PengYe Dissertation 2011

The Pennsylvania State University
The Graduate School
John and Willie Leone Family Department of Energy and Mineral Engineering
INVESTIGATION OF IMPACT OF FUEL INJECTION STRATEGY AND
BIODIESEL FUELING ON ENGINE EMISSIONS AND PERFORMANCE
A Dissertation in
Energy and Mineral Engineering
by
Peng Ye
© 2011 Peng Ye
Submitted in Partial Fulfillment
of the Requirements
for the Degree of
Doctor of Philosophy
December 2011
The dissertation of Peng Ye was reviewed and approved* by the following
André L. Boehman
Professor of Fuel Science and Materials Science and Engineering
Dissertation Advisor
Chair of Committee
Randy L. Vander Wal

Associate Professor of Energy and Mineral Engineering and Materials Science
and Engineering
Daniel C. Haworth
Professor in Mechanical and Nuclear Engineering
Yaw D. Yeboah
Professor and Department Head of Energy and Mineral Engineering
R. Larry Grayson
Professor of Energy and Mineral Engineering
George H., Jr., and Anne B. Deike Chair in Mining Engineering
Graduate Program Officer of Energy and Mineral Engineering;
*Signatures are on file in the Graduate School
ii
Abstract
Both biodiesel fueling and changes of fuel injection pressure have significant
impacts on diesel engine emissions. The investigations of their impacts on engine exhaust
NOx and particulate matter emissions were conducted with an 8-cylinder common-rail
turbocharged direct injection diesel engine using ultra low sulfur diesel fuel and soybean
methyl ester (SME) – based biodiesel blends. The engine was running at moderate speed
and different loads. Three fuel injection parameters: start of injection, fuel injection
pressure and fuel injection duration were investigated to investigate their impact on
engine emissions. With the control of fuel injection strategy, it is shown in this work that
the biodiesel engine NOx emission penalty can be eliminated.
A fuel spray, mixture stoichiometry field and lift-off length model was employed
to explain the variations of NOx emission from biodiesel fueling and change of fuel
injection strategy. Linear correlations between the average oxygen equivalence ratio of
the fuel-air mixture at the autoignition zone near the lift-off length and brake specific
NOx emissions were observed for all load conditions, regardless of fuel type. This
confirms that the dominant factor that determines NOx emissions is the ignition event
controlled by the oxygen equivalence ratio at the autoignition zone.
The impact of late in-cylinder (post) injection combustion with biodiesel on
lubricating oil dilution was investigated in this work. It is shown that this injection
strategy could effectively decrease engine NOx emissions, while increase the CO and
unburned hydrocarbon emissions. The lubricating oil dilution depends on the post
injection timing: an increase in the lubricating oil dilution can be only observed if the
post injection timing is later than 45º after top dead center.
iii
The impacts of fuel injection pressure on diesel and biodiesel soot morphology
and oxidative reactivity were investigated. It is shown that compared with engine
condition and fuel injection pressure, biodiesel has much less significant impact on soot
morphology. For soot oxidative reactivity, it is found that both diesel and biodiesel soot
from higher fuel injection pressure have higher reactivity, and biodiesel soot has higher
reactivity than diesel soot when both of them are obtained from the same injection
pressure.
The optimized apparent heat release pattern for improved engine thermal
efficiency was investigated with a zero-dimensional engine thermodynamic simulation.
The results suggest that the optimized apparent heat release is a “wide and low” peak.
The reason for this kind of heat release is that it can decrease the in-cylinder temperature
and consequently decrease heat loss.
iv
Table of Contents
List of Figures ................................................................................................................... vii
List of Tables. .................................................................................................................. xiv
Nomenclature .................................................................................................................... xv
Acknowledgements ........................................................................................................ xviii
Chapter I. Introduction ................................................................................................... 1
Chapter II. Literature Review.......................................................................................... 7
2.1 Biodiesel background ............................................................................................ 7

2.2 Impact of biodiesel fueling on engine emissions ................................................ 13
2.2.1 Nitric oxides (NOx) ....................................................................................... 15
2.2.2 Particulate Matter (PM) ................................................................................. 19
2.3 Origin of biodiesel emission effect ..................................................................... 22
2.3.1 Biodiesel NOx effect...................................................................................... 22
2.3.2 Biodiesel particulate matter effect ................................................................. 27
2.4 Fuel Injection strategy with biodiesel fueling ..................................................... 29
2.5 Post injection strategy ......................................................................................... 33
2.6 Biodiesel and fuel injection pressure on PM morphology and soot oxidative
reactivity .............................................................................................................. 36
2.6.1 PM morphology ............................................................................................. 36
2.6.2 Soot oxidative reactivity ................................................................................ 40
2.7 Impact of biodiesel fueling, injection strategy and heat release pattern on Engine
efficiency ............................................................................................................. 42
Chapter III. Technical Approach .................................................................................... 49
3.1 Experiment apparatus.......................................................................................... 49

3.2 Engine emission bench: ...................................................................................... 50
3.3 Test fuel .............................................................................................................. 51
3.4 Engine test conditions ......................................................................................... 53
3.5 Biodiesel lubricating oil dilution ........................................................................ 57
3.6 Soot characterization ........................................................................................... 60
Chapter IV. Investigation of injection strategy and biodiesel fueling on diesel engine
emissions and performance ........................................................................ 64
v
4.1 Impact of injection timing and fuel injection pressure on NOx emissions ......... 64
4.2 Impact of injection timing and fuel injection pressure on particulate matter (PM)
emissions ............................................................................................................. 68
4.3 NOx-PM trade-off ............................................................................................... 72
4.4 Brake thermal efficiency ..................................................................................... 78
4.5 Apparent heat release analysis ............................................................................ 83
Chapter V. Origin of biodiesel NOx effect ................................................................... 92
Chapter VI. Investigation of late in-cylinder injection strategy on engine emissions and
lubricant fuel dilution ............................................................................... 108
6.1 Exhaust Emissions ............................................................................................ 108

6.2 Exhaust Temperature and BSFC ....................................................................... 110
6.3 Heat Release Analysis....................................................................................... 112
6.4 Oil Dilution Level, Total Acid Number (TAN) and mfMEP ........................... 120
6.5 Analysis of Hydrocarbons from Lubricant and Exhaust................................... 123
Chapter VII. Impact of fuel injection pressure on soot morphology and oxidative
reactivity ................................................................................................... 129
7.1 Morphology analysis......................................................................................... 129

7.2 Analysis of soot oxidative reactivity................................................................. 139
Chapter VIII. Optimized heat release pattern to achieve high engine efficiency ............ 149
8.1 Model formulation ............................................................................................ 149

8.2 Computational procedure .................................................................................. 156
8.3 Optimized heat release pattern probing ............................................................ 163
Chapter IX. Conclusions and recommendations for future work ................................. 193
References…. .................................................................................................................. 199
Appendix A: Matlab code ............................................................................................... 229
Appendix B: Repeatability of thermogravimetric analysis ............................................. 236
Appendix C: Repeatability of Raman analysis .............................................................. 239
Appendix D: Repeatability of Morphology analysis ...................................................... 240
Appendix E: Repeatability of particle size distribution (SMPS) result .......................... 242
vi
List of Figures
Figure II-1: A diagram representing the transesterification process, replicated from [6].
The general scheme of the process is to modify a triglyceride (a fat or oil
which is a glyceride with three fatty acids, represented as R1, R2 andR3)
into a fatty acid methyl ester. ....................................................................... 8
Figure II-2: NOx, PM, CO and HC % changes as biodiesel blending percentage increase
as determined by the EPA through statistical regression of publically
available data on highway heavy-duty truck engines [18, 51]. Taken from
[44] ............................................................................................................. 14
Figure II-3: Diesel soot agglomerate composed of spherical primary particles, taken from
Ref. [168].................................................................................................... 37
Figure II-4: A sketch of limit-pressure cycle .................................................................... 46
Figure II-5: Indicated efficiency map with α and β .......................................................... 47
Figure III-1: FAME composition of B100 soybean oil methyl ester ................................ 52
Figure III-2: Molecular structures[217], names, abbreviations, chemical formulas, and

Chemical Abstracts Service (CAS) numbers of compounds comprising the
biodiesel used in this work. ........................................................................ 53
Figure III-3: Thermophoretic sampling apparatus ............................................................ 57
Figure III-4: FTIR spectrums for oils with different biodiesel dilution............................ 58
Figure III-5: Correlation between absorbance of Peak 1746 cm-1 from FTIR and dilution
level ............................................................................................................ 59
Figure III-6: Demonstration of curve fitting for Raman spectrum ................................... 63
Figure IV-1: NOx emissions at 25% load. Four injection timings (9˚, 5˚, 1˚ before top
dead center and 3˚ after top dead center) were tested for both ultra low
sulfur diesel (ULSD) and B40 blend. Three fuel injection pressures (52, 60
and 80 MPa) were tested at each injection timing...................................... 65
Figure IV-2: NOx emissions at 50% load. Four injection timings (9˚, 7˚, 5˚, 3˚ 1˚ before
top dead center and 1˚, 3˚ after top dead center) were tested for both ultra
low sulfur diesel (ULSD) and B40 blend. Three fuel injection pressures (72,
90 and 108 MPa) were tested at each injection timing for ULSD, and six
fuel injection pressures (72, 80, 90, 98, 108 and 118 MPa) were tested at
each injection timing for B40. .................................................................... 66
vii
Figure IV-3: NOx emissions at 75% load. Four injection timings (9˚, 5˚, 1˚ before top
dead center and 3˚ after top dead center) were tested for both ultra low
sulfur diesel (ULSD) and B40 blend. Three fuel injection pressures (87,
109 and 131 MPa) were tested at each injection timing............................. 67
Figure IV-4: PM emissions at 25% load. Four injection timings (9˚, 5˚, 1˚ before top dead
center and 3˚ after top dead center) were tested for both ultra low sulfur
diesel (ULSD) and B40 blend. Three fuel injection pressures (52, 60 and 80
MPa) were tested at each injection timing. ................................................ 69
diesel (ULSD) and B40 blend. Three fuel injection pressures (72, 90 and
108 MPa) were tested at each injection timing. ......................................... 71
diesel (ULSD) and B40 blend. Three fuel injection pressures (87, 109 and
131 MPa) were tested at each injection timing. ......................................... 72
Figure IV-7: NOx-PM emission trade-off at 25% load. The legend shows the fuel and
injection timing for each test. Three fuel injection pressures (52, 66 and 80
MPa) were tested at each injection timing for each fuel. Fuel injection
pressure is not distinguished in the plot. .................................................... 73
injection timing for each test. Three fuel injection pressures (72, 90 and
108 MPa) were tested at each injection timing for each fuel. Fuel injection
pressure is not distinguished in the plot ..................................................... 74
injection timing for each test. Three fuel injection pressures (87, 109 and
131 MPa) were tested at each injection timing for each fuel. Fuel injection
pressure is not distinguished in the plot ..................................................... 75
Figure IV-10: PM-NOx emission trade-off at the start of injection timing at 9˚ before top
dead center for ultra low sulfur diesel and B40. (a) 25% load; (b): 50% load;
(c): 75% load; (d): 75% load after Soxhlet of PM filters ........................... 77
Figure IV-11: Brake thermal efficiency at 25% load for ultra low sulfur diesel (left) and
B40 (right). The multiple points at each start of injection timing are due to
the multiple fuel injection pressures. .......................................................... 79
viii
Figure IV-14: Apparent heat release profile of different fuels and different fuel injection
pressures at 25% load. The SOI: a) 9˚BTDC, b) 5˚BTDC, c) 1˚BTDC, d)
3˚ATDC. ..................................................................................................... 85
Figure IV-15: Apparent heat release profile of different fuels and different fuel injection
pressures at 50% load. The SOI: a) 9˚BTDC, b) 5˚BTDC, c) 1˚BTDC, d)
3˚ATDC. ..................................................................................................... 88
Figure V-2: Oxygen equivalence ratio (φΩ) field for ULSD and B40 at different injection
pressures at the SOI of 7ºBTDC, mid-load ................................................ 99
Figure V-3: Oxygen equivalence ratio (φΩ) field for ULSD and B40 at different injection
pressures at the SOI of 9ºBTDC, mid-load .............................................. 100
Figure V-4: The correlation between brake specific NOx emissions and average oxygen
equivalence ratio of the fuel-air mixture at lift-off length at 25% load,
regardless of fuel type. The legend: xB is x degree before TDC; xA is x
degree after TDC. ..................................................................................... 102
degree after TDC. ..................................................................................... 103
degree after TDC. ..................................................................................... 104
Figure V-7: semi-ln plot of absolute value of slope vs. SOI .......................................... 105
Figure V-8: semi-ln plot of Y-axis intercept (at X=0) vs. SOI ....................................... 106
Figure VI-1: Emissions of B20 engine test with various post injection strategies. Samples
are differentiated with post injection conditions while rest parameters were
kept the same. The fuel for the non-ULSD test was B20. ........................ 109
Figure VI-2: Exhaust temperature and BSFC profile of the 100-hour test. Each point is
the average of one test period. The dash lines are used to divide zones of
each post injection strategy. ..................................................................... 111
Figure VI-3: Pressure traces of test with different post injection strategies at both TDC
(top) and late crank angle (bottom) .......................................................... 113
ix
Figure VI-5: Bulk cylinder temperature of tests with different post injection strategies.
The arrow points to the lowest acceptable combustion temperature. ....... 117
Figure VI-7: GC-MS results of diesel, B100, lighter cut from 432˚C distillation of fresh
and used lubricant (top); Region of interest in the GC-MS results (bottom).
The numbers at the diesel zone means the number of carbon in relevant
paraffin. .................................................................................................... 124
Figure VI-8: GC-MS results of hydrocarbons collected from exhaust gases for various
post injection strategies ............................................................................ 126
Figure VI-9: FAME composition of B100 and of the biodiesel in used lubricant. ........ 127
Figure VII-1: A transmission electronic microscopy (TEM) image of diesel soot at 30%
load and 50 MPa fuel injection pressure .................................................. 130
Figure VII-2: A transmission electronic microscopy (TEM) image of B20 soot at 30%
load and 50 MPa fuel injection pressure .................................................. 131
Figure VII-3: Primary particle sizes for both diesel and B20 soot ................................. 133
Figure VII-4: Fractal dimensions for both diesel and B20 soot...................................... 134
Figure VII-5: Number of primary particles for both diesel and B20 soot ...................... 135
Figure VII-6: Particle size distribution of diesel and B20 combustion exhaust, 30% load,
before thermodenuder............................................................................... 136
after thermodenuder ................................................................................. 137
before thermodenuder............................................................................... 138
after thermodenuder ................................................................................. 139
Figure VII-10: Thermogravimetric analysis result of 30% load diesel and B20 soot, m/m0
indicates the residual mass weight percentage. ........................................ 140
Figure VII-11: Thermogravimetric analysis result of 60% load diesel and B20 soot, m/m0
indicates the residual mass weight percentage. ........................................ 141
Figure VII-12: Apparent oxidative reactivity of 30% load diesel and B20 soot ............ 143
Figure VII-13: Apparent oxidative reactivity of 60% load diesel and B20 soot ............ 144
Figure VII-14: AD1/Ag ratio with fuel injection pressure for both diesel and B20 soot . 146
x
Figure VII-15: D1 FWHM values with fuel injection pressure for both diesel and B20
soot ........................................................................................................... 147
Figure VIII-1: comparison of simulated motor pressure trace with engine data, running at
1500 rpm, 30% load, single injection timing at 1º before top dead center,
100 MPa fuel injection pressure ............................................................... 158
Figure VIII-2: Representation of apparent heat release (AHR) at 1500 rpm, 30% load, 1º
before top dead center injection, 100 MPa fuel injection pressure with a
convolution of two peaks: one Gaussian peak and one ExpConvExp peak.
Top: difference between the engine data and fit curve; middle: comparison
of engine data and fit curve; bottom: deconvolution of artificial peaks. .. 159
Figure VIII-3: Representation of apparent heat release (AHR) at 1500 rpm, 60% load, 1º
before top dead center injection, 100 MPa fuel injection pressure with a
convolution of two peaks: one Gaussian peak and one ExpConvExp peak.
Top: difference between the engine data and fit curve; middle: comparison
of engine data and fit curve; bottom: deconvolution of artificial peaks. .. 160
Figure VIII-4: comparison of simulated pressure trace with engine data, running at 1500
rpm, 30% load, single injection timing at 1º before top dead center, 100
MPa fuel injection pressure ...................................................................... 163
Figure VIII-5: Examples of apparent heat release with: h1/h1b = 0.2, h2/h2b=1.29;
h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1 .................................. 166
Figure VIII-6: Effect of start of combustion on indicated thermal efficiency, h1/h1b = 0.2,
h2/h2b=1.29; h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1 ............ 167
Figure VIII-7: Effect of start of combustion on brake thermal efficiency, h1/h1b = 0.2,
h2/h2b=1.29; h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1 ............ 168
Figure VIII-8: Examples of apparent heat release with: h1/h1b = 1, h2/h2b=1; h1/h1b = 2,
h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4, h2/h2b=0.02 .......... 169
Figure VIII-9: Effect of start of combustion on indicated thermal efficiency, h1/h1b = 1,

h2/h2b=1; h1/h1b = 2, h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4,
h2/h2b=0.02 .............................................................................................. 170
Figure VIII-10: Effect of start of combustion on brake thermal efficiency, h1/h1b = 1,

h2/h2b=1; h1/h1b = 2, h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4,
h2/h2b=0.02 .............................................................................................. 171
Figure VIII-11: Examples of apparent heat release with: w1/w1b = 0.6, h2/h2b=1.17;
w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1 ............................... 172
Figure VIII-12: Effect of start of combustion on indicated thermal efficiency, w1/w1b =

0.6, h2/h2b=1.17; w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1 . 173
xi
Figure VIII-13: Effect of start of combustion on brake thermal efficiency, w1/w1b = 0.6,
h2/h2b=1.17; w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1 ........ 174
Figure VIII-14: Examples of apparent heat release with: w1/w1b = 1, h2/h2b=1; w1/w1b
= 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b = 4, h2/h2b=0.02 175
Figure VIII-15: Effect of start of combustion on indicated thermal efficiency, w1/w1b = 1,

h2/h2b=1; w1/w1b = 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b
= 4, h2/h2b=0.02 ...................................................................................... 176
Figure VIII-16: Effect of start of combustion on brake thermal efficiency, w1/w1b = 1,

h2/h2b=1; w1/w1b = 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b
= 4, h2/h2b=0.02 ...................................................................................... 177
Figure VIII-17: Examples of apparent heat release with different peak distances ......... 178
Figure VIII-18: Effect of start of combustion on indicated thermal efficiency with

different peak distance, ∆L = l2-l1 ............................................................ 179
Figure VIII-19: Effect of start of combustion on brake thermal efficiency with different
peak distance ............................................................................................ 180
Figure VIII-20: Examples of apparent heat release profiles for sharp peak (h1/h1b = 8)
versus low and wide peak (w1/w1b = 8) .................................................. 181
Figure VIII-21: Indicated thermal efficiency for sharp peak (h1/h1b = 8) versus low and
wide peak (w1/w1b = 8) ........................................................................... 182
Figure VIII-22: Brake thermal efficiency for sharp peak (h1/h1b = 8) versus load and
wide peak (w1/w1b = 8) ........................................................................... 183
Figure VIII-23: Heat transfer loss for sharp peak (h1/h1b = 8) versus load and wide peak
(w1/w1b = 8) ............................................................................................ 184
Figure VIII-24: Comparison of cylinder pressure for sharp peak (h1/h1b = 8) versus load
and wide peak (w1/w1b = 8) apparent heat release.................................. 185
Figure VIII-25: Indicated work for sharp peak (h1/h1b = 8) versus low and wide peak
(w1/w1b = 8) ............................................................................................ 186
Figure VIII-26: P-V diagram for sharp peak (h1/h1b = 8) versus low and wide peak
(w1/w1b = 8) ............................................................................................ 187
Figure VIII-27: Comparison of two apparent heat release (AHR) profiles with the same
start of combustion timing at 362.2º (2.2º after top dead center), indicated
efficiency of AHR1 = 49.74%, indicated efficiency of AHR2 = 47.23% 188
xii
Figure VIII-28: Comparison of cylinder pressure traces of two apparent heat release
(AHR) profiles with the same start of combustion timing at 362.2º (2.2º
after top dead center), indicated efficiency of AHR1 = 49.74%, indicated
efficiency of AHR2 = 47.23%.................................................................. 189
Figure VIII-29: Comparison of two apparent heat release (AHR) profiles with the same
start of combustion timing at 350º (10º before top dead center), indicated
efficiency of AHR1 = 50.02%, indicated efficiency of AHR2 = 50.84% 190
Figure VIII-30: Comparison of cylinder pressure traces of two apparent heat release
(AHR) profiles with the same start of combustion timing at 350º (10º
before top dead center), indicated efficiency of AHR1 = 50.02%, indicated
efficiency of AHR2 = 50.84%.................................................................. 191
Figure B-1: Repeatability of thermogravimetric analysis for diesel soot, 30% load and 75
MPa .......................................................................................................... 238
Figure E-1: Repeatability of particle size distribution of exhaust gases running with B20
at 30% load and 50 MPa fuel injection pressure ...................................... 242
xiii
List of Tables
Table I-1: Emission regulation for light and heavy duty diesel engines [3] ....................... 2
Table II-1: Chemical structure of common fatty acids, taken from [38] ............................ 9
Table II-2: Chemical composition of biodiesel derived from different feedstock, taken
from [39]..................................................................................................... 10
Table II-3: Physical and chemical properties of common fatty acid methyl esters
(FAMEs), taken from [39].......................................................................... 10
Table II-4: The effects of property differences between soy methyl ester biodiesel or its
blend and petroleum diesel on engine parameters and fuel combustion. A
positive/negative sign in the “difference” column means increase/decrease;
a positive/negative sign in the rest column means increase (or
advance)/decrease (or retarding) of the property. Taken from [44] ........... 12
Table III-1: Engine specifications ..................................................................................... 49
Table III-2: Engine operating parameters in Task 1 and 2 ............................................... 54
Table III-3: Engine operating parameters in Task 3 ......................................................... 55
Table III-4: Engine operating parameters in Task 4 ......................................................... 56
Table V-1: Input parameters during calculation ............................................................... 94
Table VI-1: Equivalence ratio ϕ, mean pressure P mean temperature T and ignition delay
τ ID calculated for each post injection case. .............................................. 118
Table VI-2: Calculated UHC emission rate and consumed HC by combustion at late
crank angle. B20 is a fuel mixture blended with 20% v/v SME based B100
and 80% v/v ultra low sulfur diesel. ......................................................... 120
Table VIII-1: Engine basic parameters used by friction model. ..................................... 155
Table C-1: Repeatability of Raman analysis for B20 soot, 30% load and 50 MPa fuel
injection pressure ...................................................................................... 239
Table D-1: Repeatability of number of primary particles and fractal dimension for B20
soot, 60% load at 75 MPa fuel injection pressure .................................... 240
Table D-2: repeatability of measurement of diameter of primary particles of diesel soot,

