area by the dynamic pressure of swirling air (x/X
= 0.15). The shape transformation caused an increasing drag area and larger interaction
between the liquid column and swirling air. Therefore, liquid ligaments were peeled off
from the core of the jet and evolved into smaller segments and particles while traveling
along the streamlines of swirling airflows. The ligaments broke down into discrete
segments, forming irregular sized droplets, as they moved into the fully atomized region.
Various drop sizes were distributed both in the breakup and fully atomized regions.
A
B
C
D
E
Figure B-4 The effect of liquid jet momentum and air momentum on the mixing
•
•
(D = 0.840 mm, SN = 0.86, θ = 35˚, x/X = 0.2). (A) M air = 0.889 N, M jet =
•
•
•
•
0.093 N; (B) M air = 0.889 N, M jet = 0.125 N; (C) M air = 0.889 N, M jet = 0.135
•
•
•
•
N; (D) M jet = 0.118 N, M air = 0.622 N; (E) M jet = 0.118 N, M air = 0.889 N.
•
In A, B and C of Figure B-4, three different liquid jet momentum rates ( M jet =
•
0.093 N, 0.125 N and 0.135 N) were applied under the same air momentum rate M air =
186
0.889 N. Because of insufficient or excessive liquid momentum in A and C, respectively,
the liquid jets were directed towards the wall and non-uniform droplet distributions were
formed.
•
In D and E of Figure B-4, the liquid momentum rate was fixed as M jet = 0.118 N,
•
•
and two different air momentum rates ( M air = 0.622 N and M air = 0.889 N) were applied.
•
•
The effect of M air / M jet on droplet distribution was obvious. All these results suggested
that there should be an optimum momentum rate ratio of air to liquid for these specified
conditions (D = 0.840 mm, SN = 0.86, θ = 35˚) in order to achieve a fast and uniform
•
•
atomization. Extended experiments were conducted to find out the optimum M air / M jet for
each injector for each swirler.
A
B
C
D
E
•
Figure B-5 Optimum atomization at the same air-liquid momentum rate ratio ( M air =
•
•
9.54 M jet , D = 0.60 mm, SN = 0.86, x/X = 0.2, θ = 35˚). (A) M air = 0.889 N,
•
M jet
•
•
•
•
= 0.093 N; (B) M air = 0.854 N, M jet = 0.088 N; (C) M air = 0.753 N, M jet =
•
•
•
•
0.079 N; (D) M air = 0.622 N, M jet = 0.066 N; (E) M air = 0.504 N, M jet = 0.054 N.
Figure B-5 also shows an example of the effect of air momentum (or air velocity)
on the atomization. Air flows with high momentum rate (i.e., Figure B-5A) generated
187
faster radial dispersion and produced smaller size particles in the interesting planes. A
liquid jet kept its shape after it exited the injector until certain breakup criteria were
satisfied. The breakup criteria varied with the flow details inside the injector as well as
air flows outside the injector.
Increased air momentum rate enhanced the surface
interactions between the air and the liquid jet, which made it easier to attain the breakup
criteria. These results verified that the jet breakup and disintegration process was very
much dependent on the air motion as long as the necessary momentum rate ratio of air to
jet was satisfied. Swirling air flows with higher momentum accelerated jet breakup more
effectively and smaller size particles were expected.
It was found that if the inclination angle was fixed at this optimum angle, an
optimum momentum rate ratio of air to jet was able to be found for each injector to make
atomization uniform and rapid. An example is given in Figure B-5. With diameter D =
•
•
0.60 mm, SN = 0.86, θ = 35˚, and M air = 9.54 M jet , tests were conducted based on the
•
•
following procedure: with a fixed air momentum rate, i.e., M air = 0.889 N in A, M jet was
•
changed until an optimum atomization was found, M jet = 0.093 N in this case; then, the
•
air momentum rate was changed to M air = 0.854 N in B, and following the same
•
•
procedure in A, another optimum M jet was found as M jet = 0.088 N; and so on, ultimately,
•
•
an optimum M air / M jet = 9.54 was found for all six air momentum rates. Using the same
•
•
method, M air / M jet = 13.47, 10.28, 7.52 and 5.94 were found for D = 0.344 mm, 0.515 mm,
•
0.84 mm and 1.19 mm, respectively. Figure B-6B shows the plotted relation of M air and
•
M jet
for optimum atomization using this swirler with SN = 0.86.