60% load at 125 MPa fuel injection pressure ........................................... 241
xiv
Nomenclature
AZ = autoignition zone
AHR = apparent heat release
CO = carbon monoxide
PM = Particulate Matter
NOx = Nitrogen Oxides
IVC = inlet valve close
EVO = outlet valve open
LCA = late crank angle
LOL = Lift-Off Length
l1,l2 = locations of artificial heat release peaks
h1, h2 = height of artificial heat release peaks
w1 = width of artificial Gaussian heat release peaks
l1b, l2b, h1b, h2b, w1b = location, height and width of apparent heat release at baseline
TDC = top dead center
TAN = total acid number
FAME = Fatty Acid Methyl Esters
SME = Soybean Methyl Esters
SMPS = scanning mobility particle sizer
SOI = Start of Injection
SOC = Star of Combustion
ULSD = Ultra Low Sulfur Diesel
UHC = Unburned Hydrocarbon
xv
mfMEP = mechanic fraction mean effective pressure
IMEPg = gross indicated mean effective pressure
τID = ignition delay
ϕ = overall equivalence raito
x = axial coordinate
r = radical coordinate
ϕ(x) = average equivalence ratio of a fuel spray at x
(A/F)st = stoichiometric air fuel ratio
ϕΩ = oxygen equivalence ratio
x+ = penetrate length scale of fuel jet
ρf = fuel density
ρa = air density
d = diameter of orifice
Ca = area-contraction coefficient
nC = nuber of carbon atoms in fuel
nH = number of hydrogen atoms in fuel
nO = number of oxygen atoms in reactants
H = lift-off length
Uf = velocity of injected fuel
Zst = stoichiometric mixture fraction of fuel
p = mean pressure
P = in-cylinder pressure
V = chamber volume
xvi
T = in-cylinder temperature
T = mean temperature
mɺ post = rate of the mass of post injected fuel
η = engine thermal efficiency
θ = crank angle
γ = Cp/Cv, ratio of constant pressure heat capacity and constant volume heat capacity
qt = total heat added
qw = engine wall heat transfer
xvii
Acknowledgements
I want to express my sincere gratitude to my academic advisor, Dr. André L.
Boehman, for his support, guidance, enthusiasm, patience and encouragement during this
study. Thanks to his excellent direction, I could explore new field of knowledge and
complete my Ph.D. study.
I would also like to express my gratitude to Dr. Daniel C. Haworth, Dr. Randy L.
Vander Wal and Dr. Yaw D. Yeboah for their valuable comments and suggestions to
improve the quality of my research and dissertation as my committee.
I also want to thank Dr. Magin Lapuerta Amigo from University of Castilla-La
Mancha, Spain and Dr. John R. Agudelo from Universidad de Antioquia, Colombia for
their generous help and instructions in the work of soot oxidative reactivity and engine
efficiency analysis.
I also want to thank Dr. Joseph M. Perez and his tribology group in the
Department of Chemical Engineering for their generous help in the lubricant analysis.
I also want to thank my group members at the Diesel Combustion and Emissions
lab for their help. In particular, I am grateful to Vince Zello, Greg Lilik and Bhaskar
Prabhakar for their help in the engine test stand, to Chenxi Sun for her help in soot TEM
xviii
imaging, to Eduardo Barrientos for his help in TGA operating, to Dongil Kang and
Vickey Kalaskar for their general help and discussions during my research.
I also want to thank Dr. Dania A. Fonseca for her generous help in GC analysis
and Joe Stitt for the Raman spectroscopy analysis in the Material Characterization Lab.
I want to give my special thanks to my family for their support during my work.
The financial support and valuable discussions from Infineum, USA, GE
Transportation, Volvo Powertrain and Oak Ridge National Lab are heartily
acknowledged.
xix
Chapter I. Introduction
Since it was first developed in the late 19th century, the diesel engine has been
widely used in heavy duty transportation for its higher output torque and brake thermal
efficiency than gasoline engine [1]. Through the past century, although the fundamental
design of the diesel engine hasn’t been significantly changed, the regulation for diesel
engine exhaust emissions, including carbon monoxide (CO), unburned hydrocarbon
(UHC), nitrogen oxides (NOx) and particulate matter (PM), are becoming increasingly
strict. Among these exhaust gases, the amount of CO and UHC from conventional engine
operating mode are much lower than NOx and PM, therefore the latter are more of
interest to regulators and researchers. Table I-1 presents the current and future planned
emission regulations for both light duty and heavy duty vehicles from authorities
including environmental protection agency (EPA) and California government. Significant
difficulties can be expected for current emission control technologies. For example, the
EPA Tier 2 Bin 5 NOx regulation is likely beyond the reach of today’s de-NOx
aftertreatment considering cold start, and in 2010 only 30% of current new cars in
California are able to meet the SULEV program [2]. For heavy duty transportation, very
few engines before 2010 are able to meet the US 2010 regulation [3], i.e., most on-road
engines need to be improved in emission control.
1
Table I-1: Emission regulation for light and heavy duty diesel engines [3]
NOx PM CO HCHO
(g/mile) (g/mile) (g/mile) (g/mile)
Bin 5 (2009 0.07 0.01 4.2 0.018
Light duty
effective)
(EPA Tier
Bin 4 0.04 0.01 2.1 0.011
2)
Bin 3 0.03 0.01 2.1 0.011
Bin 2 0.02 0.01 2.1 0.004
Bin 1 0 0 0 0
Light duty SULEV 0.02 0.01 1 0.004
(California) program
NOx PM CO NMHC
Heavy duty (mg/kWh) (mg/kWh) (mg/kWh) (mg/kWh)
US 2010 260 13 - 182
In addition to the emission regulation, current engine technologies are also facing
the challenge of sustainable development. Given that it is widely believed that the fossil
energies – coal and petroleum, which have supported the past 200 year’s industrial
development – may not be sufficient to satisfy the increased need for energy in the future,
studies on energy sustainability are more and more popular. Typically, sustainable
energy development involves three aspects: energy saving on demand side, replacement
of fossil fuel with renewable fuels and efficiency improvement in energy applications [4].
For engineering practice, the evaluations of applications of renewable fuels and
efficiency improvements are especially important, which is therefore the focus of the
present work. With respected to diesel engines, evaluation of biodiesel (primarily on the
emission impact as discussed previously) as a renewable fuel and engine thermal
efficiency improvements are investigated.
Biodiesel is a fuel which is comprised of monoalkyl esters of long chain fatty
acids derived from vegetable oils or animal fats [5], and it has been studied as an
alternative fuel to petroleum diesel for compression-ignition engines, both in neat form
2
(i.e., 100% esters) or in blended form with diesel fuel [6]. Compared with petroleum
diesel fuel, biodiesel has many advantages: it is renewable and biodegradable; it has a
high flash point which improves transportation safety [7]; and it can reduce exhaust
levels of some regulated emissions including CO, UHC and PM [8-11]. However, some
disadvantages of biodiesel are limiting the growth of its usage, such as decreased
oxidative stability [12], degraded cold-flow performance compared with petroleum diesel
[7], and increased emissions of NOx [8-11, 13]. Among these, the increase of NOx
emissions has especially drawn researcher’s attention and attempts to explain the origin
of the biodiesel NOx effect have been made [14-18], which may contribute to potential
approaches to biodiesel NOx reduction.
Traditionally, there are two ways for control of tailpipe NOx emissions: control of
in-cylinder combustion, and after treatment devices like NOx absorber catalysts and
selective catalytic reduction [19, 20]. The former can be affected by a variety of factors
including the physical and chemical properties of fuel, the engine control strategy (e.g.
fuel injection control), exhaust gas recirculation (EGR) ratio, and boost pressure. The
latter can be affected by the type of catalyst, exhaust gas composition and temperature of
the treatment device. The interest of this work focuses mainly on the engine in-cylinder
combustion control via fuel injection strategy, i.e., the impact of start of injection (SOI),
fuel injection pressure, fuel injection duration and multiply injections on the combustion
and, consequently, engine emissions.
Experiments conducted by previous researchers have shown that fuel injection
strategy can significantly affect engine emissions [21, 22]. For instance, it has been
shown that advancing the SOI can increase NOx emissions and decrease PM emissions
3
[21, 23]. Note this observation was also claimed to contribute to the biodiesel NOx effect
in traditional mechanical (pump line nozzle type) fuel injection systems [23] given that
biodiesel has a higher bulk modulus of compressibility. Nevertheless, the modern
common-rail injection system will not be affected by the fuel compressibility [13]. Fuel
injection pressure can also significantly affect the engine emissions. Increased fuel
injection pressure can increase NOx emissions and decrease PM emissions [19, 20]. This
observation is of special interest with biodiesel fuel. Since biodiesel has a lower heating
value than petroleum diesel [24], a higher brake specific fuel consumption (increased
injection quantity per unit time) for biodiesel is observed [6], and the default engine
calibration will achieve the increased injection quantity by increasing both the injection
pressure and the injection duration. Therefore, this work also intends to investigate which
is the more significant factor for the biodiesel NOx effect: biodiesel’s chemical properties,
increased injection pressure or increased injection duration. Furthermore, it has been
reported by many researchers that multi-injection strategy, especially multi-injection with
post injection [a fuel injection considerable after top dead center (ATDC)], can improve
engine emissions [13, 25-30]. Therefore, a thorough investigation of the post injection
strategy is also conducted in this work.
The other aspect of sustainable energy development – efficiency improvement in
energy applications (i.e., brake thermal efficiency of a diesel engine) is also an interesting
topic. Although biodiesel fueling introduces higher brake specific fuel consumption, the
brake thermal efficiency of a diesel engine operating with biodiesel has been observed to
be the same level as with diesel fuel [31]. Meanwhile, variation of fuel injection strategy,
which consequently affects combustion phasing, can be a significant factor to affect
4
engine brake thermal efficiency [22]. Theoretical investigation by Hoppie et al. [32]
suggests that neither the Otto nor the Diesel cycle is optimum for realistic engines, and an
optimum rate of heat release may be attainable via hypergolic combustion. However, the
artificial heat release profiles employed by them are not suitable for practical engineering.
Hence in this work, the author applies an analysis of optimized heat release pattern,
which can be controlled by fuel properties and engine injection strategy, intending to
provide suggestions for high efficiency engine design. While it is acknowledged that
other factors such as turbo, EGR system and aftertreatment devices can also affect the
total brake thermal efficiency of a diesel engine, they are not included in the discussion of
this work.
In summary, this work investigates the impact of biodiesel fueling and fuel
injection strategy, including start of injection, fuel injection pressure/duration and split
injection on diesel engine emissions and performance. The origins of the biodiesel NOx
effect and the change of NOx emissions by fuel injection strategy are investigated in this
work. In addition, a theoretical attempt to achieve high engine efficiency by optimizing
the apparent heat release pattern is also conducted in this work.
Objective: in this work the intention is to understand the relation between engine
emissions, biodiesel fueling and fuel injection strategy. The significance of factors
contributing to the biodiesel NOx effect is evaluated. In this work it is also intended to
understand the impact of post injection on lubricating oil dilution with biodiesel fueling,
and the impact of fuel injection pressure on diesel engine soot characteristics. The
optimized heat release pattern toward high engine thermal efficiency is also investigated.
5
Hypothesis:
Main hypotheses of this work include:
• Biodiesel NOx penalty can be eliminated by engine fuel injection control
• The impact on NOx emission from engine fuel injection strategy control and
biodiesel fueling can be explained by the equivalence ratio at autoignition
zone near the flame lift-off length
• Biodiesel fueling can introduce lubricating oil dilution with post injection
strategy
• Soot from high injection pressure has higher oxidative reactivity since the
decreased soot precursor concentration can prolong the soot inception time.
• Optimized heat release design is a sharp peak at top dead center
To achieve the objectives of this research plan, the study is divided into 5 tasks:
Task 1: Investigation of the impact on engine emissions of both biodiesel fueling
and injection strategy at different engine speeds and loads.
Task 2: Investigation of the origin of engine NOx emission variations from
biodiesel or injection strategy by evaluating the adapted spray model.
Task 3: Investigation of the impact of post injection on engine emissions and
lubricating oil dilution.
Task 4: Investigation of the impact of fuel injection pressure on diesel engine soot
oxidative reactivity with both diesel and biodiesel blends.
Task 5: Investigation of optimized heat release pattern to achieve high engine
thermal efficiency.
6
Chapter II. Literature Review
2.1 Biodiesel background
Throughout this thesis biodiesel implies a fuel composed of glyceride-free mono-
alkyl esters of long-chain fatty acids converted from triglycerides such as biologically-
based fats and oils [33-35]. In fact, attempts to use biologically derived hydrocarbons
(not only biodiesel) in internal combustion engines have been conducted for more than a
century. For instance, in 1900 when Dr. Diesel exhibited his great invention – the Diesel
engine – at the Paris World Exposition for the first time, the engine was running on 100%
peanut oil [33, 34]. From the 1920s to the 1940s numerous applications of vegetable oils
as diesel engine fuels were reported simply based on the theme of providing tropical
colonies of European countries with an independent source of fuel [36]. The more
appropriate form of biologically-derived fuel (ethyl esters of palm oil), which meets the
present definition of biodiesel, was initially included in a Belgian patent issued to
Chavanne as early as 1937 [37]. However, the petroleum derived diesel fuel immediately
became the primary fuel for diesel engines, and the concept of biologically-derived fuel
hadn’t drawn people’s attention until the 1970s due to the energy crises. Since then
vegetable oil-based fuels have been reconsidered among the alternative energy sources.
In addition, the feedstocks for biodiesel have also been expanded to a variety of vegetable
oils including soybean, rapeseed, sunflower, etc., which also creates a variation in both
physical and chemical properties of biodiesel [34].
The production of biodiesel mainly follows the transesterification process, which
can crack the large molecules (triglyceride) of vegetable oils/fats and transform them to
7
fatty acid methyl esters (FAME) with similar molecular weight as diesel fuel. Figure II-1
presents a traditional transesterification process for biodiesel production. This reaction
requires methanol as a reactant and alkali or acid as catalyst, and the by-product of the
reaction is glycerol. Demirbas [35] proposes that the same reaction can also be promoted
with supercritical methanol, with which no catalysts are needed, and consequently
purification of the final product can also be simplified.
Figure II-1: A diagram representing the transesterification process, replicated from

[6]. The general scheme of the process is to modify a triglyceride (a fat or oil which
is a glyceride with three fatty acids, represented as R1, R2 andR3) into a fatty acid
methyl ester.
8
With different structures (length, saturation) in the fatty acid, the physical and
chemical properties of FAME may vary. Table II-1 presents the chemical structure of
common fatty acids.
Table II-1: Chemical structure of common fatty acids, taken from [38]
Fatty acid Systematic name Structure* Formula

Lauric Dodecanoic 12:0 C12H24O2
Myristic Tetradecanoic 14:0 C14H28O2
Palmitic Hexadecanoic 16:0 C16H32O2
Stearic Octadecanoic 18:0 C18H36O2
Arachidic Eicosanoic 20:0 C20H40O2
Behenic Docosanoic 22:0 C22H44O2
cis-9-Hexadecenoic 16:1 C16H30O2
Palmitoleic Acid
Oleic cis-9-Octadecenoic 18:1 C18H34O2
cis-9,cis-12- 18:2 C18H32O2
Linoleic Octadecadienoic
cis-9,cis-12,cis-15- 18:3 C18H30O2
Linolenic Octadecatrienoic
- cis-11-Eicosenoic 20:1 C20H38O2
* xx:y indicates xx carbons in the fatty acid chain with y double bounds
Although these FAMEs are found in biodiesel, the chemical composition of
biodiesel is different if the biodiesel is derived from different feedstock, as shown in
Table II-2. It can be seen that generally C16:0, C18:0, C18:1 and C18:2 are the major
FAMEs species in biodiesel. In the United States, the primary feedstock for biodiesel is
soybean oil [12], which predominantly contains unsaturated FAME – C18:1, C18:2 and
C18:3, as well as, a small amount of saturated FAME – C16:0 and C18:0. The difference
in physical and chemical properties of FAME can also be significant, as shown in Table
II-3. Saturated FAME with a long chain always leads to higher melting point and
kinematic viscosity, which is expected. However, the degree of saturation can
significantly affect certain properties including melting point, oil stability index and
cetane number.
9
Table II-2: Chemical composition of biodiesel derived from different feedstock, taken from [39]
Fatty acid composition (%)

feedstock
14:0 16:0 18:0 20:0 22:0 16:1 18:1 18:2 18:3 20:1 Other
Canola oil 0 4 2 0 0 0 61 22 10 1 0
Palm oil 1 45 4 0 0 0 39 11 0 0 0
Soybean oil 0 11 4 0 0 0 23 54 8 0 0
Sunflower oil 0 6 5 0 1 0 29 58 1 0 0
Corn oil 0 11 2 0 0 0 28 58 1 0 0
Cottonseed oil 1 23 2 0 0 1 17 56 0 0 0
Chicken fat 1 25 6 0 0 8 41 18 1 0 0
Beef tallow 4 26 20 0 0 4 28 3 0 0 14
Table II-3: Physical and chemical properties of common fatty acid methyl esters (FAMEs), taken from [39]
Kinematic
Melting pointa Heat value Oil stability indexc Lubricitye (µm) at
FAME viscosityb at 40 ᵒC Cetane No.d
(ᵒC) (MJ/kg) (h), at 110 ᵒC 60 ᵒC
(mm2/s)
C14:0 19 39.45 3.3 >40 - 353
C16:0 31 39.45 4.38 >40 86 357
C16:1 −34 39.3 3.67 2.1 51 246
C18:0 39 40.07 5.85 >40 101 322
C18:1 −20 40.09 4.51 2.5 59 290
C18:2 −35 39.7 3.65 1 38 236
C18:3 −52 39.34 3.14 0.2 23 183
a: Ref. [7, 40]; b: Ref. [41]; c: Ref. [7]; d: Ref. [7]; e: Ref. [42]
10
Although cetane numbers of the C18:2 and C18:3 FAMEs are relatively low, as a
mixture, biodiesel is often observed to have a higher cetane number than petroleum diesel
[39], which improves its performance in diesel engines. High unsaturation can improve
the cold weather performance, but also significantly decrease the oxidation stability.
Chemical properties of a fuel (e.g., hydrocarbon structure, aromatic content,
cetane number) dominate the fuel’s effect on combustion and pollutant formation [43].
Physical properties (e.g., viscosity, density) generally affect injection timing, fuel
atomization and fuel evaporation, which eventually have indirect effects on combustion
and emission formation. In other words, diesel engine may have a different response
when biodiesel is used. Table II-4 presents a summary by Sun et al. [44] on the impact of
soy methyl ester (SME) based biodiesel and its blend with petroleum diesel on physical
and chemical properties and engine response. This summary also reveals the difficulty in
biodiesel research that multi-parameters can change simultaneously with biodiesel
fueling, so that the understanding of each parameter’s impact is not straightforward. For
instance, the lower heating value of biodiesel can lead to an increase injection pressure
and duration by the default engine calibration, which may also have impact on engine
emissions [21, 22]. This issue will also be addressed in this work.
11
Table II-4: The effects of property differences between soy methyl ester biodiesel or its blend and petroleum diesel on engine
parameters and fuel combustion. A positive/negative sign in the “difference” column means increase/decrease; a
positive/negative sign in the rest column means increase (or advance)/decrease (or retarding) of the property. Taken from [44]
Difference Injection Injection Fuel spray Fuel spray Fuel spray Ignition Heat Combustion
timing pressure penetration angle atomization delay release temperature
Physical Properties
Liquid densitya + + + - - +
Bulk modulus of + + + +
compressibilityb
Speed of soundb + + + +
a
Liquid viscosity + + - - +
a
Surface tension + + + -
Vapor pressurea - - - +
c
Volatility - -
Liquid specific - + -
a
heat
Vapor specific - + -
a
heat
Heat of + -
a
vaporization
Chemical Properties
Chain lengthd + -
Oxygen contentc + + - + +
e
Aromatics content - + + +
f
Sulfur content - +
Saturation (iodine - + +
d
value)
Cetane numberg + -
c
Heating value - -
a: Ref [45]; b:Ref. [23, 46]; c: Ref. [47]; d: Ref.[48]; e: Ref. [49]; f: Ref.[15]; g: Ref. [50]
12
at the same time, the “sustainability” concept of biodiesel has also been criticized
since the methanol needed for biodiesel production majorly comes from petroleum [35].
Nevertheless, approximately 95% of the carbon in biodiesel are derived from renewable
sources such as oil and fat [33]. Hence, it is still of interest to investigate the impact of
biodiesel on engine emissions and performance to approach the future of sustainable
energy.
2.2 Impact of biodiesel fueling on engine emissions
The review by Graboski and McCormick in 1998 [6] and a more recent review by
Lapuerta et al. in 2008 [15] show consistently that the majority of diesel engine tests with
biodiesel fueling show reductions of exhaust levels of some regulated emissions
including unburned hydrocarbons (UHC), carbon monoxide (CO), and particulate matter
(PM), but increases of emissions of nitric oxides (NOx). For example, Figure II-2 shows
the statistical analysis of trends of the impact of biodiesel blending on emissions from
heavy-duty engines, conducted by the Environmental Protection Agency (EPA) [51].
However, despite the “statistical analysis” by the EPA, the impact of biodiesel on engine
emissions as observed in Figure II-2 may be limited to certain engine conditions: in this
case it is a heavy-duty truck engine on highway. The understanding of biodiesel’s impact
on diesel engine emissions requires a thorough investigation of previous engine studies.
13
Figure II-2: NOx, PM, CO and HC % changes as biodiesel blending percentage
increase as determined by the EPA through statistical regression of publically
available data on highway heavy-duty truck engines [18, 51]. Taken from [44]
In addition, as discussed in the introduction, since NOx and PM emissions from
current technology diesel engines are close to the limits permitted by regulation and both
limits will be even more stringent in the near future, these two emissions will be critical
factors in the development of new diesel engines and aftertreatment systems. For the
other regulated emissions, CO and UHC, no further development in engine technology
seems to be necessary to meet future limits [15].
14
2.2.1 Nitric oxides (NOx)
In their comprehensive review, Lapuerta et al. [15] states that a majority of the
literature report an increase of NOx emissions with biodiesel use and a minority of other
observations. In this sense, they divide the impact of biodiesel fueling on NOx emissions
into 4 groups:
• Group I: Increase.
• Group II: Increase only in certain operating conditions
• Group III: No significant difference
• Group IV: Decrease with biodiesel fueling
Similar categorization is also adapted in another review by Sun et al. [44], except
that they claim that the majority of the literature they reviewed reported inconsistent
trends in the effect on NOx emissions from the use of biodiesel, in other words,
considering the distribution of engine configurations and operating conditions, it is
difficult to arrive at a deterministic conclusion on the impact of biodiesel fueling on NOx
emissions. Nevertheless, observations associated with engine configurations and
operating conditions are still meaningful for biodiesel research.
Group I: In Graboski and McCormick’s review [6] of ten older two-stroke engines
running with biodiesel derived from soy and rapeseed, all engines running with rapeseed-
oil biodiesel show a consistent increase in NOx emissions, and two third of the engines
running with soybean-oil biodiesel show a consistent increase in NOx emissions. The
remaining one third engines running with soybean-oil biodiesel showed both increases
and decreases in NOx emissions depending on the operating conditions, which should be
categorized into Group II. FEV Engine Technology [52] carried out an experimental
15
study in a 7.31L Navistar engine running the 13-mode US Heavy-Duty test cycle using
different soybean blends and observed a maximum of 8% increase in NOx emissions
with B100. Marshall et al. [53] tested a Cummins L10E engine under steady conditions
with different biodiesel blends, and a maximum of 16% increase in NOx emissions was
observed with B100. However, in their transient test condition B20 blend gave a higher
increase in NOx emissions than B30. Szybist et al. [14] tested a Yanmar L70 EE air-
cooled, four-stroke, single cylinder direction injection diesel engine operating at steady
3600 rpm and 75% load and reported around 10% increase in NOx emissions with pure
soybean-oil biodiesel. Other experiments have reported increased NOx emissions with
biodiesel fueling [18, 54, 55].
Group II: Some authors claimed that the effect of biodiesel on NOx emissions
depends on the type of engine and its operating conditions. Serdari et al. [56] tested on-
road emissions from three different vehicles using high sulfur diesel fuel (1800ppm) and
10% sunflower-oil biodiesel blends. They found both increases and decreases in NOx
emissions, and attributed such differences to the different engine technology and
maintenance conditions. Hamasaki et al. [57] tested a single-cylinder engine at 2000 rpm
and different loads with three waste-oil biodiesel fuels. They measured slight decreases in
NOx emissions at low loads but increases at high loads. Similar observations have also
been reported by Zhang and Boehman [13]. They conducted an experiment in a
DDC/VM Motori 2.5L, 4-cylinder, turbocharged, common rail, direct injection light-duty
diesel engine and found that NOx increases with biodiesel fuel were only observable at
high engine load. At low engine load, no significant difference in NOx emissions can be
observed between the baseline diesel, B20 and B40. Yehliu et al. [25] conducted
16
experiments in the same engine with a thorough investigation of different load and speed
conditions, and found that at high speed and low load, biodiesel fueling decreased NOx
emissions. Staat and Gateau [58] tested a 6-cylinder engine following the ECE R49 test
cycle and an urban transient cycle named AQA F21 established by the French Agency of
Air Quality. They observed a 9.5% increase in NOx emissions in the ECE R49 cycle, yet
a 6.5% reduction in the transient urban cycle. The above observations appear to be
consistent with the review by Tat [59], that NOx emissions with biodiesel fuels are
usually higher than those from diesel fuel when they are measured in an engine test bench
but not when they are measured from vehicles. Tat [59] attributed this observation to the
reason that engine loads are usually lower in vehicles than those imposed in experimental
test rigs. Similar observations have also been reported by McCormick [60, 61] who
measured NOx emission reductions of around 5% when using 20% soybean-oil biodiesel
blends.
Attempts have been made to explain these “conditional” biodiesel NOx effects. Li
and Gulder [62] claimed that the higher cetane number of biodiesel has a more sensible
effect on NOx emission at low load than at high load. However, this understanding may
no longer work when high-cetane ultra low sulfur diesel (ULSD) fuel is used instead of
the conventional low sulfur diesel fuel, i.e., blending of biodiesel with ULSD may not
significantly change the cetane number of fuel. Tat [59] proposed another explanation:
the injection pump tended to advance the injection timing at low load, but he observed
that this advance was higher with diesel than with biodiesel fuel in a certain load range,
leading to increased NOx emissions with diesel fuel at these load condition. This
explanation, however, is not applicable for diesel engines with a common-rail injection
17
system, whose start of injection is well controlled and consistent for both petroleum
diesel and biodiesel, as in [13, 25, 63]. Although the same authors attributed the NOx
reduction with biodiesel to the specific engine configuration, it might arise from the
decreased apparent heat release from premixed combustion with biodiesel [63], due to its
lower energy content and a shortened ignition delay.
Group III: Durbin et al. [64] tested four different engines with diesel, pure
biodiesel and a 20% biodiesel blend. The engines were chosen to represent a wide variety
of heavy-duty engines: turbocharged and naturally aspirated, direction and indirect
injection, and no significant differences in NOx emissions were found. Mandpe et al. [65]
tested a Euro III emission certified common-rail diesel engine on road with Jatropha
Curcas seed oil biodiesel and found no significant difference in NOx emissions between
biodiesel and petroleum diesel. Nabi et al. [66] conducted an experiment with a single-
cylinder 9.8 kW engine at steady-state with neem-oil biodiesel at different EGR ratios.
Although they observed decreased NOx emissions with increased EGR ratio, they found
no significant differences in NOx emissions between diesel and biodiesel. As suggested
by Lapuerta et al.[15], the observation by Nabi et al. [66] may be actually due to the low
unsaturation level of neem oil, since Graboski et al. [67] have shown that NOx emissions
increased with unsaturation.
Group IV: A minor number of papers have reported decreases in NOx emissions
when using biodiesel fuels. Armas et al. [63] conducted an experiment in a DDC/VM
Motori 2.5L, 4-cylinder, turbocharged, common rail, direct injection light-duty diesel
engine at high speed and low load, with different start of injection. For each start of
injection they also fine-tuned the ECU to match the combustion phasing between
18
biodiesel blends and baseline diesel. They found slight NOx reduction with biodiesel
consistently for all start of injection timings. Lapuerta et al. [68] tested pure and blended
biodiesel from sunflower and cardoon oils in an indirection injection 1.9L engine
operating at five selected steady modes, and they observed a slight decrease in NOx
emissions with biodiesel.
Overall, although no unanimity has been found, the majority of observations in
the literature show a slight increase in NOx emissions with biodiesel fueling. Under
certain conditions, no significant difference or decreased NOx emissions were observed
with biodiesel, which was likely due to specific engine configuration or fuel composition,
which suggests that such results are not generally representative.
2.2.2 Particulate Matter (PM)
It has been consistently reported in the literature that a noticeable decrease in
diesel PM emissions can be observed with biodiesel fueling [6, 8, 9, 11, 58, 69-72]. The
Environmental Protection Agency collected a number of engine emissions data with
biodiesel [51] and proposed a statistical correlation between PM emissions level and
biodiesel blending:
PM
= e −0.006384⋅% B (1)
PM D
PM
where indicates the PM emission ratio between biodiesel and diesel, and %B is the
PM D
biodiesel blending level in percentage. The shape of Equation (1) can be also seen in
Figure II-2.
19
This trend proposed by EPA suggests that the maximum PM emission reduction is
around 50% with pure biodiesel. This observation is consistent with several other
investigations [54, 69, 73-75]. Other authors found even higher PM emission reductions
with biodiesel. Van Gerpen and Canakci [76] observed as high as 65% reductions of PM
emissions with soybean oil and waste oil based biodiesel fuels. Schumacher et al. [77]
tested PM emissions from a DDC series 60 diesel engine and reported more than 60%
PM emission reduction with biodiesel.
The reductions in PM emissions have been shown in general, as being more
effective with lower biodiesel concentrations in the blends [15]. In other words, the
reduction of PM emissions for each percentage of biodiesel blending, i.e., relative
reduction, has been observed to be the highest for partial blends [78-81].
The effects of biodiesel on PM emissions together with other parameters, such as
engine configurations and load conditions have also been investigated. In the review by
Krahl [82] on diesel engine tests with rapseed-oil biodiesel, the authors found that the PM
emission reductions were lower ( or there were no significant reductions) in heavy-duty
engines than in light-duty engines. Lapuerta et al. [79] investigated the PM emissions of a
typical European passenger car diesel engine operating at low-middle load conditions
and found that the PM emission reduction with biodiesel is larger at low load. Contrary
observations have also been reported [83]. Leung et al. [84] tested a single-cylinder
engine with a diesel fuel and a pure rapeseed oil based biodiesel at different load
conditions. They found larger decreases for biodiesel at high load. In addition, the
property of diesel fuel can also affect the NOx reduction by biodiesel. Boehman et al. [75]
reported a PM emission of 20% blending soybean oil based biodiesel with both low
20
sulfur diesel (325 ppm sulfur) and ultra low sulfur diesel (15 ppm sulfur) and found that
the reduction of PM emissions by biodiesel is more significant for low sulfur diesel.
Similar observations have also been reported [51].
A small number of studies did not find significant reductions (or even found
slightly increases) in PM emissions. Armas et al. [63] found increased PM emissions in
their experiment with a DDC/VM Motori 2.5L, 4-cylinder, turbocharged, common rail,
direct injection light-duty diesel engine at high speed and low load. Boehman et al. [75]
found a slight increase of PM emission for soybean oil based biodiesel blending with
ultra low sulfur diesel. Similar observations that PM emissions were increased with
biodiesel have also been reported in other places [74, 85]. This phenomenon has been
generally attributed to the higher soluble organic fraction (SOF) of the particulate [63,
75], which is widely accepted to be increased when using biodiesel [6, 67, 69, 74, 77] due
to its lower volatility of the unburned hydrocarbons (Table II-4). Meanwhile, the
insoluble fraction (ISF) was decreased with biodiesel fueling [75].
In summary, it is widely accepted that biodiesel fueling can decrease the PM
emissions of diesel engine. Therefore, it has been suggested that this impact of biodiesel
on PM emissions may be combined with engine injection strategy to optimize the NOx-
PM emission trade-off [15].
21
2.3 Origin of biodiesel emission effect
2.3.1 Biodiesel NOx effect
NOx formation chemistry:
To identify the factors that can affect NOx emissions, it is necessary to understand
the mechanism of NOx formation. Although a large number of reactants and reaction
pathways have been found to participate during NOx formation [86, 87], the key
reactions at the pressure, temperature and time scales in a diesel engine consist of three
major reactions [86], referred as to the extended Zeldovich mechanism:
N 2 + O 
k1
→ NO + N
O2 + N 
k2
→ NO + O (2)
OH + N 
k3
→ NO + H
Typically the time scales in a diesel engine do not allow these reactions to reach
equilibrium. Therefore, longer residence time of in-cylinder gases at high temperature
(above 1800K since the thermal NO formation below 1800K is generally negligible [87])
can lead to higher NOx emissions. A simple second order kinetic expression can be used
to describe the NO formation when its concentration is below equilibrium:
d [ NO ]
= k1[O ][ N 2 ] + k 2 [O2 ][ N ] + k3[OH ][ N ] (3)
dt
where [NO], [O], [N2] and [OH] are the molar concentrations of NO, O, N2 and OH,
respectively. The rate constant k can be generally expressed as [88]:
Eai
ki = BiT α i exp( − ), i = 1, 2,3 (4)
RuT
22
where Bi, αi, Ru and Eai are constants for each ki. Equations 2 and 3 suggest that the key
parameters that control the NOx formation include: 1) concentration of O-atom, N2, N-
atom and OH, 2) in-cylinder temperature, or 3) the residence time of the in-cylinder
mixture at high temperatures. Hence, any variation in NOx emissions must be related to
one or more of the above factors.
Biodiesel NOx effect:
To date, a significant number of studies have been conducted in the attempt to
explain the NOx variation and to identify the key parameters that affect it when
alternative fuels (e.g., biodiesel) are used, both from experiments [14, 15] and numerical
simulation [16, 17, 89]. Literature reviews by Mueller et al. [18], Lapuerta et al. [15] and
Sun et al. [44] have discussed the various proposed hypotheses that may account for
engine NOx variation with biodiesel fueling, which can be summarized into seven
categories:
(1) Injection timing. Alternative fuels have different compressibility than
petroleum diesel. For example, biodiesel has higher bulk modulus of
compressibility while FT diesel has lower bulk modulus of compressibility [23].
A higher bulk modulus can lead to a faster traveling of the pressure wave in the
fuel line of pump-line-nozzle fuel injection systems, causing an advance in the
start of injection (SOI) that can consequently lead to a longer residence time
and/or higher in-cylinder temperature. The advanced SOI of biodiesel and
retarded SOI of FT diesel have been observed to affect the NOx emissions [23, 90,
91]. Although this factor has been considered as the most solid argument for the
23
biodiesel NOx effect [15], the injection timing of a modern common-rail injection
system is not affected by the bulk modulus of compressibility of the fuel [13].
(2) Injection pressure and fuel spray. The lower heating value of biodiesel
results in an increase in the brake specific fuel consumption, which will be treated
as an increase of the fuel injection pressure/duration according to default engine
calibration [55]. This change in fuel injection parameters can yield an increase in
NOx emissions [21, 22]. Furthermore, the change of fuel properties, e.g. viscosity,
density and surface tension when biodiesel is used can increase the average
droplet diameter [92], leading to a lengthened ignition delay as well as a possibly
longer diffusion combustion period [6].
(3) Combustion phasing. That biodiesel can increase NOx emissions also can
be due to its impact on combustion phasing. The higher cetane number of
biodiesel can advance the start of combustion (SOC), which may lead to longer
residence time at high temperature for the combustion products [18]. However, as
discussed previously for FT diesel, some authors also claim that the NOx
reduction is due to its higher CN [25, 63, 91, 93] since very high CN indicates
short ignition delay, leading to less premixed burning [94, 95]. This inconsistency,
that both biodiesel and FT diesel have increased CN, leads to different direction in
change of NOx emissions, suggests that the impact of high cetane number on
NOx emissions is a combination of two effects.
(4) Premixed-burn fraction. The fact that biodiesel contains oxygen increases
the premixing quality (since fewer oxidizer molecules are required to reach
stoichiometric proportion) during the ignition delay, resulting in better
24
combustion and higher temperature. The oxygen content in biodiesel also
increases the local oxygen concentration, which can promote NOx formation as
discussed previously. Song et al. [96, 97] tested oxygenated diesel fuel (via
oxygenated fuel additives) and found increases in NOx emissions. However, this
explanation is questioned by other authors [68, 69, 89]. In addition, this
explanation is not suitable for FT diesel since it does not contain oxygen in the
fuel.
(5) Kinetics. Different chemical kinetic pathways for NOx formation exist for
different fuels. For example, it is argued that an increased level of CH exists at the
autoignition zone (AZ) during biodiesel combustion [98], which leads to the
production of N-atoms in the jet core as well as the promotion of NOx formation.
(6) Adiabatic flame temperature. Some authors claimed that biodiesel has a
slightly higher adiabatic flame temperature than conventional diesel, so thermal
NOx production is higher [16, 69, 99]. Although the decreased NOx emissions for
FT diesel appears to be consistent to this theory: FT fuel has a lower adiabatic
flame temperature [100], due to its lower C/H ratio [101], than the petroleum
diesel fuel. No unanimity can be found for the biodiesel since others state that the
adiabatic flame temperature is higher for diesel fuel [44, 66, 102].
(7) Radiation heat transfer. Soot radiation is the primary means of heat loss
from an in-cylinder flame. It has been suggested that since biodiesel can decrease
the PM emissions, the actual flame temperature is higher, leading to an increase
of NOx formation [103]. However, it has been suggested that soot emissions may
25
not correlate with flame luminosity and the soot radiation appears to be smaller
than other effects [18].
Mueller et al. [18] also pointed out that the biodiesel NOx increase is related to a
number of coupled mechanisms, which also have influences on each other, rather than a
single phenomenon. Note that many of those mechanisms are affected by both fuel
properties and engine operating conditions. For example, the premixed-burn quality can
be affected by the droplet size, which is controlled not only by the injection strategy and
fluid surface tension [1], but also by the difference in the rate of mass transfer during the
injection of the multi-component fuel [104]. Among those mechanisms, Mueller et al. [18]
suggest that the most important ones responsible for the biodiesel NOx increase are
longer residence time and higher in-cylinder temperature from either advances in
combustion phasing or lower radiative heat loss. They proposed, based on the models
established by Siebers and Naber [105, 106], that the origin of the NOx increase with
biodiesel is based on reacting mixtures that are closer to stoichiometric (less rich) for
biodiesel-containing fuels: a) during ignition (i.e., during the premixed volumetric
autoignition event from the start of combustion to the end of premixed burning), and b) in
the standing premixed autoignition zone (AZ) near the flame lift-off length (LOL) at
higher loads. According to Mueller et al. [18], this is so far the most generalized
explanation which is consistent with the observations from biodiesel, high cetane number
diesel and engine load conditions. In additional to their heavy-duty single-cylinder engine
test, in this work this theory will be tested with a light-duty diesel engine for the impact
of both biodiesel and engine fuel injection strategy.
26
2.3.2 Biodiesel particulate matter effect
As discussed previously, it is widely accepted that biodiesel can decrease the PM
emissions. It is also believed that the factors contributing to the PM emission reductions
include:
• Oxygen content in biodiesel: the oxygen content of the biodiesel molecule,
which can promote complete combustion even in regions of fuel-rich diffusion
flames, has been believed to contribute to the PM emission reduction with
biodiesel [107]. Song et al. [96, 97] tested oxygenated diesel fuel (via oxygenated
fuel additives) in a VW turbocharged direct injection diesel engine, and found
decreased PM emissions with increased oxygen content of fuel for both engines.
Similar observation has also been reported by Frijters and Baert [108]. Sison et al.
[109] tested different oxygenates, including biodiesel, with diesel fuel in a single-
cylinder optically accessible diesel engine and observed decreased soot formation.
The kinetic model developed by Flynn et al. [110] based on the conceptual model
of Dec [111] suggested that a sharp decrease in the formation of soot precursors
would be observed when the oxygen content in the fuel increased: as the oxygen
content of the fuel increased, larger fractions of the fuel carbon were converted to
CO in the rich premixed region, rather than to soot precursors.
• Absence of aromatics in biodiesel: The aromatic content in petroleum
diesel fuel has been considered to serve as soot precursors [69, 71, 112, 113]. This
can be investigated in two ways: 1) decrease the aromatic content in petroleum
diesel, (Schmidt and Van Gerpen [113] artificially blended diesel fuel with
octadecane (C18H38) and observed a significant reduction in PM emissions.); 2)
27
increase the aromatic content in biodiesel. Mueller et al. [18] doped the pure
biodiesel with phenanthrene, a 3-ring aromatic species, and compared the
emissions at both decreased oxygen environment and normal air environment
with a heavy-duty single-cylinder optically accessible diesel engine. They found
increased PM emissions for all load conditions for the doped biodiesel in a
decreased oxygen environment. With a normal air environment, the doped
biodiesel has an increased PM emission only at low-moderate load condition.
Other authors have treated both the oxygen content and absence of aromatics in
the biodiesel to the PM emission reductions [114, 115].
• Higher oxidative reactivity of biodiesel soot: Boehman et al. [75] collected
the particulate matter from the exhaust of a six-cylinder Cummins ISB 5.9L direct
injection turbocharged diesel engine, with both diesel fuel and a 20% blend
soybean oil based biodiesel, and then they employed thermogravimetry to analyze
the oxidative reactivity of the soot particles (de-volatized particulate matter).
They found that the soot particles for B20 are more reactive than for diesel fuel,
which suggests that the particulate matter will be more rapidly consumed by the
oxidation in the exhaust, leading to decreased PM emissions. Song et al. [116]
expanded this observation to pure biodiesel and used high resolution transmission
electronic microscopy (TEM) technique to confirm the faster oxidation of the soot
for pure biodiesel. Yehliu et al. [117] applied a counting algorithm to identify the
initial nanostructure of soot particles from high resolution TEM images, and they
found more short and curvy graphene layers in soot from biodiesel combustion,
28
i.e., it has a higher oxidative reactivity due to its higher degree of disorder [118-
120].
There have been other explanations proposed for the biodiesel PM effect, as well.
For instance, biodiesel can introduce an advance of combustion either due to advanced
injection timing [90] or higher cetane number [102] can prolong the residence time of
soot particles in a high-temperature atmosphere, which in the presence of oxygen
promotes further oxidation [72, 112]. Others also claimed that the usually lower final
boiling point of biodiesel provided lower probability of soot or tar being formed from
heavy hydrocarbon fractions unable to vaporize [69]. High sulfur content of diesel fuel
has also been suggested as a reason for its high PM emissions [121]. Nevertheless, these
explanations have been believed to be insignificant [108, 115, 122], as well as not
applicable since new fuel standards came into effect.
2.4 Fuel Injection strategy with biodiesel fueling
Fuel injection strategy is an important aspect in the optimization of diesel engine
missions. For instance, in earlier discussion the significant impact of injection timing and
injection pressure on engine NOx emissions has been shown. Some studies have
attempted to optimize diesel engine combustion with biodiesel fuel by adjusting the
injection strategy to improve exhaust emissions [123, 124]. Generally speaking, for
conventional diesel combustion, the control of fuel injection strategy includes three
aspects 1) injection timing (or start of injection); 2) injection pressure or injection
duration; 3) number of fuel injection. Among those, although the advance of injection
29
timing with biodiesel fueling was a concern for earlier engine research, it is now believed
as an insignificant factor for modern engines with common-rail fuel injection systems
[13]. The application of common-rail fuel injection systems also provides availability of
multiple fuel injection design, which has become a fuel injection strategy. A large
number of researchers have found that multiple fuel injections can reduce engine NOx
emissions [125-129]. However, some authors observed increased hydrocarbon emissions
and slight decreased engine thermal efficiency with multiple fuel injections [25]. In
addition, the complex combustion behavior of multiple fuel injections is not suitable for
fundamental engine combustion study, and therefore it is not a main focus in the fuel
injection strategy investigated here. Nevertheless, multiple fuel injections with post
injection, which is a fuel injection event considerably after top dead center (TDC) will be
investigated in this work and discussed in a later section.
Meanwhile, fuel injection pressure is affected by fluid properties such as density,
compressibility, and speed of sound [76, 130]. Biodiesel fueling in a diesel engine can
also affect the fuel injection pressure or/and injection duration, regardless of the type of
fuel injection system. As shown by Heywood [1], the fuel mass flow rate thought the
injection nozzle can be related to:
∆θ
m f = C D An 2 ρ f ∆P (5)
360 N
where m f is the mass of fuel delivered per cycle, C D is the injector tip nozzle discharge
coefficient, An is the nozzle flow area, ρ f is the fuel density, ∆P is the pressure
difference across the nozzle, ∆θ is the injection duration, and N is the engine speed.
Hence, for an engine running at steady state (i.e., N is constant), ρ f , ∆P and ∆θ control
30
the mass fuel flow. Steady state operating also requires a constant engine consumption,
and since biodiesel has around 13% heat value lower than petroleum diesel fuel, with
albeit a slight increase (around 3%) in fuel density [34], an increased fuel injection
volume (i.e., the value of ∆P∆θ ) is needed with biodiesel fueling [107].
The fuel injection pressure and ambient gas pressure, together with fuel properties
including density, surface tension, bulk modulus and boiling point can affect fuel spray
atomization and penetration, which consequently affect air mixing and fuel evaporation
rate [131-134]. Allen and Watts [48] stated that the Sauter mean diameter [135] of methyl
ester biodiesel, which can be used to describe the spray droplet size, varies from 5% to 40%
higher for biodiesel than petroleum diesel fuel, which was suggested to reduce the
premixed phase of combustion, causing an increase in the diffusion phase of combustion
[6]. Similar observations have also been reported by Suh et al. [136] They concluded that
the higher viscosity and surface tension of biodiesel were the reason for the its higher
droplet size, lower axial mean velocity and lower droplet number during single injection.
Meanwhile, no impact from viscosity or surface tension on spray penetration length was
found. Naber and Siebers [105, 106] experimentally measured the spray penetration
length of diesel fuel in various environments and found that it is dominantly controlled by
fuel injection pressure, fuel density and ambient gas density. These changes of spray
atomization and formation can affect fuel combustion and consequently gaseous
emissions. Although a significant number of studies attempted to investigate the
combustion character with computational fluid dynamics simulations of the fuel spray
process [137-141], the investigations have been computationally intensive and time
consuming. From this consideration, Siebers et al. [105, 106, 142-145] have
31
comprehensively developed a simple and fast mathematical model that describes the
equivalence ratio distribution in a fuel spray and the flame lift-off length based on
mathematical modeling and experimental correlations. Even without full implementation
of the complex physical and chemical process of fuel spray injection and ignition, this
model has been shown to be able to effectively describe fuel combustion and gas
emissions [18]. Hence, it will be adapted in this work in the attempt to explain the
variation of the engine NOx emissions, and a more detailed description of the model will
be presented in a later section.
In addition, given the aforementioned fuel injection behavior, the default
calibration of diesel engine fuel injection system tends to increase both fuel injection
pressure and duration when biodiesel is used. As discussed earlier, this “system response”
would also contribute to biodiesel’s emission effect. It has been shown that the increase
of fuel injection pressure leads to increased NOx emissions and decrease PM emissions
[21, 22] since the formation and duration of the fuel spray can significantly affect
combustion phasing [146]. A question still remains that for biodiesel usage in a common-
rail diesel engine, which aspect, including fuel chemistry, the increased injection pressure
and duration, is the most significant factor that account to biodiesel’s emission effect.
This question will also be addressed in this work.
32
2.5 Post injection strategy
As discussed earlier, multiple fuel injection strategy has been investigated by
many researchers as an approach to control engine exhaust gas emissions [125-129]. The
more conventional way to conduct multiple fuel injections is usually through a split fuel
injection strategy: a pilot fuel injection with a small quantity of the fuel starting at from
35º to 15º before top dead center (TDC) and a main injection with the majority quantity
of the fuel starting at around TDC [25, 136]. This split fuel injection will appear as two
stage combustion events in the apparent heat release profile. Since the fuel mass injected
at each event is lower than that from a single injection event, the local combustion
temperature and residence time at high temperature can be reduced, leading to decreased
engine NOx emissions.
At the same time, it is acknowledged that this kind of split injection: pilot + main
injection does not significantly alter the fuel combustion behavior of diesel engine,
resulting in an insignificant improvement in NOx-PM emission trade-off [25]. To
improve the emissions trade-off, researchers have attempted to take advantage of
multiple-injection capacity in new combustion methods, which generally represent
advanced compression-ignition combustion modes. Popular methods proposed for
advanced combustion modes include homogeneous-charge compression-ignition (HCCI)
and low-temperature combustion (LTC) [147], both of which rely extensively on engine
fuel injection control. For instance, HCCI combustion can be achieved by early and
multiple-stage fuel injection to create a homogenous (or close to homogeneous) fuel/air
mixture before ignition [148-152]. However, HCCI combustion has been observed to
have difficulty in extending operating ranges, short ignition delay, non-homogeneous
33
mixture formation and other problems [147]. From this consideration, some introduced a
new method used to depress NOx and soot formation by controlling the combustion
temperature [153]. It employed a very high EGR ratio to dilute the oxygen content in
engine intake air to reduce the combustion temperature. Later this approach was named
as low-temperature combustion (LTC) [154, 155]. Although LTC is formed on the
similar basis to HCCI, that the fuel/air mixture is more uniform than the conventional
diesel combustion, it still possesses moderate inhomogeneity of the whole mixture
compared to HCCI. This leads to a combustion process that is easier to control with a
flexible fuel injection strategy. In fact, LTC can be achieved by either very early injection
before TDC or late post injection after TDC [22, 147, 156]. However, the late post
injection often showed much better engine thermal efficiency than the early injection
method [157], and the higher exhaust temperature from post injection can also be utilized
to re-generate aftertreatment devices [19]. Therefore, a large number of authors have
investigated the application of post injection in diesel engines [26, 28-30, 158, 159]. Yun
and Reitz [28] tested the post injection strategy combined with a main injection in a
single-cylinder version of a 2.4L five-cylinder light-duty common-rail DI diesel engine
and found reductions of both NOx and PM emissions. Fang and Lee [159] tested a single
pulse post injection in an optically accessible single-cylinder high speed direction-
injection (HSDI) diesel engine and also found improvement in both NOx and PM
emissions.
At the same time, it has been observed that post injection can introduce increased
CO and unburned hydrocarbon (UHC) emissions [13, 25-30]. Therefore, a combination
of biodiesel fueling and post-injection strategy appears to be attractive. The increased
34
NOx emissions with biodiesel (as discussed earlier) can be reduced by using a post
injection strategy, while biodiesel’s lower CO and UHC emissions can compensate the
disadvantages of post injection strategy. Fang and Lee [160] tested biodiesel in an
optical-accessible single-cylinder high-speed direction-injection (HSDI) diesel engine
with an early pilot and post main injection strategy, and they found with biodiesel fueling
both NOx emissions and soot luminosity decreased. However, a few potential issues may
also arise. For example, unexpected lubricant dilution may be introduced [161] since the
emissions of UHC, which may be dissolved into the oil film on the cylinder wall, are
significantly increased. Fuel impingement on the cylinder wall has also been observed for
post injection, and it becomes stronger with biodiesel fueling [162]. In addition, the
complexity of the lubricant dilution challenge may increase as a multitude of
hydrocarbon species from exhausts can be observed in an engine using post injection
with diesel fuel [19]. This issue will become more severe when biodiesel fueling is used,
since biodiesel tends to concentrate in the lubricant [20, 163, 164] due to its higher
boiling point [20] than diesel fuel. High levels of oil dilution by biodiesel has a negative
impact on engine performance since it can increase mechanical friction [161], engine
wear [165], and metal corrosion [166], which may pose additional challenges for
lubricant development. However, a systematic understanding of these issues is not
available due to the lack of published studies combining both biodiesel and post injection
strategy.
Since the impact of post injection on engine performance and fuel oil dilution,
especially with biodiesel fueling, is still not well understood, the work presented here will
investigate the impact of post injection with biodiesel fueling on exhaust emissions, fuel
35
combustion and lubricating oil dilution. In addition, the hydrocarbon species in both
exhaust and used lubricant were also characterized.
2.6 Biodiesel and fuel injection pressure on PM morphology and soot