188
2.4
SN = 0.49
2.0
Mair (N)
1.6
D=
D=
D=
D=
D=
1.2
0.8
1.19 mm
0.84 mm
0.60 mm
0.515 mm
0.334 mm
0.4
0.0
0
0.04
A)
0.08
0.12
Mjet (N)
1.0
SN = 0.86
Mair (N)
0.8
0.6
D=
D=
D=
D=
D=
0.4
0.2
1.19 mm
0.84 mm
0.60 mm
0.515 mm
0.334 mm
0.0
0
0.04
B)
0.08
0.12
0.16
Mjet (N)
0.5
SN = 1.48
Mair (N)
0.4
0.3
0.2
0.1
6
D=
D=
D=
D=
D=
1.19 mm
0.84 mm
0.60 mm
0.515 mm
0.334 mm
0.0
0
C)
0.04
0.08
0.12
0.16
Mjet (N)
Figure B-6. Correlation between nozzle diameter and air-liquid momentum
rate ratio with A) SN = 0.49, B) SN = 0.86, C) SN = 1.48
189
Similar results were also found for swirlers with SN = 0.49 and SN = 1.48, as
•
•
shown in Figure B-6A and Figure B-6C, respectively. Correlations of M air and M jet for
•
•
optimum atomization with SN = 0.49 are plotted in Figure B-6A. M air / M jet = 66.6, 36.7,
33.9, 26.0 and 20.6 were found for injector diameter D = 0.344 mm, 0.515 mm, 0.60mm,
•
•
0.84 mm and 1.19 mm, respectively. Figure B-6C shows the correlations of M air and M jet
•
•
for optimum atomization under SN = 1.48, which are M air / M jet = 8.0, 5.60, 5.21, 3.88
and 3.16 for injector D = 0.344 mm, 0.515 mm, 0.60 mm, 0.84 mm and 1.19 mm,
respectively.
A
B
D
C
E
•
Figure B-7. Effect of injector diameter (SN = 0.86, θ = 35˚, M air = 0.889 N, x/X = 0.2):
•
•
(A) D=0.344mm, M jet = 0.068 N; (B) D=0.515mm, M jet = 0.088 N; (C)
•
•
D=0.60mm, M jet = 0.093 N; (D) D=0.84mm, M jet = 0.118 N; (E) D=
•
1.19mm, M jet = 0.15 N
It was found that injector diameter has an important effect on jet breakup and
atomization. An example is given in Figure B-7, which shows the mixing images of
190
•
different injector with same air momentum rate, M air = 0.889 N, swirler number, SN =
0.86 and inclination angle θ = 35˚. With the increase of injector diameter, more ligament
and large size particles were observed at 25.4 mm downstream of the injection (i.e.,
Figure B-7 E). As already shown in Figure B-6A, B-6B and B-6C, linear correlations of
momentum rate ratio of air to jet were found as a function of the injector diameters and
swirl number. Under each swirler configuration, the coefficient of Rmom increases with
the decreasing of diameter. These correlations indicated that in LDWI, the injector
diameter plays an important role on the atomization.
It was also observed in our experiments that the swirl number had strong impact
on atomization phenomena. Stronger swirling air motion enhanced surface interactions
between air and liquid and caused a faster jet breakup and produced smaller size particles.
A liquid jet in a higher intensity swirling flow had a much quicker spread-out. However,
the pressure drop through the swirler increased with the swirl number, which directly
affected the maximum air momentum rate the test system can provide. In our case, all
experimental air was supplied by a 3.5 kW blower; increasing the swirl number from 0.49
to 1.48 decreased the maximum air momentum rate from 2.373 N to 0.432 N, almost an
82% drop.
Although under the same air momentum rate (0.432 N), the effect of
increasing swirler intensity on enhancing jet breakup was obvious. It is noted that use of
a large swirl number in future gas turbine combustors is not always desirable due to the
pressure drop.
The above discussions show that some parameters, such as injection angle,
momentum rate ratio of air to liquid, jet diameter and swirl number, have important
influences on the jet breakup process. For design and preliminary calculations of LDWI
191
combustion chambers and other applications, it is necessary to have reliable correlations
for evaluating the atomization phenomena under the criteria of quick and uniform
atomization as a function of these parameters.
The following equations were found for each swirler configuration by applying
the term of ( Dr )1.5 to each linear relation of Figure B-6A, B-6B and B-6C. The modified
D
correlations are shown in Figure B-8A, B-8B and B-8C, where Dr = 1 mm is as a
reference diameter employed to cancel out units in the right side of the equations.
•
α=30°
M air
•
M
•
•
M
•
•
M
= 6.70 ( DDr )1.5
(B-4)
= 3.57( DDr )1.5
(B-5)
jet
M air
α=60°
(B-3)
jet
M air
α=45°
= 23 .3( DDr )1.5
jet
Equations (B-3), (B-4) and (B-5) indicate the correlation between optimum
•
•
atomization and air momentum rates ( M air ), jet momentum rates ( M jet ) and injector
diameter (D). The fact that swirler configurations play an important role in the jet
breakup process leads us to the following correlation, being expressed as equation (B-6)
and plotted in Figure B-9.