oxidative reactivity
2.6.1 PM morphology
Diesel soot agglomerates are composed of primary particles with geometries very
similar to spheres. Although both the mean diameter of the primary particles and their
standard deviation have small variations within the engine operating conditions [167],
diesel agglomerates are usually treated as uniform in size (monodisperse) [168].
Meanwhile, the soot agglomerates often show irregular shape which may be not easy to
describe with traditional terminology (Figure II – 3).
36
Figure II-3: Diesel soot agglomerate composed of spherical primary particles, taken
from Ref. [168]
The description of this kind of irregular formation presents difficulties for
making quantitative observations. Recently researchers have applied fractal geometry
[169] in the investigation of diesel engine soot agglomerates and to obtain quantitative
observations [168, 170, 171]. Generally speaking, the morphology analysis of diesel soot
agglomerates includes three aspects: 1) primary particle size 2) fractal dimension of the
agglomerates 3) particle size distribution of exhaust gases. The first two analyses can be
conducted based on transmission electronic microscopy (TEM) images of diesel
particulate matter obtained from thermophoretic sampling [168], and the particle size
distribution information can be obtained with scanning mobility particle sizer (SMPS)
[63]. The detailed description of test procedure will be provided in a later section.
37
These morphology characterizations essentially reflect diesel particle
agglomerates’ ability to adsorb unburned hydrocarbons. For instance, for agglomerates
with smaller primary particle size or lower fractal dimension, it will possess higher
surface area per unit mass, which may adsorb more hydrocarbons and possess higher
active surface area for soot oxidation [172]. It has been observed that both fuel properties
and engine operating condition have significant impacts on diesel particulate matter
morphology [73, 170, 173-176]. Lapuerta et al. [177] investigated the impact of engine
operating conditions on the size of primary particles of diesel soot agglomerates, and they
found the primary particle size decreases with increased air/fuel ratio, and increased
engine speed. No significant impact from EGR ratio was observed though. Smekens et al.
[178] investigated the impact of biodiesel fueling on primary particle size of diesel
engine soot agglomerates and found decreased primary particle size with biodiesel.
Mathis et al. [176] investigated the effect of fuel injection pressure on primary particle
size and found a reduction in primary soot particle diameter when fuel injection pressure
increases or start of injection advances. Lapuerta et al. [170] found that the decrease of
air/fuel ratio also decreases the fractal dimension of soot agglomerates, however, engine
speed did not have a significant impact on the fractal dimension. Zhu et al. [179]
investigated the fractal dimension of soot agglomerates from both a light-duty and a
heavy-duty diesel engines. They found that soot agglomerates from light-duty diesel
engine have lower fractal dimension. Neer and Koylu [174] tested particulate
morphology with a medium-duty John Deere 5059T diesel engine and found neither
engine speed nor load can significantly affect agglomerate fractal dimension. Similar
observations have also been reported by Soewono and Rogak [175]. They tested a VW
38
1.9L 4-cylinder turbocharged direct injection diesel engine and found no significant
impact on fractal dimension from engine speed or load. They also conducted experiments
with B20 blend biodiesel and found no impact on fractal dimension either. However,
Lapuerta et al. [170] found an increase in fractal dimension with increased engine load in
a 2.2L 4-cylinder direction-injection diesel engine (Nissan YD22). Other researchers
investigated the impact on particle size distribution from fuel properties and engine
operating conditions. Tsolakis [173] measured the particle size distribution with a heavy-
duty single-cylinder direction injection diesel engine and found decreased mean particle
diameter with high engine speed. Mathis et al. [176] investigated the impact of fuel
injection pressure on particle size distribution. They found the mean particle diameter
significantly decreases with high fuel injection pressure. Maricq [180] investigated the
impact of biodiesel fueling on particle size distribution and found a slight increase in the
number of smaller particles. Similar observations have also been reported in other studies.
[73, 173]
In summary, the number of existing studies on the impact of biodiesel fueling and
engine fuel injection strategy on the morphology of soot agglomerates is still limited, and
no unanimous conclusion can be summarized from previous work. Hence, the impact of
biodiesel fueling and engine fuel injection strategy on particulate morphology will be
investigated in this work.
39
2.6.2 Soot oxidative reactivity
As discussed earlier, it is widely known that biodiesel fueling can increase the
oxidative reactivity of diesel engine soot [75, 116, 117, 181]. This is due to increase of
degree of disorder of the soot [118-120], given a carbon atom at an edge site is much
more reactive than one in the basal plane of a graphene layer. As explained by Vander
Wal and Mueller [182], the increase of fuel oxygen content can increase the degree of
amorphous nanostructure with narrower fringe length distributions and broader and larger
tortuosity. Engine operating conditions can also significantly affect soot oxidative
reactivity. Braun et al. [183] tested the soot nanostructure with a two-stroke diesel engine.
They found that soot from an idling condition is less ordered in the crystalline structure.
Similar observations have also been reported by [175]. Al-Qurashi and Boehman [184]
investigated the impact of EGR ratio and observed that EGR soot is less ordered than
non-EGR soot. Yehliu [185] compared the oxidative reactivity of soot from single
injection and split injection, and found higher oxidative reactivity for soot from a single
injection strategy. Yehliu also found increased oxidative reactivity with retarded injection
timing. Although they suggested that fuel injection pressure had an impact on soot
oxidative reactivity, detailed investigation hasn’t been conducted. Zheng et al. [186]
investigated the impact of fuel injection duration on soot nanostructure, and they
observed no significant difference in the degree of disorder between different injection
durations.
It is believed that soot nanostructure depends upon its formation conditions, such
as temperature and residence time [187]. Vander Wal and Tomasek [188] observed that
soot formed at 1250ºC in a furnace appeared to be amorphous in its nanostructure, while
40
soot formed at 1650ºC shows a graphite-like nanostructure. In fact, Al-Qurashi and
Boehman [184], in the observation of higher disorder of EGR soot compared with non-
EGR soot, and Barun et al. [183], in the observation of higher disorder of idling soot,
both claimed that cylinder temperature is the key factor that determines soot crystal
structure. However, Seong and Boehman [189] thoroughly investigated the impact of
engine cylinder temperature and adiabatic flame temperature on soot oxidative reactivity
by intake oxygen enrichment. Although they found increased oxidative reactivity for soot
from lower adiabatic flame temperature combustion in some cases, this statement is not
generally valid. For some cases soot from similar adiabatic flame temperature
combustion possesses significantly difference in oxidative reactivity. Meanwhile, in
another investigation of the degree of disorder of soot collected from flame burner by the
same authors [190], they concluded that it was soot inception time that determined the
soot degree of disorder: the longer the soot inception time, the more amorphous the soot
particles. This statement is consistent with various experimental observations since soot
inception time is shorted with increasing flame temperature due to the enhanced soot
formation, and oxygenated fuel combustion has prolonged soot inception time due to
decreased soot precursor concentration [191-193]. In the present work we also believed
this is the key parameter that determines the soot oxidative reactivity, although Seong
[190] did not provide a theoretical justification. On the other hand, Hurt et al. [194]
investigated the nanostructure of primary soot particles with a thermodynamic approach
and found that during soot formation, the nanostructure has a “self-organization” feature,
i.e., it will transform itself into a more ordered state to decrease the total energy, given
that the soot precursors always possess more disordered structure than mature soot
41
particles [195, 196]. Hence, the reason that longer soot inception time corresponds to
higher degree of disorder of soot particles may be due to a decreased time for soot self-
organization, which can leave more amorphous (original state of soot precursors) soot
particles.
In summary, studies on the impact of fuel injection pressure on soot oxidative
reactivity is limited. Since higher fuel injection pressure can significantly decrease soot
emissions, it is expected that the soot inception time will increase. According to the
aforementioned theory, soot from higher fuel injection pressure should possess higher
oxidative reactivity, which is one of the hypotheses of this work.
2.7 Impact of biodiesel fueling, injection strategy and heat release

pattern on Engine efficiency
Engine thermal efficiency is the ratio between the power output and the rate of
energy consumption of fuel, which is a product of fuel consumption speed and lower
heating value. Since it is customary to use the brake power for the determination of
thermal efficiency in experimental engine studies, the efficiency obtained is really a
brake specific thermal efficiency. This parameter is more appropriate than brake specific
fuel consumption in comparing engine performance with different fuels.
With biodiesel fueling, the majority of existing literature shows no significant
difference in engine brake specific thermal efficiency [10, 76, 173, 197-201]. Graboski et
al. [67] tested a large number of methyl esters from different feedstocks and even pure
esters in a 11.1L, 254 kW engine following the transient cycle for heavy-duty engines (40
CFR Part 86 subpart N). From the results obtained, they showed that neither the oil origin,
42
nor the length of the carbon chain, nor the number of double bonds provides significant
differences in thermal efficiency. Nabi [202] investigated the impact of fuel oxygen
content on engine thermal efficiency, and found that efficiency decreases with increased
oxygen content only when oxygen content of the fuel is higher than 20%, otherwise the
impact was insignificant. A minor number of authors have found some decrease [63] or
some improvement [203] in thermal efficiency with biodiesel fueling. Armas et al. [63]
tested pure biodiesel derived from soybean oil in a 2.5L 4-cylinder turbocharged
common-rail direction injection DDC/VM Motori light-duty diesel engine at a steady
high speed and low load mode, with three start of injection timings. They found
decreased brake thermal efficiency for all injection timings. Some authors also observed
a conditional impact of biodiesel fueling on engine thermal efficiency. Bittle et al. [204]
tested palm olein biodiesel in a turbocharged common-rail direct injection medium-duty
diesel engine at different loads and speed, and only found decreased thermal efficiency at
high load, regardless of engine speed. From low to middle load, no significant differences
in thermal efficiency were found. Other than those, some authors also claimed synergies
when blending biodiesel. Labeckas and Slavingkas [205] tested a 4.75L engine under
different steady modes using 5%, 10%, 20%, 35% blends and pure rapeseed oil biodiesel.
The thermal efficiency appeared to reach a maximum for 5-10% blends. Murillo et al. [11]
conducted similar study with biodiesel from used cooking oil in a marine outboard 3-
cylinder naturally aspirated engine and found decreased thermal efficiency with blends
while highest thermal efficiency with pure biodiesel. Nevertheless, generally it is
accepted that biodiesel fueling won’t significantly affect diesel engine brake specific
thermal efficiency.
43
Engine thermal efficiency also relies on engine operating condition and control
strategy. Kowalewicz [206] showed that the engine brake specific thermal efficiency
increased with load with experiments in a single-cylinder naturally aspirated direct
injection diesel engine for both diesel and rapeseed oil methyl ester. Similar observations
have also been reported by Bittle et al. [204], and they also found that at low load the
brake thermal efficiency decreased with increased engine speed, but maintained at stable
level regardless engine speed at mid-high loads. Aoyagi et al. [207] investigated the
impact of boost pressure on engine thermal efficiency in a single-cylinder direct injection
diesel engine, and they found that increased boost pressure led to higher brake thermal
efficiency. Reddy and Ramesh [208] also investigated the impact of gas swirl on engine
thermal efficiency in a single-cylinder direct injection diesel engine at steady state, and
they observed no significant impact of gas swirl on thermal efficiency. Meanwhile, fuel
injection strategy has been observed to exert a significant impact on engine thermal
efficiency [63, 207]. Armas et al. [63] investigated engine thermal efficiency at different
single injection timings of -2º, 0 and 2º after TDC and observed decreased thermal
efficiency with retarded injection timing. Karra et al. [22] tested a John Deere 4045
HF475 4-cylinder turbocharged common-rail direction injection diesel engine using a
single injection strategy with start of injection ranging from -20º to 5º after TDC at steady
moderate speed and load. They found that the highest engine thermal efficiency was
observed when the start of injection is in the range of 10º to 5º before TDC, which led to
an apparent heat release peak at around TDC. Similar observations have also been
reported by Shudo et al. [209]. They conducted hydrogen assisted combustion in a 4-
cylinder spark ignition gas engine modified from an automobile gasoline engine (Nissan
44
CA 20S) and observed highest thermal efficiency at an apparent heat release peak close
to TDC. Fuel injection pressure can also affect engine thermal efficiency. Theoretical
development by Nagase and Fukushima [210] suggested that an increase in fuel injection
pressure can lead to higher thermal efficiency. However, Reddy and Ramesh [208] tested
various fuel injection pressure at different engine loads and observed that the maximum
engine thermal efficiency happened at moderate injection pressure, i.e., too high or too
low fuel injection pressure would have decreased thermal efficiency. In addition, thermal
efficiencies with post injection strategy have also been investigated and were often
reported to decrease from conventional engine combustion [160, 211]. Nevertheless, it
can be also improved by adjusting fuel injection strategy. In the low temperature
combustion work done by Fang et al. [212], an improvement of thermal efficiency was
observed when the starts of injection was retarded to 3º after TDC.
Although it is possible to investigate the strategy to improve engine thermal
efficiency through extensive engine tests by varying engine controls such as fuel injection
strategy, it requires tremendous effort to conduct such a test program, as well as, engine
modification for un-conventional combustion modes. Hence, theoretical development
might be a more appropriate approach in the attempt to improve engine thermal
efficiency. For example, Heywood [1] suggests that modern diesel engine pressure-
volume diagram can be represented with a “limit-pressure cycle”, as shown in Figure II-
4:
45
Figure II-4: A sketch of limit-pressure cycle
This limit-pressure cycle can be seen as a combination of Otto cycle and Diesel
cycle. It illustrates the maximum ideal thermal efficiency. From [1] one can calculate the
ideal indicated efficiency as:
1 αβ γ − 1
η f ,i = 1 − γ −1
[ ] (6)
rc αγ ( β − 1) + α − 1
where: rc = V4 / V2 is the compression ratio
γ = c p / cv is the ratio of heat capacity
and: α = p3 / p2 , β = V3b / V3a
Although the ratio of heat capacity γ is a weak function of temperature, the scale
of its variation is insignificant for this specific investigation. With an assumption of
46
constant γ value of 1.35 one can prepare an indicated thermal efficiency map as shown in
Figure II-5:
0.65
0.6
0.55
0.5
ηf,i
0.45
0.4
0.35
1
3
1.9
5 1.7
7 1.5
1.3
9 1.1
α
β
Figure II-5: Indicated efficiency map with α and β
It appears from Figure II-5 that β has a more significant impact on engine
efficiency than α. It can also be observed that: 1) increase of α can improve thermal
efficiency, although its effect is less obvious, 2) decrease of β close to 1 can increase
thermal efficiency significantly. Based on the above observations, an intuitive conclusion
can be drawn: to improve engine thermal efficiency, one should tend to reduce the
difference between V3a and V3b , i.e., adjust the limit-pressure cycle to Otto cycle.
This is actually suggesting that a maximum engine thermal efficiency may be
observed if the combustion heat release rate is a shape with its peak located at TDC, and
47
this understanding of location appears to be consistent with the highest thermal efficiency
locations of several experimental studies [22, 209]. However, engine energy conversion
is a much more complex process including combustion efficiency, heat loss, friction and
leakage. In fact, Hoppie et al. [32] showed that neither the Otto nor the Diesel cycle is
optimum for realistic engines. Meanwhile, the majority of existing theoretical researches
[210, 213-216] focused on parametric studies in energy efficiency. Therefore, in this
work an investigation of optimized heat release pattern design to improve engine thermal
efficiency is included, the conclusion of which could be used to direct the design of both
conventional and advanced combustion modes.
48
Chapter III. Technical Approach
3.1 Experiment apparatus
Engine tests were performed with an 8-cylinder 6.4L Ford “Powerstroke”
turbocharged direct injection diesel engine. The engine was equipped with two variable
geometry turbochargers and a common-rail fuel injection system (180 MPa maximum
injection pressure, 5 maximum injections within one combustion cycle). The injection
strategy was controlled through an electronic interface. The engine has a maximum brake
power of 261 kW at 3000 rpm, and a peak torque of 881 Nm at 2000 rpm. The
compression ratio is 17.2. The original engine calibration complied with EPA Tier 2
emission standards. Table III-1 summarizes the engine specifications:
Table III-1: Engine specifications
item Specifications
Engine type Direction injection 8-cylinder diesel engine
Bore 9.82 mm
Stroke 10.50 mm
Displacement 795.25 × 8 cm3
Compression ratio 17.20
Vc 49.09 cm3
Connect rod length 17.60 cm
Crank radius 5.25 cm
Injection system Common-rail, 180MPa maximum injection
pressure, 5 maximum injection pulse per cycle
Air charging Variable geometry turbocharger
Peak torque 880 N.m at 3000 rpm
49
3.2 Engine emission bench:
Gaseous emissions: The exhaust gases went through a heated filter and a heated
sample line, with both held at 190 ˚C. Exhaust gases contents were measured with an
AVL CEB-II emissions bench: NOx emissions were measured using an EcoPhysics
chemiluminescence analyzer, unburn hydrocarbons were measured using ABB flame
ionization detector and CO was measured by Rosemount infrared analyzer. For each
condition, the engine was operated for fifteen minutes to stabilize at the operating
condition. Data were recorded every fifteen seconds for a period of 20 minutes. The
measurements were performed three times for each condition to assess repeatability and
experimental uncertainty.
Particulate matter (PM) emissions: PM emissions were sampled through a
Sierra Instruments BG-3 micro-dilution tunnel using a dilution ratio of 10:1 at a total
flow rate of 75 slpm and a sample flow rate of 7.5 slpm for 5 min. The PM samples were
collect on Pallflex 47mm filters. The filters were weighed on a Sartorius M5P electronic
microbalance before and after sampling. The scale was located in a constant relative
humidity (45%) chamber. The filters were placed in the chamber 48 hour prior to mass
analysis before and after sampling. Three sample filters were taken per mode. The
average and standard deviation of the three samples was used to represent the result error.
Scanning mobility particle sizer (SMPS): A TSI 3936 SMPS was used to
analyze the size distribution of the PM. The SMPS instrument included a TSI series 3080
electrostatic classifier with a differential mobility analyzer (DMA), a series 3776
condensation particle counter (CPC), and a TSI series 3065 thermal denuder. A PC
running Aerosol Instrument Manager Software collected and managed the sampled data.
50
SMPS sampled diluted exhaust gas after BG-3 dilution tunnel, and it had an internal
dilution ratio of 10:1. The final dilution of the SMPS is 100:1 compared with raw exhaust
gas. Three samplings were conducted for each mode. The reported result is the average
value of three samples. In addition, a thermodenuder was used to remove the
hydrocarbon in the exhaust gases.
3.3 Test fuel
The baseline fuel for this test is ultra-low sulfur diesel fuel (ULSD). Tasks 1 and 2
used B40 blends (v/v %) using soybean methyl ester (SME)-based biodiesel obtained
from PeterCremer, L.P. Task 3 and 4 used B20 blends (20% v/v). Figure III-1 shows the
fatty acid methyl ester (FAME) composition of the SME (as neat biodiesel, B100)
measured by gas chromatography – mass spectrometry (GC-MS). The B100 used in the
experiment primarily consisted of methyl linoleate (C18:2), methyl oleate (C18:1),
methyl stearate (C18:0) and methyl palmitate (C16:0), the molecular structures of which
are shown in Figure III-2.
51
Figure III-1: FAME composition of B100 soybean oil methyl ester
52
Figure III-2: Molecular structures[217], names, abbreviations, chemical formulas,
and Chemical Abstracts Service (CAS) numbers of compounds comprising the
biodiesel used in this work.
3.4 Engine test conditions
Tasks 1 and 2: To simplify the problem of exploring the relationship between
injection strategy, biodiesel fueling and NOx emissions, a single fuel injection strategy
was used in Tasks 1 and 2. The engine was operated at the speed of 1500 rpm at 25%, 50%
and 75% of the maximum torque. The same injection timings were used for both test
fuels, and a total of seven injection timings were studied. At each start of injection, three
fuel injection pressures and injection durations were tested with each fuel. Table III-2
53
summarizes the engine operating parameters for Tasks 1 and 2. The fuel injection
pressures were chosen based on: 20% lower than default, default and 20% higher than
default when the engine was running on ultra low sulfur diesel (ULSD). The injection
duration was automatically adjusted by the electronic control unit (ECU) given that the
load and fuel injection pressure were both fixed, i.e., a higher fuel injection pressure led
to shortened injection duration. When B40 is used, both the injection pressure and
duration are increased by default because of the higher rate of fuel consumption to
achieve the same power.
Table III-2: Engine operating parameters in Task 1 and 2
Parameter Value
Speed (RPM) 1500
EGR ratio (%) 10.6 ± 2.5
Number of Injection 1
Boost pressure (Bar) 1.947 ± 0.1
Load (N.m) 220 475 660
Start of Injection (˚ATDC) -9,-5,-1,3 -9,-7,-5,-3,-1,1,3 -9,-5,-1,3
52 72 87
Injection pressure, with ULSD (MPa) 66 90 108
80 108 131
52 72, 80±1 87
Injection pressure, with B40 (MPa) 66 90, 98±1 108
80 108, 121±1 131
Task 3: one 100-hour test was performed with the same engine. During the test
period, the engine was operated at the condition of 1500 rpm with 270 N.m torque (30 %
load). A stable EGR level around 10% was used during this test. The engine started with
fresh lubricant, which is obtained from Infineum USA L.P. The engine operation was
divided into multiple time blocks ranging from 3 to 8 hours to accumulate the 100 hour
test duration. At the end of each test block 50 ml oil sample was taken. The engine
control setting for the post injection was adjusted during the test, as shown in Table III-3.
54
The default engine control only included a pilot and a main injection. The introduction of
post injection brought the total number of injections to three, and the injection quantities
and timings of the pilot and main injections were controlled by the engine’s default
calibration. The timing of the post injection ranged from 45º to 70º after top dead center
(TDC) to cover the zone at which both late combustion and lubricant dilution were
observed. Engine states, e.g., exhaust temperature and brake specific fuel consumption
(BSFC), were recorded every two hours, and the data were averaged within one test block.
An extra post injection experiment at 45º after top dead center (ATDC) with ULSD was
conducted separately to provide a comparison to the biodiesel fuel.
Table III-3: Engine operating parameters in Task 3
Engine Speed 1500 rpm

Engine Torque 270 N.m (30% load)
0h – 32h 45º ATDC at 5mg/stk
Post Injection
Configurations
56h – 75h 70º ATDC at 5 mg /stk
75h – 100h NONE
Task 4: The same engine tested in Tasks 1 – 3 will also be used in Task 4.
Similar to Task 3, B20 soybean oil based biodiesel and ULSD blends will also be used in
this task. Table III-4 summarizes the operating conditions. Two load conditions were
tested. Single injection at 1º before top dead center was used (the injection timing was
chosen based on default engine calibration). Three fuel injection pressures for each load
were tested.
55
Table III-4: Engine operating parameters in Task 4
Engine Speed 1500 rpm