•
M
jet
•
M
= 0 . 41 (
air
= 0 . 37 (
D
)
D r
D
)
D r
2 / 3
( SN
)1
/ 4
(sin
α )3
R
R
4
[
R
1 − (
R
1 − (
2 / 3
(cos
α )
−1 / 4
(sin
α ) 13
/
h
)3
]1
v
h
v
)
2
/ 4
(B-6)
192
2.8
Dr = 1 mm
SN = 0.49
2.4
Mair (D / Dr) 2/3
2.0
Mair (D / Dr)2/3 = 23.3 Mjet
1.6
D = 1.19 mm
D = 0.84 mm
D = 0.60 mm
D = 0.515 mm
D = 0.344 mm
1.2
0.8
0.4
0.0
0
0.04
A)
0.08
0.12
Mjet (N)
1.2
Dr = 1.00 mm
SN = 0.86
Mair (D / Dr)
2/3
1.0
0.8
Mair (D / D r)
2/3
= 6.7 Mjet
0.6
D = 1.19 mm
D = 0.84 mm
D = 0.60 mm
D = 0.515 mm
D = 0.344 mm
0.4
0.2
0.0
0
0.04
B)
0.08
0.12
0.16
0.12
0.16
Mjet (N)
0.6
Dr = 1.00 mm
SN = 1.48
Mair (D / Dr) 2/3
0.5
0.4
Mair (D / Dr) 2/3 = 3.57Mjet
0.3
0.2
0.1
0.0
0
C)
0.04
0.08
Mjet (N)
Figure B-8. Modified Correlation between nozzle diameter and air-liquid
momentum rate ratio
193
•
In Figure B-9, the vertical axis indicates the calculated M jet based on equation (B•
6) and tested M air shown in Figure B-6 and the horizontal axis indicates the measured
•
experimental M jet .
0.16
D r = 1.0 mm
(N)
0.12
Mjet, Calc.
0.08
SN = 1.48
SN = 0.86
SN = 0.49
0.04
0.00
0
0.04
0.08
M jet, Act.
0.12
0.16
(N)
Figure B-9. Generalized correlation of air/liquid momentum rate ratio
for optimum atomization
Similar to the linear correlation between air momentum rate and liquid jet
momentum rate, the air mass flow rate also has a linear relation with liquid jet mass flow
rate for each specified injector diameter for optimum atomization. For example, for swirl
number SN = 1.49, with the increase of diameter from D = 0.344 mm to D = 1.19 mm,
the mass flow rates of air to jet decreases from 27.1 to 3.93, which indicates larger liquid
jet mass flow rates are needed for larger injector diameter. In order to achieve a lean
optimum mixture, it is necessary to use a certain range of injector diameter.
The
194
correlation of mass flow rate between air and liquid jet will finally decide the equivalence
ratios in a real combustor. Choice of injector diameter should consider a combination of
droplet size (small diameter may produce small size droplet) and equivalence ratio.
Multiple injectors may be used to maximize the particle dispersion if a certain droplet
size range is desired, and to reach the desired equivalence ratio.
B.5. Closure
The breakup and subsequent atomization of radially injected liquid jets in a
swirling flow were investigated by using laser-based flow visualization. The current
experimental investigation of liquid jet injection into a confined swirling flow over a
wide range of parameters allows us to make the following conclusions:
(1) For LDWI, the atomization phenomena are sensitive to the parameters such as jet
inclination angle, momentum rate ratio of air to jet, swirl number, and injector
diameter.
(2) There are optimum jet inclinations angles at which uniform atomization can be
quickly reached. In this study, optimum jet inclination angles were found as 32°,
35° and 42° under swirler vane angles α = 30°, 45° and 60°.
(3) For the three different swirler configurations tested, each injector exhibited a linear
relations between air momentum rate and liquid jet momentum rate for optimum
atomization. Five different injectors were tested for each swirler and a modified
correlation that collapsed the data for each swirler was found.
195
(4) It was possible to develop a generalized correlation between air momentum rate and
liquid jet momentum rate for optimum atomization based on swirler configuration
and injector diameter.
The atomization phenomena of liquid jets in LDWI were provided in the present
study. However, the discussions were only based on image observation and hence are
qualitative. Further image processing and analysis is needed to quantitatively determine
the atomization parameters in a confined geometry, such as centrality of particles, degree
of spread of particles and total area ratio of particles. Droplet size also has a strong
impact on the combustor design. Further cold flow experiments need to be conducted at
high air pressure and velocity to verify the accuracy of the correlations of the present
study before they can be considered adequate for LDWI combustion at high velocity and
high temperature.
B.6. Literature Cited
1
Tacina, R. R., “Combustor Technology for Future Aircraft,” AIAA Paper 90-2400,
July 1990.
2
Gupta, A. K., and Lilley, D. G., “Combustion and Environmental Challenges for Gas
Turbines in the 1990s,” Journal of Propulsion and Power, Vol.10, No.2, 1994, pp.137147.