Engine Torque 270 N.m (30% load) and 530 N.m (60%
load)
Start of Injection 1º before top dead center, single injection
Fuel injection pressure 50, 75 and 100 MPa for 30% load, 75, 100
and 125 MPa for 60% load
A vacuum pump was used to draw exhaust gas through a teflo filter for soot
collection. Before sampling, the engine was maintained at steady state for 15 minutes.
The time for each collection was 45 min. Soot was gently scratched off the filter after the
collection.
Diesel particulate were collected onto transmission electronic microscopy (TEM)
grids with a thermophoretic sampling apparatus, as shown in Figure III-3. The TEM grid
was mounted at the tip of the apparatus, which will be forced into raw exhaust gases
during engine running. The sampling time is around 0.6 second, and the grid was
controlled to be parallel to the flow direction of prevent direct particle compact.
56
Figure III-3: Thermophoretic sampling apparatus
Task 5: This task was conducted with a zero-dimensional thermodynamic
simulation with Matlab. A detailed description will be presented in a later chapter.
3.5 Biodiesel lubricating oil dilution
Biodiesel oil dilution was measured with a Fourier transform infrared
spectrometer (FTIR). It was conducted with a Nicolet Avatar 360 FTIR system. First of
all, a background (or air) spectrum was obtained. 1 ml sample of used oil was then placed
on the device crystal and the obtained spectrum was subtracted with the background
spectrum. The scanning region is from 500cm-1 to 4000 cm-1. The raw spectrum was then
mathematically converted to the absorption spectrum. FTIR can measure the level of
biodiesel dilution in lubricant and identify the functional groups in lubricating oil.
57
Figure III-4 shows the FTIR spectrums of engine oil with different artificially
biodiesel content. It can be observed that with higher biodiesel dilution, the intensities of
peaks at around 1746 cm-1 and 1180 cm-1 tend to increase. This observation can be used
to identify and quantitatively measure the biodiesel dilution of any used oil.
Figure III-4: FTIR spectrums for oils with different biodiesel dilution.
The authors artificially mixed 3%, 5%, 8% and 10% v/v diluted lubricant
standards with B100 and measured their IR absorbance. A linear correlation between the
absorbance of the peak at 1746 cm-1 and dilution level was observed, as shown in Figure
III-5, Equation (7) was regressed from the correlation and used to calculated the oil
dilution level.
Dilution = 33.161 × I1746 − 0.192 (7)
58
10
Y = 33.161*x-0.192
R-Square = 0.9977
8
Dilution Level (%)
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Absorbance of Peak 1746 cm-1
Figure III-5: Correlation between absorbance of Peak 1746 cm-1 from FTIR and
dilution level
Total acid number (TAN) of the oil sample was measured according to ASTM
standard method D974 based on a color indication titration method. In addition, the used
lubricant after 100-hour test was distilled at 432 ˚C (which was the highest boiling point
of the biodiesel that was used, determined by simulated distillation method). The lighter
cut was collected and analyzed with GC-MS. Hydrocarbons from the engine exhaust
were collected in chilled dichloromethane and also analyzed with GC-MS. A detailed
description of the collection method can be found in Ref.[218]
Mechanical friction mean effective pressure (mfMEP) during engine operation
was calculated with Equation (8), which was based on the suggestion by Heywood [1].
mfMEP = IMEPg − BMEP − PMEP (8)
59
where IMEPg is the gross indicated mean effective pressure, BMEP is brake mean
effective pressure and PEMP is the pumping loss.
3.6 Soot characterization
3.6.1 Removal of organic fraction
Since raw soot particles collected from engine exhaust often adsorb certain
amount of hydrocarbon on the surface, which may affect instrument, a thermal treatment
procedure was employed before the analysis of soot particle. The raw particulates were
placed in a Perkin Elmer TGA 7 thermogravimetric analysis under nitrogen environment.
The temperature was ramped at a speed of 5ºC/min to 450º, and maintained at isothermal
condition for 30 min. The residual soot after this thermal treatment can be used to
conduct other experiments.
3.6.2 Thermogravimetric analysis (TGA)
An SDT Q600 from TA Instruments thermogravimetric analyzer was employed to
evaluate soot oxidative reactivity. Each diesel raw PM sample was placed in a 90 µL
alumina sample crucible and was heated at a rate of 10ºC/min to 500ºC and kept for 60
min to drive off volatile compounds under N2 gas in a flow rate of 100 cm3/min. After
30º min isothermal treatment, the sample was heated at a rate of 5ºC/min to 550ºC, and
then N2 gas was switched to ultra zero air (99.8% purity) at a flow rate of 100 cm3/min
for isothermal experiments. The mass loss of each sample during isothermal oxidation
was normalized with respect to the weight after thermal treatment, and the normalized
mass loss was compared for each sample to evaluate soot oxidative reactivity. Two
60
repeated tests were conducted for each sample, and the reported result was an average of
the two tests.
3.6.3 Transmission Electron Microscopy (TEM)
The grids collected from thermophoretic sampling were loaded onto a single-tilt
holder and analyzed in the TEM mode using a JEOL EM-2010F with field emission
source at 200 kV accelerating voltage. All images were taken in a Gatan
DigitalMicrograph with a slow scan CCD camera using 150 µm of condenser aperture
size.
3.6.4 Morphology analysis
The procedure for morphology analysis follows the work by Lapuerta et al. [168],
where a more detailed description can be found. Given this analysis requires a large
amount of sample to be statistically significant, TEM samples were examined to produce
more than 25 images at random locations on the samples to characterize many soot
agglomerates. Furthermore, 10 of the total images for each sample, which tend to have
clear structure of primary particles, were chosen. For each chosen images, 10 of random
primary particles were chosen to measure the size. A totally 100 primary particles were
measured, and the average and derivation were calculated.
The fractal dimension analysis tested all of the images obtained. These images
require a de-background procedure first. After removal of the background, with input on
primary particle size combined with an iteration procedure, the fractal dimension of each
sample can be obtained.
61
3.6.5 Raman Spectroscopy
A confocal Raman spectrometer (WITec CRM200) was used to determine the
degree of disorder of the soot samples. The excitation laser was an Argon ion laser
(λ=514.5nm), focused onto the sample using a microscope objective lens (100x or 40x).
The scattered light was collected by the same objective lens collected in a backscattering
configuration (180 degree). The microscope is equipped with a XY stage, driven by
piezo-electric actuators, for selecting the sample area of interest. The spectrum collection
method is configured by setting the integration time (seconds), software accumulations,
and hardware accumulations in the software WITec Project. All spectra were collected at
the same source power level, and post-processed via a software package (Igor Pro 6.22,
Wavemetrics Inc.). The software was used for the determination of spectral parameters.
The Raman spectra were first corrected by a linear baseline. Then, the corrected spectra
were curve fitted by three Lorentzian-shaped bands (1580, 1350, 1200 cm-1) and one
Gaussian-shaped band (1500 cm-1) based on the approach by Seong and Boehman [189,
190], as shown in Figure III-6. The ratio of AD1/AG and D1 FHWM can represent the
density of edge sites [219] and the distribution of crystalline sizes [220], respectively.
According to an early research [190], those two values correlate with the soot oxidative
reactivity, hence they will be employed for the investigation in this work.
A total of five locations for each sample were investigated. The reported results
were averaged data from the four locations.
62
Figure III-6: Demonstration of curve fitting for Raman spectrum
63
Chapter IV. Investigation of injection strategy and biodiesel
fueling on diesel engine emissions and performance
4.1 Impact of injection timing and fuel injection pressure on NOx

emissions
Figure IV-1 presents the NOx emissions at 25% load for ULSD and B40 both
with different start of injection timings and fuel injection pressures. Consistent with
previous research [21, 22, 55], at the same SOI the NOx emissions for both B40 and
ULSD increase with fuel injection pressure. Meanwhile, at the same fuel injection
pressure the advance of SOI can significantly increase NOx emissions, which is also
consistent with previous studies [23]. At the same SOI and fuel injection pressure, B40
has significantly higher NOx emissions than ULSD, which is also expected [10, 61]. One
can observe that the difference in NOx emissions due to the change of fuel injection
pressure tends to decrease with retarded SOI, and at some SOIs (e.g., 9˚ and 5˚ BTDC)
the NOx emissions for B40 at low fuel injection pressure can be lower than that of ULSD
at higher fuel injection pressure.
64
Figure IV-1: NOx emissions at 25% load. Four injection timings (9˚, 5˚, 1˚ before
top dead center and 3˚ after top dead center) were tested for both ultra low sulfur
diesel (ULSD) and B40 blend. Three fuel injection pressures (52, 60 and 80 MPa)
were tested at each injection timing
Figure IV-2 presents the NOx emissions at 50% load. To increase statistical
significance, more SOIs were tested in the range of interest (9˚ BTDC to 3˚ ATDC).
Similar to 25% load, at the same SOI, higher fuel injection pressure leads to increased
NOx emissions, and at the same fuel injection pressure, advancing SOI leads to increased
NOx emissions. B40 significantly increases NOx emissions compared with ULSD with
the same injection strategy. The additional fuel injection pressures considered for B40
were for the purpose of comparing the effects of fuel injection pressure and injection
duration. For instance, to compare the NOx emissions for ULSD at 90 MPa fuel injection
pressure, both 90 and 98 MPa fuel injection pressure for B40 are tested: matched fuel
injection pressure can lead to prolonged injection duration, and matched injection
duration requires increased fuel injection pressure, which is due to the increased fuel
65
consumption with biodiesel. It can be seen that matched injection duration increases NOx
emissions more than matched fuel injection pressure, which suggests that fuel injection
pressure plays a more significant role in the impact on NOx emissions from injection
strategy. The difference of NOx emissions from the change of fuel injection pressure also
tends to decrease with retarded SOI, which is consistent with 25% load. Furthermore,
lower NOx emissions for B40 than ULSD can also be observed for some cases at the
same SOI due to the difference in fuel injection pressure, e.g., at the SOI of 9˚ and 7˚
BTDC, NOx emission of B40 at 72 MPa fuel injection pressure is lower than that of
ULSD at 108 MPa fuel injection pressure.
Figure IV-2: NOx emissions at 50% load. Four injection timings (9˚, 7˚, 5˚, 3˚ 1˚
before top dead center and 1˚, 3˚ after top dead center) were tested for both ultra
low sulfur diesel (ULSD) and B40 blend. Three fuel injection pressures (72, 90 and
108 MPa) were tested at each injection timing for ULSD, and six fuel injection
pressures (72, 80, 90, 98, 108 and 118 MPa) were tested at each injection timing for
B40.
66
Figure IV-3 presents the NOx emissions at 75% load. The impacts of SOI, fuel
injection pressure and biodiesel on NOx emissions are consistent with those at 25% and
50% loads. The difference in NOx emission due to changes of fuel injection pressure
decreases with retarded SOI, and at certain SOI the NOx emissions of B40 is also lower
than ULSD as the fuel injection pressure decreases. For instance, at the SOI of 9˚ and 5˚
BTDC, NOx emission of B40 at 87 MPa fuel injection pressure is lower than that of
ULSD at 131 MPa fuel injection pressure.
Figure IV-3: NOx emissions at 75% load. Four injection timings (9˚, 5˚, 1˚ before
top dead center and 3˚ after top dead center) were tested for both ultra low sulfur
diesel (ULSD) and B40 blend. Three fuel injection pressures (87, 109 and 131 MPa)
were tested at each injection timing.
Overall, the impact of SOI, fuel injection pressure and biodiesel on NOx
emissions are consistent regardless of engine load. The general observations include: 1)
for both ULSD and B40, NOx emissions increase with advanced SOI; 2) NOx emissions
67
increase with increased fuel injection pressure; 3) B40 has higher NOx emissions than
ULSD at the same SOI and fuel injection pressure; 4) the difference in NOx emissions
due to fuel injection pressure becomes smaller with retarded SOI; 5) at advanced SOI (e.g.
before 3˚ BTDC) it is possible to reduce the NOx emission of B40 below ULSD by
decreasing the fuel injection pressure.
4.2 Impact of injection timing and fuel injection pressure on

particulate matter (PM) emissions
The level of particulate matter emissions is also an important concern for diesel
engines. Figure IV-4 presents the total PM emissions at 25% load for ULSD and B40,
both at different start of injection timings and fuel injection pressures. The general trend
is that the PM emissions increase with retarded SOI or decreased fuel injection pressure,
which is consistent with previous research [21] and consistent with the trend for NOx
emissions (i.e. PM-NOx trade-off [15]). At the same SOI and fuel injection pressure, the
PM emissions for B40 are lower than ULSD, which is also expected [6, 72, 201]. An
interesting observation is that the reduction of PM emissions with biodiesel at lower fuel
injection pressure is more significant than that at higher fuel injection pressure, and the
reduction of PM emissions at retarded SOI is also more significant than that at advanced
SOI. This is surprising given that such differences have not been observed for NOx
emissions, but it may suggest that the reduction of PM emissions by biodiesel is more
significant during slower combustion (a low and wide peak in the apparent heat release).
One may note the unusual decrease of PM emissions when retarding SOI from 1˚ BTDC
68
to 3˚ ATDC at a fuel injection pressure of 52 MPa. This could be due to a difference in
the soluble organic fraction (SOF) [15].
Figure IV-4: PM emissions at 25% load. Four injection timings (9˚, 5˚, 1˚ before top
dead center and 3˚ after top dead center) were tested for both ultra low sulfur diesel
(ULSD) and B40 blend. Three fuel injection pressures (52, 60 and 80 MPa) were
tested at each injection timing.
An interesting observation is that the retarding of SOI timing from 1 ºBTDC to 3º
ATDC slightly decreased the PM emissions for both ULSD and B40 at the lowest fuel
injection pressure. Since the PM emissions come from a combined adverse effects of PM
formation and oxidation, and a retarding of SOI timing could increase the exhaust
temperature and consequently promote PM oxidation while increasing PM formation.
These combined competing effects may be the reason of slight decrease of PM effect at
retarded SOI timing (3 ºATDC to 1 ºBTDC).
Figures IV-5 and IV-6 show the PM emissions at 50% and 75% load,
respectively. A consistent observation with 25% load is that the PM emissions increase
69
with retarded SOI or decreased fuel injection pressure. However, contrary to 25% load,
no significant difference in PM emissions can be observed between B40 and ULSD. At
75% load with fuel injection pressures of 109 and 131 MPa, the PM emissions for B40
are even slightly higher than ULSD. Similar observations have also been reported.
Lapuerta et al.[79] showed that at low load biodiesel could introduce a larger PM
decrease. In their review on rapeseed oil biodiesel, Krahl et al.[82] noted that the PM
emissions reductions were lower (or no significant reductions) in heavy-duty engines
than in light-duty engines. This may also suggest that the reduction of PM emissions by
biodiesel is more significant during slower combustion. The slightly higher PM emissions
of B40 at 75% load may be due to the increase of SOF with biodiesel,[6, 69, 74, 77]
which is probably caused by the lower volatility of the unburned hydrocarbons. It also
appears that the impact of SOI or the fuel injection process on PM emission becomes less
significant with increased load. For instance, the impact of SOI and fuel injection
pressure on PM emissions for B40 is less significant and consistent at 75% load (Figure
IV-6).
70
71
4.3 NOx-PM trade-off
An important evaluation of diesel engine emissions is obtained by comparing the
NOx-PM emissions trade-off. If the trade-off is not improved, a reduction of one
emission always leads to an increase of the other emission. Figure IV-7 presents the
NOx-PM emissions trade-off at 25% load for ULSD and B40 both at different start of
injection timings and fuel injection pressures. It can be seen that at this load, the NOx and
PM emissions generally fall into the same trade-off curve, regardless of fuel, start of
injection or fuel injection pressure. The trade-off curve shows a hyperbolic behavior, i.e.,
at high PM (NOx) emissions points, a significant change in the PM (NOx) emissions by
72
controlling the fuel injection pressure only has an insignificant impact on the NOx (PM)
emissions. The lowest trade-off curve location (closest to the origin) is that at the SOI of
3˚ ATDC: 1) for B40 the best fuel injection pressure is moderate, or 66 MPa in this case;
2) for ULSD the best fuel injection pressure is moderate to high, or 66 – 80 MPa in this
case.
Figure IV-7: NOx-PM emission trade-off at 25% load. The legend shows the fuel
and injection timing for each test. Three fuel injection pressures (52, 66 and 80 MPa)
were tested at each injection timing for each fuel. Fuel injection pressure is not
distinguished in the plot.
Figure IV-8 presents the NOx-PM emissions trade-off at 50% load. One can
observe that the hyperbolic feature of the trade-off curve is weakened, i.e., the test points
become more scattered. Figure IV-9 presents the NOx-PM emissions trade-off at 75%
load. The scattering becomes more significant, and the trade-off seems able to be
improved with engine injection control. For instance, at 50% load ULSD at 1˚ BTDC SOI
73
with fuel injection pressure of 108 MPa has both lower NOx and PM emissions than at 9˚
BTDC SOI with fuel injection pressure of 90 MPa. At 75% load ULSD at 1˚ BTDC SOI
with fuel injection pressure of 90 MPa has both lower NOx and PM emissions than at 9˚
BTDC SOI.
and injection timing for each test. Three fuel injection pressures (72, 90 and 108
MPa) were tested at each injection timing for each fuel. Fuel injection pressure is
not distinguished in the plot
74
and injection timing for each test. Three fuel injection pressures (87, 109 and 131
MPa) were tested at each injection timing for each fuel. Fuel injection pressure is
not distinguished in the plot
From the previous NOx emission analysis (Figures IV-1 to IV-3), it can be seen
that decreasing the fuel injection pressure is an available method to counter biodiesel
NOx effect with the start of injection timing at 9˚ BTDC. Although this manipulation in
injection strategy may introduce an increase of PM emissions, it may then be countered
by biodiesel’s PM effect. To more clearly understand its impact, the NOx –PM trade-off
for only the SOI timing at 9˚ BTDC at different loads are summarized, as shown in
Figure IV-10. At 25% load (Figure IV-10(a)), the decrease of fuel injection pressure
does not significantly increase the PM emissions for B40, and B40 with a fuel injection
pressure of 52 MPa has the same NOx and total PM emissions as ULSD with a fuel
injection pressure of 66 MPa. At 50% load (Figure IV-10(b)), the total PM emissions for
75
B40 are generally lower than those of ULSD. Therefore, B40 with decreased fuel
injection pressure appears to be the lowest NOx-total PM emissions combination for this
condition. The generally higher total PM emissions for B40 at 75% load (Figure IV-
10(c)) may be due to the higher hydrocarbons adsorbed on soot surfaces. Hence, the
authors performed a Soxhlet process to clean the solute organic fraction (SOF) of the PM.
Figure IV-10(d) presents the NOx-soot emission (both PM filters for ULSD and B40
have been cleaned with Soxhlet) trade-off for both fuel. It can be seen that the soot
emissions for B40 is generally lower than those for ULSD, verifying a higher SOF, i.e.,
hydrocarbons adsorbed on the soot surface for B40. Similar to cases at lower loads, the
decrease of fuel injection pressure for B40 is an effective way to counter the biodiesel
NOx penalty without resulting of higher soot emissions than ULSD.
76
Figure IV-10: PM-NOx emission trade-off at the start of injection timing at 9˚ before top dead center for ultra low sulfur
diesel and B40. (a) 25% load; (b): 50% load; (c): 75% load; (d): 75% load after Soxhlet of PM filters
77
Overall, for the case at early SOI timing, decreasing the fuel injection pressure for
biodiesel is an effective way to solve the biodiesel NOx penalty while maintaining the
level of PM emissions. It provides a fast and easily applicable way to implement
biodiesel in a modern engine without a significant alteration of engine emission profile.
Although the NOx-PM trade-off at early SOI may not be the best trade-off point (Figures
IV-7 to IV-9 show the best trade-off points are around the SOI timing at 1˚ BTDC), the
significance of earlier SOI will be more apparent if the engine brake thermal efficiency is
considered, which will be discussed in the following section.
4.4 Brake thermal efficiency
The brake thermal efficiency investigated here is the percentage of fuel energy
that is converted into engine brake energy. Figure IV-11 presents the thermal efficiency
at 25% load for ULSD and B40 both at different start of injection timings and fuel
injection pressures. At the same SOI, the variation of fuel injection pressure investigated
in this work (52 to 80 MPa) does not have a significant impact on thermal efficiency for
either ULSD or B40. Although it appears that B40 has a slower brake thermal efficiency
than ULSD, the differences are in the standard deviation zone. Meanwhile, the thermal
efficiency tends to decrease with retarded SOI after 5˚ BTDC, suggesting decreased
efficiency at retarded SOI.
78
Figure IV-11: Brake thermal efficiency at 25% load for ultra low sulfur diesel (left) and B40 (right). The multiple points at
each start of injection timing are due to the multiple fuel injection pressures.
79
The same trend has also been observed for higher load. Figures IV-12 and IV-13
present the thermal efficiency of ULSD and B40 at 50% and 75% load, respectively. It
can be shown that the highest engine brake efficiency seems to be at the SOI between 9˚
to 5˚ BTDC, and the efficiency decreases with retarded SOI later than 5˚ BTDC.
Similarly, no significant difference in thermal efficiency between ULSD and B40 can be
observed, nevertheless, engine brake thermal efficiency seems to be higher for B40 at
high load. In addition, the impact on thermal efficiency from variation of SOI is more
significant at higher load (Figures IV-12 and IV-13).
80
81
82
Given no significant difference in brake thermal efficiency from the variation of
fuel injection pressure and biodiesel fueling, and higher brake thermal efficiency can only
be observed with early SOI, for a modern engine which intends to run at a high thermal
efficiency the biodiesel NOx penalty can be fixed with decreased fuel injection pressure.
4.5 Apparent heat release analysis
The apparent heat release profile can provide valuable information about
combustion phasing. Figure IV-14 presents the apparent heat release profiles for ULSD
and B40 at 25% load with different SOI and fuel injection pressures. The single peak in
apparent heat release profile suggests that the fuels are primarily consumed by premixed
combustion. The shift of heat release profile is consistent with the change of SOI. Similar
to previous observations [55], at the same SOI, variation of fuel injection pressure can
significantly alter the heat release profile, i.e., higher fuel injection pressure can lead to
earlier start of combustion (SOC) and higher, as well as, narrower heat release peak. The
earlier SOC and higher heat release peak yields longer residence times and/or higher in-
cylinder temperature, leading to an increase in NOx emissions [18]. This can be
explained by the increased fuel droplet velocity [105, 106] and decreased droplet size
[221] due to increased fuel injection pressure, which can lead to better overall mixing
between fuel and air and shortened ignition delay. Higher heat release rate can introduce
higher in-cylinder temperature, yielding increased NOx emissions. It further confirms the
significance of residence time and higher in-cylinder temperature at lower load [18, 55].
e.g., the timing of maximum cylinder temperature [14].
83
When the fuel is injected before TDC, the apparent heat release profiles are
similar with the same fuel injection pressure regardless of SOI, i.e., a similar highest heat
release rate with a noisy long tail (diffusion flame). However, significant differences can
be observed for the fuel injected after TDC. At the same fuel injection pressure, the
maximum heat release rate of fuel with an SOI after TDC is apparently higher than those
with an SOI before TDC. Furthermore, the “tail” after the main combustion event is also
smaller than those with SOI before TDC. These observations suggest that the fraction of
premixed combustion of fuel with a SOI after TDC is higher than those with the SOI
before TDC. This can be explained in that the decreasing in-cylinder temperature and
pressure due to piston motion during expansion stroke can prolong the ignition delay,
resulting in longer time for premixing.
84
Figure IV-14: Apparent heat release profile of different fuels and different fuel injection pressures at 25% load. The SOI: a)
9˚BTDC, b) 5˚BTDC, c) 1˚BTDC, d) 3˚ATDC.
85
Figure IV-14 also shows the impact of biodiesel on the heat release profile. At
the same fuel injection pressure, B40 tends to have an earlier SOC, and this advancing of
SOC is more significant with later SOI. For instance, at the SOI of 9˚ BTDC the SOC for
B40 and ULSD are similar to each other, but at the SOI of 3˚ ATDC the SOC for B40 is
earlier than for ULSD. This may be attributed to the slightly higher cetane number of
biodiesel, which will lead to a shorter ignition delay, and its impact will be more
observable at prolonged ignition delay condition, i.e., with the SOI after TDC. In addition,
the maximum heat release rate of B40 seems to be lower than that of ULSD at the same
SOI and fuel injection pressure. This decreased heat release is widely observed for
biodiesel in premixed combustion [34, 91, 222, 223], which is mostly attributed to its
lower volatility in addition to the shorter ignition delay [44]. It is also consistent with the
theory proposed by Mueller et al. [18] given the oxygen content in biodiesel can make the
mixture at autoignition zone closer to stoichiometric during ignition, which can lead to
more rapidly increased in-cylinder temperature and consequently to a more advanced
combustion event (shorter ignition delay), as well as higher NOx emissions.
Figure IV-15 presents the apparent heat release profiles of ULSD and B40 at 50%
load. Unlike at 25% load, the diffusion burn constitutes the dominate fraction of
combustion at 50% load. The heat release profiles at different SOIs are generally similar,
and the shifting of the profile is consistent with the variations in SOI as well. At the same
SOI, the impact of fuel injection pressure is also similar to 25% load, i.e., higher fuel
injection pressure can lead to higher maximum heat release rate and narrower heat release
peak (for both premixed and diffusion burns). Higher fuel injection pressure can also
advance the SOC, and the work presented here confirms it by expanding the applicable
86
range of SOI from 9˚ ~ 5˚ BTDC [55] to 9˚ BTDC ~ 3˚ ATDC. Similar to 25% load, the
heat release peak of premixed burn is higher at the SOI of 3˚ ATDC, suggesting a
prolonged ignition delay. Meanwhile, the heat release profiles for B40 and ULSD with
the same SOI and injection pressure are mostly overlapping with each other. This
suggests that biodiesel does not have a significant impact on the combustion phasing. The
same observation has been reported by Zhang and Boehman [13], where the same authors
have shown that the SOI will not be affected by the biodiesel using a common-rail
injection system, and the slight difference in cetane number between ULSD and B40 will
not be strong enough to alter the combustion phasing. In addition, from the presented
work, it is suggested that the difference in combustion phasing between B40 and ULSD
is smaller at higher load.
87
9˚BTDC, b) 5˚BTDC, c) 1˚BTDC, d) 3˚ATDC.
88
Figure IV-16 presents the apparent heat release profiles for ULSD and B40 and
75% load. The fraction of diffusion burn is further increased. Similar to Figures 16 and
17, the heat release profiles at different SOIs are generally similar, and the shifting of the
profile is consistent with the variations in SOI. The fraction of premixed burn is stable
with the SOI from 9˚ to 1˚ BTDC, but higher with the SOI of 3˚ ATDC. At the same SOI,
higher fuel injection pressure can lead to advanced SOC, higher maximum heat release
rate and narrower heat release peak (for both premixed and diffusion burns). In addition,
the heat release profiles for B40 and ULSD are also similar to each other with the same
SOI and fuel injection pressure, suggesting the insignificant impact of biodiesel on
combustion phasing is valid for higher load as well.
89
˚BTDC, b) 5˚BTDC, c) 1˚BTDC, d) 3˚ATDC
90
In summary, it is observed that both fuel injection strategy and biodiesel
properties can significantly affect diesel engine NOx emissions. Advance of start of
injection or increase in the fuel injection pressure can significantly increase NOx
emissions. Both matched fuel injection pressure and matched fuel injection duration tests
with biodiesel fueling significantly increase NOx emission, and the matched fuel
injection duration (higher fuel injection pressure) has more increased NOx emissions than
the matched fuel injection pressure case. This suggests that fuel injection pressure has a
more significant impact on the NOx emissions than the fuel injection duration with
biodiesel fueling. In addition, the impact of biodiesel’s properties is even most significant
factor since a big increase in NOx emissions can be observed between diesel and B40
with the same fuel injection pressure at the same start of injection. It is also observed that
for engine with advanced start of injection (to achieve high engine thermal efficiency),
the decrease of fuel injection pressure can be an effective method to decrease the NOx
emissions when biodiesel is used. This approach can reduce the NOx emissions of engine
fueling with biodiesel blends back to the same level of NOx emissions of engine fueling
with diesel fuel. Because biodiesel generates less particulate matter (PM) emissions,
although the decrease of fuel injection pressure may introduce increased PM emissions, it
will still be at the same level as / or even lower than the PM emissions of engine fueling
with diesel. However, this approach is more effective for engine running at conventional
high efficiency mode (fuel injection timing earlier than 5º before TDC). Therefore, the
hypothesis that biodiesel NOx emission penalty can be eliminated by fuel injection
strategy is conditionally validated.
91
Chapter V. Origin of biodiesel NOx effect
As discussed in the Introduction, the behavior of the fuel spray in the autoignition
zone may explain the origin of the NOx increase in this experiment. Siebers et al. [105,
106, 142-144] have comprehensively developed a numerical model that describes the
equivalence ratio distribution in a fuel spray at the end of the premixed combustion.
Based on their analysis, the cross-sectional average equivalence ratio in a non-reacting,
isothermal jet at any position, x, along its axis is given by:
2( A / F ) st
φ ( x) = (9)
x
1 + 16( + ) 2 − 1
x
where (A/F)st is the stoichiometric ambient-gas/fuel ratio by mass and x+ is the
penetration length scale for the jet, which is defined as:
ρ f d Ca
x+ = (10)
ρ a a tan(θ )
2
where ρf is the density of injected fuel, ρa is the density of air, d is the orifice diameter, Ca
is the area-contraction coefficient of the orifice which varies with injection pressure, a is
a constant with a value of 0.75[145], and θ/2 is the spreading half-angle of the jet. The
spreading half-angle of the jet has been shown to be a function of ρf and ρa [105]:
θ ρ ρ
tan( ) = c[( a ) 0.19 − 0.0043 f ] (11)
2 ρf ρa
Here c is a constant for a defined orifice diameter which is correlated with
injection pressure [105]. The equivalence ratio from Equation (9), however, does not
92
account for oxygen heteroatoms, if any, present in the fuel. Therefore the oxygen
equivalence ratio φΩ is used to characterize the mixture stoichiometry. The oxygen
equivalence ratio is defined as the amount of oxygen required to convert all carbon atoms
to CO2 and all hydrogen atoms to H2O, divided by the amount of total oxygen available
in both reactants:
1
2nC + nH
φΩ = 2 (12)
nO
where nC, nH and nO are the numbers of carbon, hydrogen and oxygen atoms in all the
reactants, respectively. Therefore, the oxygen equivalence ratio and equivalence ratio
have this relation:
φ
φΩ = (13)
1−φ
1−
φΩ , f
where φΩ,f is the oxygen equivalence ratio of the fuel alone. It can be shown that for
diesel fuel, φΩ,f is infinitely large, or φΩ = φ. Table V-1 summarizes the input parameters
for the calculation.
93
Table V-1: Input parameters during calculation
Parameter ULSD B40

a
Chemical Structure C14.48H26.23 C16.15H30.38O0.8
3 b
ρf (kg/m ) 843 856
Zstc 0.0099 0.0089
dd (mm) 0.111
Cae 0.91+7.5785*10-5*(Pf-720)
Cdf 0.8
g
c 0.255
ρah (kg/m3) at 40˚C 1.14
Lower heating value
42.7 40.6
(MJ/kg)i
Cetane numberj ~45 ~47
Kinematic viscosity(cSt) 2.25 3.01
a: B100 obtained from GC-MS results, the chemical formula of diesel fuel was calculated
by the suggestion of C/H = 0.552[18] and the average molecular weight of around 200
g/mol[224]; b: density; c: stoichiometric mixture fraction; d: orifice diameter; e: obtained
and correlated with injection pressure (Pf) from[105]; f: discharge coefficient; g:
spreading angle constant; h: air density; i: obtained from GREET fuel cycle analysis
model[225]; j: based on the data from[18]
The local gas density at fuel injection is calculated as:
PI V
ρa = ⋅ ρ aφ ⋅ B (14)
Pa VI
where PI is intake pressure, Pa is atmospheric pressure, ρaφ is air density at 40˚C at 1 bar.
VB is the chamber volume at bottom dead center; VI is the chamber volume at the
injection crank angle. It is assumed that at the intake the air is ideal gas and the boost
temperature is constant at 40˚C.
The radial variation of equivalence ratio can be obtained given a known centerline
axial variation based on the suggestions by various researchers [105, 142, 226] as:
r ln(0.08)
φΩ ( x, r ) = 1.3φΩ ( x) exp[( )2 ] (15)
x tan 2 (θ / 2)
94
where φΩ ( x) was calculated from Equations (9) and (13), r is the radial coordinate, and
tan(θ/2) was calculated with Equation (11). The lift-off length is calculated based on
Siebers and Higgins’s model[145, 227]
H = 7.04 *108 Tg −3.74 ρ a −0.85d 0.34U f Z st −1 (16)
where Tg is cylinder temperature at fuel injection, which was obtained by heat release
calculation, ρa is the air density from Eq.(14), d is the orifice diameter, Zst is the
stoichiometric mixture fraction for fuel, and Uf is the velocity of injected fuel based on
Bernoulli’s equation:
Pf − Pa
U f = Cv 2( ) (17)
ρf
where Cv is the ratio of discharge coefficient Cd and area-contraction coefficient Ca [105].
The results of the computation are shown as follows. Figure V-1 illustrates the
oxygen equivalence ratio variation of the fuel spray and lift-off length for diesel and B40
at different injection pressures at the SOI of 5 ºBTDC. Figure V-1 shows that the change
of injection pressure does not have a significant impact on the oxygen equivalence ratio
field. The lift-off length, however, significantly increases when the injection pressure
increases due to the increase of injected fuel velocity, which results in an oxygen
equivalence ratio that is closer to stoichiometric at the autoignition zone near the lift-off
length. This explains the NOx increase when the injection pressure was increased, which
is also consistent with the observation of higher apparent heat release rate. When B40
was used at the same injection pressure, the equivalence ratio field shrinks and the lift-off
length slightly increases, leading to an equivalence ratio that is closer to stoichiometric at
autoignition, which is consistent with the observations of Mueller et al.[18] regarding the
95
biodiesel NOx increase. Note that the oxygen equivalence ratio at AZ of diesel at 108
MPa injection pressure is close to that of B40 at 72 MPa injection pressure, which may
explain the similarity of NOx emissions level between these two conditions.
96
φΩ) field for ULSD and B40 at different injection pressures at the SOI of 5ºBTDC, mid-
Figure V-1: Oxygen equivalence ratio (φ
load. The isolines divide relevant equivalence zones. The lift-off length (LOL) for each fuel and injection pressure is
represented with a horizontal line.
97
Figure V-2 illustrates the oxygen equivalence ratio variation of the fuel spray and
lift-off length for diesel and B40 at the SOI of 7ºBTDC. Compared with the case at the
SOI of 5ºBTDC, the lift-off lengths were slightly increased and the equivalence ratio
field was slightly extended due to lower cylinder temperature and gas density at the
injection region, resulting in similar equivalence ratios at the AZ. This suggests that the
difference in residence time at high cylinder temperature, which is not included in the
present model, may be the primary reason for the NOx difference when the SOI is
changed. Nevertheless, given the same residence time, the change of oxygen equivalence
ratio at the AZ is consistent with the change of NOx emissions: both for the increase of
injection pressure and the change of fuel to B40 led the oxygen equivalence ratio at the
AZ to be closer to stoichiometric. One can also observe that diesel at 108 MPa injection
pressure has a closer to stoichiometric equivalence ratio at the autoignition zone than B40
at 72 MPa injection pressure, which is also consistent with the higher NOx emissions for
diesel at 108 MPa than for B40 at 72 MPa injection pressure. Figure V-3 illustrates the
oxygen equivalence ratio variation of the fuel spray and lift-Off length for diesel and B40
at the SOI of 9ºBTDC. The equivalence ratio field was further extended and the lift-off
lengths were further increased. The observations of the relation between equivalence ratio
at AZ and NOx emissions for 9 ºBTDC are consistent with those at the SOI of 5 and 7
ºBTDC.
98
load
99
load
100
Consequently, one can estimate the average oxygen equivalence ratio at the AZ
with Equations (9) and (16) and construct a φΩ ( H ) – BSNOx emission correlation
diagram. Figure V-4 shows the correlation between brake specific NOx emissions and
average oxygen equivalence ratio of the fuel-air mixture at the autoignition zone (location
near lift-off length) at 25% load, regardless of fuel type, i.e., both B40 and ULSD are
presented with the same symbol. The variation for the same fuel and SOI is due to the
change of fuel injection pressure, the effect of which has been thoroughly investigated
[55]. At the same SOI, linear correlations between NOx emissions and average oxygen
equivalence ratio at the autoignition zone can be observed for all of the four SOI cases.
When the mixture is less rich, i.e., the oxygen equivalence ratio of the mixture is closer to
stoichiometric, and the NOx emissions are higher. It can also be seen that with retarded
SOI, the correlated straight line shifts downward, and the slope decreases, which suggests
that the impact of oxygen equivalence ratio at the AZ is more significant on NOx
emissions with earlier SOI. The relation between the slope, Y-axis intercept (at X=0) and
SOI will be discussed later.
101
Figure V-4: The correlation between brake specific NOx emissions and average
oxygen equivalence ratio of the fuel-air mixture at lift-off length at 25% load,
regardless of fuel type. The legend: xB is x degree before TDC; xA is x degree after
TDC.
Figures V-5 and V-6 present the correlation between brake specific NOx
emissions and average oxygen equivalence ratio of the fuel-air mixture at AZ 50% and 75%
load, respectively. At the same SOI, the linear correlations are still valid regardless of
fuel type for moderate and high load. Similar to low load, the slope and Y-axis intercept
(at X=0) is larger with earlier SOI, suggesting the same trend of impact of oxygen
equivalence ratio at AZ on NOx emissions at all load conditions.
102
TDC.
103
TDC.
Overall, this work validates the significance of oxygen equivalence ratio of the
fuel-air mixture at AZ in a light duty diesel engine with single injection extended to a
broader range of SOI and loads. Combined with the work by Mueller et al.[18] in a heavy
duty single-cylinder diesel engine, the present work further confirms that, at the same
SOI, the most important factor determining engine exhaust NOx emissions is the ignition
event that is controlled by the oxygen equivalence ratio of the fuel-air mixture at the
autoignition zone.
To probe the impact of SOI, one can construct the plots of slope – SOI and Y-axis
intercept (at X=0) – SOI from the linear correlations in Figures V-4 to V6. Especially, if
one constructs a semi-ln plot of the slope and SOI, i.e., a ln scale of the absolute value of
104
slope in Y-axis and a linear scale of the SOI in X-axis, as in Figure V-7, extremely
strong linear correlation can be observed for all three load conditions. The extra sample
points in the moderate load case are intended to increase the accuracy of the regression.
Figure V-8 shows the semi-ln plot of the Y-axis intercept (at X=0) and SOI. Extremely
strong linear correlations can be observed for all load conditions as well.
Figure V-7: semi-ln plot of absolute value of slope vs. SOI
105
Figure V-8: semi-ln plot of Y-axis intercept (at X=0) vs. SOI
Therefore, it is possible to establish a general form of an equation to indicate the
NOx emissions with SOI and the oxygen equivalence ratio at AZ for each load. After one
describes the slope and Y-intercept with the exponential correlation of SOI, the brake
specific NOx emissions can be predicted as:
At 25% load
BSNOx = −0.9574e −0.1127 SOI φΩ + 6.7849e −0.0897 SOI (18)
At 50% load
At 75% load
106
where BSNOx is in the unit of g/kWh and φΩ is the average oxygen equivalence ratio of
the fuel-air mixture at AZ. Equations (18) to (20) can be used to predict engine NOx
emissions given known fuel properties and engine injection strategy. Note the
contribution of start of injection (SOI) is interesting since it essentially follows a form of
first order kinetics, which is suggesting that the role of SOI may be a time-relevant factor.
However, this is beyond the scope of this work, therefore it may be included in the future
work. Although the specific parameters in Equations (18) to (20) may be not generalized
to be applied for any diesel engine, the correlation between NOx emissions and average
oxygen equivalence ratio of the fuel-air mixture at autoignition zone should hold
regardless of engine type. In addition, this model should not be applied to engine
operating at advanced combustion mode.
In summary, at the same injection time, strong correlations can be observed
between engine NOx emissions and the average oxygen equivalence ratio at the
autoignition zone near flame lift-off length regardless fuel type. Hence, the hypothesis
that the impact on NOx emission from engine fuel injection strategy control and biodiesel
fueling can be explained by the equivalence ratio at autoignition zone near the flame lift-
off length is validated.
107
Chapter VI. Investigation of late in-cylinder injection
strategy on engine emissions and lubricant fuel dilution
6.1 Exhaust Emissions
Figure VI-1 shows the impact of post injection strategy on NOx, CO and UHC
emissions. NOx emissions were decreased by around 50% when the post injection was
introduced. Compared with the introduction of post injection, changes of NOx emissions
from varying post injection timing, however, were insignificant in this experiment.
Reasons that may contribute to the NOx reduction when post injection was used includes
a retarding of the main start of injection (SOI) and a decrease in the peak heat release
[18]. A more detailed discussion will be provided in a later section.
108
25
UHC
20
15
10
5
Brake emissions (g/kWh)
0
CO
10
0
4 NOx
3
2
1
0
NONE AD(5mg/stk) AD(5mg/stk) AD(8mg/stk) AD(5mg/stk) g/stk) ULSD
45C 60C 60C 70C (5m
45CAD
Figure VI-1: Emissions of B20 engine test with various post injection strategies.
Samples are differentiated with post injection conditions while rest parameters were
kept the same. The fuel for the non-ULSD test was B20.
Both UHC and CO emissions had similar tendency with the variation of post
injection. An increase was observed for both UHC and CO emissions when the post
injection was introduced. When retarding the timing of post injection from 45º to 60º
ATDC, both UHC and CO emissions significantly increased. The highest emissions of
both UHC and CO were observed with the post injection at 60º ATDC and 8mg fuel
injection per stroke. Further retarding the post injection timing from 60º to 70º ATDC
while setting the injection quantity back to 5mg/stk increased UHC emissions but
decreased CO emissions. This suggests that fewer hydrocarbons from the post injection
were oxidized. Compared with diesel fuel at the post injection of 45º ATDC (5mg/stk),
109
B20 decreased UHC and CO emissions but increased NOx emissions, which confirms the
impact of biodiesel fuel on emissions reported in many previous studies [6, 8-11, 13].
6.2 Exhaust Temperature and BSFC
Exhaust temperature and BSFC are sensitive indicators of engine performance.
Figure VI-2 illustrates the exhaust temperatures and BSFC during the 100-hour test.
When there was no post injection during the 75 – 100 hour period, both the exhaust
temperature and BSFC were the lowest. The highest exhaust temperature was confirmed
at the beginning of the test when the post injection was set at 45º ATDC (5mg/stk). The
BSFC also increased when the post injection was set at 45º ATDC (5mg/stk), which was
consistent with the increased exhaust temperature and UHC emissions (Figure VI-1).
110
360
400 60CAD(5mg/stk)
70CAD(5mg/stk) 350
380 340
330
360
Exhaust Temperature (C)
320
BSFC (g/kWh)
340 310
45CAD(5mg/stk) 300
320
290
300
None 280
280 270
260
260 Exhaust Temperature -left Y
60CAD(8mg/stk) 250
BSFC - Right Y
240 240
0 10 20 30 40 50 60 70 80 90 100
Time (h)
Figure VI-2: Exhaust temperature and BSFC profile of the 100-hour test. Each
point is the average of one test period. The dash lines are used to divide zones of
each post injection strategy.
While retarding the post injection from 45º to 60º ATDC without changing the
injection quantity, the exhaust temperature decreased, but the BSFC still increased, which
suggested that more fuel escaped the cylinder. This observation was also consistent with
the trend of UHC emissions (Figure VI-1). Increasing the post injection quantity to
8mg/stk slightly increased the exhaust temperature but significantly increased BSFC,
suggesting that more fuel escaped the cylinder.
Setting the post injection quantity back to 5mg/stk but retarding the injection
timing to 70º ATDC decreased both exhaust temperature and BSFC. Compared with the
60º ATDC (5mg/stk), exhaust temperature was lower but BSFC was significantly higher.
111
These trends were also consistent with UHC emissions. This suggested that retarding post
injection timing decreased the combustion efficiency of the post injected fuel, which had
a significant effect on UHC emissions.
6.3 Heat Release Analysis
Heat release analysis is critical in the understanding of engine behavior. Figure
VI-3 illustrates the pressure traces of tests with different post injection strategies at
around TDC and late crank angle (LCA), i.e., later than 40º ATDC. It is obvious that the
peak pressure for no post injection was the highest, and the peak pressure for 45º ATDC
(5mg/stk) post injection including both B20 and diesel fuel was the lowest. This decrease
in peak pressure is consistent with the fact that a portion of work was done at LCA by the
combustion of post injected fuels, and tests with 45º ATDC (5mg/stk) post injection
generated the most LCA work. This is also confirmed by the pressure traces at LCA
(Figure VI-3), because tests with post injection had higher pressures in the later stage of
the combustion cycle. Thus, to maintain a constant power the engine decreased the work
by fuel from pilot and main injections. The pressure traces of 60º ATDC (8mg/stk) had a
lower peak pressure than 60º ATDC (5mg/stk) but higher LCA pressure, indicating more
energy output from the burning of increased post injected fuel. Late crank angle pressure
of 70º ATDC (5mg/stk) was similar to the test of no post injection, which suggested little
burning of post injected fuel.
112
None
9 45CAD(5mg/stk)
60CAD(5mg/stk)
60CAD(8mg/stk)
70CAD(5mg/stk)
8 45CAD(5mg/stk) ULSD
Crank Angle (deg.)
440 448 456 464 472 480
Crank Angle (deg.)
100
90
Cylinder Pressure (bar)
80
70
60
None
45CAD(5mg/stk)
50 60CAD(5mg/stk)
60CAD(8mg/stk)
70CAD(5mg/stk)
45CAD(5mg/stk) ULSD
40
350 355 360 365 370 375 380 385 390
Crank Angle (deg.)
Figure VI-3: Pressure traces of test with different post injection strategies at both
TDC (top) and late crank angle (bottom)
113
Apparent heat release and cumulative heat release are shown in Figure VI-4. The
combustion of fuel from pilot and main injections for each test are obvious, but for the
post injected fuel, the apparent heat release can only be observed in tests with post
injection at 45º ATDC (5mg/stk). This combustion, which consumed most of the post
injected fuel, explained the increase of exhaust temperature and pressures at LCA. The
slight increase of BSFC and UHC emissions suggested that this combustion at LCA was
not as complete as which occurred around TDC. The lack of significant combustion at
LCA in the other post injection tests also explains their significant increase of UHC
emissions and BSFC.
From cumulative heat release data, apart from the evidence of combustion at LCA
of fuel injected at 45º ATDC with 5mg per stroke, the energy adsorption from liquid
vaporization at around 60º – 80º ATDC in both tests with 60º ATDC post injection can be
clearly observed. The 60º ATDC (8mg/stk) post injection absorbed more energy due to
its higher injected quantity of fuel, which may explain why the increase of UHC
emissions was higher than the increase of quantity of fuel injected (more than twofold
increase in UHC compared with 60% increase in the amount of post injected fuel). Note
that the cumulative heat release increased at later crank angle, which suggested that fuel
still reacted with oxygen and generated thermal energy, though not in the same way as
the combustion at LCA in the test of 45º ATDC (5mg/stk). In contrast, the test of 70º
ATDC(5mg/stk) didn’t show a significant increase in its cumulative heat release before
140º ATDC, which suggests less reaction of post injected fuel at LCA.
114
100
NONE
80 45CAD(5mg/stk)
60CAD(5mg/stk)
60CAD(8mg/stk)
70CAD(5mg/stk)
Apparent Heat Release (J/deg.)
45CAD(5mg/stk) ULSD
60
40
20
-20
340 360 380 400 420 440
Crank Angle (deg.)
1200
1000
800
Cumulative Heat Release (J)
600
400
NONE
45CAD(5mg/stk)
200 60CAD(5mg/stk)
60CAD(8mg/stk)
70CAD(5mg/stk)
-200
364 392 420 448 476
Crank Angle (deg.)
Figure VI-4: Apparent (top) and cumulative heat releases (bottom) of tests with
different post injection strategies
115
Thus, retarding post injection timing significantly decreased the combustion
efficiency at late crank angle. Furthermore, these results suggest that between 45º – 60º
ATDC there is a post injection timing threshold that determines whether combustion at
LCA consuming the posted injected fuel can occur. Figure VI-5 illustrates the computed
bulk cylinder temperature. Apparently the 45º ATDC (5mg/stk) cases had significant
temperature increase at LCA due to the combustion event. Though the tests with 60º
ATDC post injection timing didn’t have similar temperature increase as the test with 45º
ATDC injection, they still had gradual increases in temperature at LCA compared with
no post injection. The temperature increase at 70º ATDC (5mg/stk) was not as significant
as in the other cases. Hence, the bulk cylinder temperatures suggest that the post injected
fuel from different injection timings went through different combustion paths: that the
45º ATDC (5mg/stk) had high temperature combustion, that 60º ATDC (5mg/stk), 60º
ATDC (8mg/stk) and 70º ATDC (5mg/stk) only followed low temperature combustion
mechanisms [218, 228, 229].
116
2200
NONE
45CAD(5mg/stk)
2000 45CAD(5mg/stk)
60CAD(8mg/stk)
70CAD(5mg/stk)
Cylinder Temperature (K)
1600
1400
1200
1000
800
340 360 380 400 420 440 460 480 500
Crank Angle (deg.)
Figure VI-5: Bulk cylinder temperature of tests with different post injection
strategies. The arrow points to the lowest acceptable combustion temperature.
Cylinder temperature is the most critical parameter that determines the path of
combustion at LCA [14]. The fuel injected into the cylinder experiences an atomization –
combustion process, and the time between SOC and SOI is defined as ignition delay.
Here a predictive model (Equation 21) suggested by Assanis et al. [230] was employed to
predict the ignition delay of the post injected fuel in the combustion chamber of diesel
engine.
 Ea 
τ ID = 2.4φ −0.2 p −1.02 exp  ? (21)
 RuT 
117
where τID is ignition delay in millisecond (ms), ϕ is the overall equivalence ratio,
and p and T are the mean pressure and temperature during the ignition delay interval,
respectively. The interval was also corrected with outputs from the iterative ignition
delay calculation. Ea/Ru = 2100 was taken to be, as in [230].
Since the main combustion already finished before the SOI of post injection for
each test (Figure VI-4), the overall equivalence ratio of post injected fuel can be
calculated from the exhaust oxygen concentration, as shown in Equation (22).
O O ɺ
    m post
 F  sto  F  sto
φ= = (22)
O MAF + FC
* CO2 * ρO2
F ρexh
where ϕ is the overall equivalence ratio, (O/F)sto is the stoichiometric oxidizer/fuel
ratio, mɺ post is the rate of the mass of post injected fuel, MAF is mass air flow rate, FC is
fuel consumption, ρexh is exhaust gas density, ρO2 is exhaust oxygen density and CO2 is
exhaust oxygen concentration. The calculation assumed a complete combustion of fuel
from the main injection and neglected the impact of post injected fuel, given its small
quantity, on the total mass of exhaust and the oxygen concentration. Table VI-1 presents
the equivalence ratios, mean pressures, mean temperatures and ignition delays.
Table VI-1: Equivalence ratio ϕ, mean pressure P mean temperature T and ignition
delay τ ID calculated for each post injection case.
Post injection ϕ ഥ
P (bar) ഥ (K)
T τ୍ୈ (ms)
None - - - -
45º ATDC (5mg/stk) 0.134 20.43 1555.10 0.64
60º ATDC (5mg/stk) 0.125 10.50 1328.54 1.61
60º ATDC (8mg/stk) 0.210 10.76 1337.42 1.40
70º ATDC (5mg/stk) 0.123 7.50 1205.18 2.67
118
For post injection at 45º ATDC (5mg/stk), the injection timing was 405º, and the
ignition delay was 0.64 ms, which is 5.7º for an engine operating at 1500 rpm. This
timing was consistent with its SOC at LCA (Figure VI-4). Furthermore, it also
represented the timing with a critical temperature – 1500K. According to Sjӧberg and
Dec [231], 1500K is the lowest acceptable bulk cylinder temperature that high
temperature combustion can happen before the combustion is quenched by piston
expansion. As explained by the Sjӧberg and Dec, OH, which is required for the
dominating CO oxidation reaction (Reaction I, which produces 78% of total CO2 [231]),
becomes too low to initiate the reaction below 1500K.
CO + OH ⇒ CO2 + H (23)
For the post injection at 60º ATDC, the injection timing was 420º. The ignition
delay was 1.61 ms (14.5º) for the injection quantity of 5mg/stk, and 1.40 ms (12.6º) for
the injection quantity of 8mg/stk. These timings were consistent with the start of
temperature rise in Figure VI-5. In addition, the bulk cylinder temperatures at the start of
temperature rise were only around 1300K, which explains the significant change in the
combustion behavior when the post injection was retarded to 60º ATDC. For the post
injection at 70º ATDC (5mg/stk), the injection timing was 430º. The ignition delay was
2.67 ms, or 24.0º. The cylinder temperature when the fuel was ignited was about 1150K.
This further decreased the rate of oxidation of the post injected fuel.
Table VI-2 presents the calculated results of consumed hydrocarbons from the
post injected fuel, assuming the UHC from the pilot-main injection were negligible,
based on the injection quantities and emission measurements. The combustion of 45º
ATDC (5mg/stk) post injected fuel consumed the most hydrocarbons while the
119
combustion of 70º ATDC (5mg/stk) post injected fuel consumed the least, which is
consistent with the measured UHC emissions. The rates of fuel consumption in both tests
with post injection at 60º ATDC were similar due to their similar temperature
environment, which suggested that at the post injection timing of 60º ATDC, the injection
quantity didn’t significantly affect the fuel consumption.
Table VI-2: Calculated UHC emission rate and consumed HC by combustion at late
crank angle. B20 is a fuel mixture blended with 20% v/v SME based B100 and 80%
v/v ultra low sulfur diesel.
Studied cases UHC emission HC Post Consumed HC by