3
Choi, K.J. and Tacina, R. R., “Lean Direct Wall Fuel Injection Method and Device,”
US Patent (5,680,765), Oct. 1997.
4
Ahmad, N. T., Andrews, G. E., Kowkabi, M., and Sharif, S. F., “Centrifugal
Atomization in Gas and Liquid Fuelled Lean Swirl Stabilized Primary Zones,” Int. J.
Turbo and Jet Engines, Vol. 3, 1986, pp. 85-92.
196
5
Huh, J. Y., “Studies on the Atomization of Liquid Jets and Pre-atomized Spray in
Confined Swirling Air Flows for Lean Direct Injection Combustion,” Ph. D thesis,
Drexel University, July, 1998.
6
Choi, K. J., Huh, J. Y., and Tacina, R. R., “Study on Well-Stirred Atomization of
Liquid Droplets in a Lean Direct Injection Mode,” 11th Annual Conference on Liquid
Atomization and Spray Systems, ILASS-America 98, Sacramento, CA, 1998,
pp.
273-277
7
Tacina, R. R., Way, C., and Choi, K. J., “Flame Tube NOx Emissions Using a LeanDirect-Wall-Injection Combustor Concept,” AIAA Paper 2001-3271, July 2001.
8
Syred, N., and Beer, J. M., “Combustion in Swirling Flows: A Review,” Combustion
and Flame, Vol. 23, 1974, pp. 143-201.
9
Ahmed, S. A., and Nejad, A. S., “Velocity Measurements in a Research Combustor
Part 1: Isothermal Swirling Flow,” Experimental Thermal and Fluid Science, Vol. 5,
1992, pp. 162-174.
10
Sheen, H. J., Chen, W. J., and Jeng, S. Y., “Recirculation Zones of Unconfined and
Confined Annular Swirler Jets”, AIAA Journal, Vol. 34, No. 3, 1996, pp. 572-579.
11
Young, D. L., Liao, C. B., and Sheen, H. J., “Computations of Recirculation Zones
of a Confined Annular Swirler Flow,” Int. J. Numer. Meth. Fluids, Vol. 29, 1999, pp.
791-810.
12
Inamura, T., and Nagai, N., “Spray Characteristics of Liquid Jet Traversing Subsonic
Airstreams,” Journal of Propulsion and Power, Vol.13, No. 2, 1997, pp. 250-256
13
Baranovsky, S. I., and Schetz, J. A., “Effect of Injection Angle on Liquid Injection in
Supersonic Flow,” AIAA Journal, Vol. 18, No. 6, 1980, pp. 625-629.
14
Chen, T. H., Smith, C. R., Schommer, D. G., and Nejad, A. S., “Multi-Zone
Behavior of Transverse Liquid Jet in High-Speed Flow,” AIAA Paper, 93-0453, Jan.
1993.
15
Laredo, D., Levy, Y., and Timna, Y. M., “Study of Two-Phase Flow for a Ramjet
Combustor,” Journal of Propulsion and Power, Vol. 7, No. 5, 1991, pp. 724-731.
16
Fuller, R. P., Wu, P., Kirkendall, K. A., and Nejad, A. S., “Effects of Injection Angle
on Atomization of Liquid Jets in Transverse Airflow,” AIAA Journal, Vol. 38, No. 1,
2000, pp. 64-72.
17
Xin, J., and Choi, K. J., “Study of the Velocity Field and Droplet Distribution in a
Confined Swirl Flow,” 5th ILASS-Americas, San Ramon, CA 1992.
197
VITA
Xiaohui Gong was born on January 11, 1971 in Yiwu, Zhejiang Province, P. R.
China. He spent his childhood in Jinhua, Zhejiang and graduated from Jinhua No.1 high
school in 1989.
Xiaohui Gong attended Tianjin University at Tianjin, P.R. China as an
undergraduate student majoring in Mechanical Engineering, and with the specialization
in Internal Combustion Engines. He obtained his Bachelor of Engineering degree in
1993. After one year’s work in a ship manufacturing factory, Xiaohui Gong went back to
the same school for his Master degree. He received his Master of Engineering degree in
March 1997, with an emphasis on diesel engine combustion and after-treatments. Since
then Xiaohui Gong worked in National Engine Combustion Laboratory located in
Tianjing University as a mechanical engineer.
In the fall of 1999, Xiaohui Gong came to the United States for his Ph. D. degree
in Mechanical Engineering at Drexel University. He started with a study on a new gas
turbine combustor concept, and then he spent most of his efforts on the study of the
autoignition and oxidation of hydrocarbon mixtures. He has co-authored 7 technical
papers (with one currently in press). He is a member of SAE and AIAA.
After receiving his Ph. D. degree, Xiaohui Gong plans to work in industry in the
area of IC engine.
198