Fuel
(ATDC) (g/m) injected (g/m) Combustion (g/m)
NONE B20 0.11 0 -
45º (5mg/stk) B20 1.03 15 13.97
60º (5mg/stk) B20 6.13 15 8.87
60º (8mg/stk) B20 15.47 24 8.52
70º (5mg/stk) B20 11.88 15 3.12
45º (5mg/stk) ULSD 1.64 15 13.36
6.4 Oil Dilution Level, Total Acid Number (TAN) and mfMEP
Figure VI-6 illustrates the profiles of total acid number, dilution level of lubricant
and mfMEP during the 100-hour test. It shows that no oil dilution can be observed when
the post injection was at 45º ATDC (5mg/stk). Retarding the post injection timing to 60º
ATDC immediately increased oil dilution. However, the subsequent changes of post
injection timings and quantities did not have a significant influence on the rate of dilution.
When the post injection was disabled, the dilution stopped increasing. The slight decrease
observed at the beginning of the test without post injection may be due to the fuel
vaporization from the lubricant. Using the Clausius-Clapeyron equation [232] based on
the suggestion of normal boiling point from [233] and enthalpy of biodiesel from [234],
120
one can estimate the boiling point of biodiesel to be around 550˚C at 20 bar, around
500˚C at 10 bar, and around 470˚C at 7.5 bar. Therefore, the small quantity of post
injected fuel quickly vaporized since their boiling points were much lower than the local
temperature into which they were injected, which may suggest that the primary
mechanism of biodiesel dilution is the dissolution of gas phase hydrocarbons into the
lubricant film on cylinder wall. Furthermore, given the similar cylinder pressure and
cylinder wall temperature in the various tests, the hydrocarbon concentration in the liquid
would be similar regardless of the hydrocarbon concentration in the gases, if it was
simplified as a binary system (light hydrocarbon and heavy oil) and the oil film was
saturated [235]. This explains the observation that the change of hydrocarbon
concentration didn’t affect the quantity of fuel diluted into the oil, as shown by the oil
dilution profile that the change of post injection strategy didn’t have a significant
influence on the rate of dilution.
121
8
TAN - Left Y Dilution - Right Y mfMEP - Right Y
60CAD(5mg/stk) 140
7
4
Total Acid Number (mgKOH/g)
120
6
mfMEP (KPa)
Dilution (%)
100
5
2
45CAD(5mg/stk)
4 80
None
3 0 60
70CAD(5mg/stk)
60CAD(8mg/stk)
2 40
0 10 20 30 40 50 60 70 80 90 100
Time (h)
Figure VI-6: Total acid number (TAN), dilution level and calculated mechanic
friction mean effective pressure (mfMEP) profiles of the 100-hour test.
As shown in Figure VI-6, the total acid number (TAN) showed an initial decrease
and remained stable until around 30 hours. It then increased slightly and stayed at
constant level until around 60 hours. After that, the TAN kept increasing until the end of
the test, which was consistent with the effect of the acid control additives in the lubricant.
At 0 – 30 hours, the mfMEP showed an initial increase then a decrease, which was
attributed to a break-in process [236]. After 30 hours, there was no distinct trend. This
may be due to the increase of oil sludge, deposit and particulate matter content during
engine operation. Besides that, a weak correlation between the mfMEP and oil dilution
122
level can be observed, which may represent an increase of friction due to the decrease of
oil viscosity [236] from dilution.
6.5 Analysis of Hydrocarbons from Lubricant and Exhaust
Figure VI-7 illustrates the GC-MS results of the diesel base fuel, B100, and the
lighter cut from a 432˚C distillation of fresh and used lubricant. The region of interest in
the mass spectrum includes hydrocarbons that do not originate from the lubricant. The
results of used lubricant show evidently a portion of the paraffins from diesel and the
fatty acid methyl esters (FAME) components from biodiesel. This confirms that both
diesel and biodiesel were diluted into the lubricant. Other hydrocarbons in the used
lubricant have not been detected from the GC-MS analysis, which suggests that fuel
rather than combustion by-products was the primary source of lubricant dilution.
123
Diesel
B100
100h used lubricant
New lub
0 10 20 30 40 50 60 70 80
Ret.Time (min)
14
15
16
Diesel
17
18
19
20
21 22 23
C18:1
C18:2
B100 PN02738366
C16:0
C18:0
C20:1 C20:0
100h used lubricant
30 35 40 45 50 55
Ret.Time (min)
Figure VI-7: GC-MS results of diesel, B100, lighter cut from 432˚C distillation of
fresh and used lubricant (top); Region of interest in the GC-MS results (bottom).
The numbers at the diesel zone means the number of carbon in relevant paraffin.
124
Figure VI-8 shows the GC-MS analysis of hydrocarbons collected from the
exhaust for the tests with different post injection strategies. Consistent with the UHC
emission results, the exhaust hydrocarbons in the case without post injection and the case
of 45º ATDC (5mg/stk) were hardly detectable, but significant in the other three cases.
The majority of hydrocarbon components were paraffins that evaporated from the diesel
fuel, FAMEs from biodiesel and smaller molecular by-products from combustion. These
by-products such as benzene and toluene, however, have too low boiling points to be
diluted into the lubricant. Further smaller molecules were not detected by the approach
used. This also confirms the absence of high temperature combustion for the cases of 60º
ATDC (5mg/stk), 60º ATDC (8mg/stk) and 70º ATDC (5mg/stk).
125
(19)
(1) (11) (12) (17)
(20)
(2) (13)
(4) (8) (9) (10)
(7) (14) (15)
(3) 70CAD(5mg/stk)
(5) (6) (16) (18) (21)
60CAD(8mg/stk)
60CAD(5mg/stk)
45CAD(5mg/stk))
none latecycle
0 20 40 60
Ret.Time (min)
(1): C6H14 (5): C7H8 (9): C11H24 (13): C15H32 (17): C17H34O2 (21): C19H38O2
(2): C5H8O (6): C8H18 (10): C12H26 (14): C16H34 (18): C19H40
(3): C6H6 (7): C9H20 (11): C13H28 (15): C17H36 (19): C19H34O2
(4): C7H16 (8): C10H22 (12): C14H30 (16): C18H38 (20): C19H36O2
Figure VI-8: GC-MS results of hydrocarbons collected from exhaust gases for
various post injection strategies
Additionally, only paraffins with carbon numbers larger than 14 were able to be
diluted into the lubricant. This suggests an average operating temperature of the oil film
at around 400˚C according to the boiling point of C14H30 at 10 bar [233]. Smaller
molecules would vaporize under this temperature rather dilute into the lubricant.
Compared with B100 fuel, as shown in Figure VI-9, the FAME composition of
the biodiesel diluted in the used lubricant had a significant decrease in the most
unsaturated hydrocarbon methyl linoleate but various increases in the other FAME
126
components, the highest of which was methyl stearate. This can be explained by the
higher oxidation instability of unsaturated FAME compounds and the lower boiling point
of methyl palmitate.
Original Biodiesel
60
in the 100h used lubricant 53.6
50
Comparative Mass Fraction (%)
44.2
40
30 26.7 25.8
21.0
20 17.8
10 8.2
2.8
0
C16:0 C18:0 C18:1 C18:2
Biodiesel constituent
Figure VI-9: FAME composition of B100 and of the biodiesel in used lubricant.
Overall, post injection strategy can be an effective approach to reduce NOx
emission. However, it will increase the fuel consumption as well as CO and unburned
hydrocarbon emissions. The combustion behavior and oil dilution significantly depend on
the post injection strategy, including the start of injection and quantity of fuel post
injected. With biodiesel fueling, no significant oil dilution can be observed if the post
injection timing is earlier than 45º after top dead center. The hypothesis that biodiesel
fueling can introduce lubricating oil dilution with post injection strategy is a conditional
127
valid statement: if the post injection timing is earlier than 45º after top dead center, no
lubricating oil dilution can be observed; post injection timing later than 45º after top dead
center will introduce lubricating oil dilution.
128
Chapter VII. Impact of fuel injection pressure on soot
morphology and oxidative reactivity
7.1 Morphology analysis
Figures VII-1 and VII-2 show two examples of TEM images for soot
agglomerates from diesel and B20 combustion, respectively. Apparently these
agglomerates are composed of spherical primary particles, and the appearances of
agglomerates from diesel combustion and biodiesel combustion are very similar to each
other based on subjective observation.
129
Figure VII-1: A transmission electronic microscopy (TEM) image of diesel soot at 30%
load and 50 MPa fuel injection pressure
130
Figure VII-2: A transmission electronic microscopy (TEM) image of B20 soot at 30%
load and 50 MPa fuel injection pressure
At the same time, with the fractal analysis approach, a quantitative comparison
can be conducted. Figure VII-3 presents the comparison of primary particle size between
diesel and B20 soot. The result suggests that neither engine load nor fuel injection
pressure can significantly affect the primary particle size. This insignificant impact of
engine load on primary particle size is inconsistent with previous observations by
Lapuerta et al. [177], who claimed that the higher fuel/air ratio increases the size of
primary particle due to weakened mixing. However, they also suggested that higher in-
cylinder temperature can decrease the primary particle size. Hence, a combined effect of
131
higher fuel/air raito and higher in-cylinder temperature may be the reason for the
insignificance in the change of primary particle size with engine load. The insignificant
impact of fuel injection pressure is also inconsistent with previous observation by Mathis
[176], who claimed a reduction in primary particle size with increased fuel injection
pressure. Nevertheless, it is acknowledged that the number of studies on fractal analysis
of diesel engine soot is limited, and it is also possible that the impact of engine conditions
on soot primary particle size is engine-dependent.
Figure VII – 3 also compares the primary particle sizes of agglomerates between
diesel and B20 soot. For B20 soot particles, no significant impact of fuel injection
pressure or engine load on primary particle size can be observed either. However,
inconsistent with previous observation by Smekens et al. [178], who claimed decreased
primary particle size with biodiesel fueling, the primary particle sizes of B20 soot appear
to be higher than diesel soot. Again, it is possible that the impact of biodiesel fueling on
soot primary particle size is engine-dependent, and further investigation is needed to
thoroughly understand the impact of engine conditions and biodiesel fueling on soot
primary particle size.
132
60
55
30% load, diesel
60% load, diesel
Diamter of primary particles (nm) 50 30% load, B20
45 60% load, B20
40
35
30
25
20
15
10
0
50 75 100 125
fuel injection pressure (MPa)
Figure VII-3: Primary particle sizes for both diesel and B20 soot
Figure VII-4 compares the fractal dimension between diesel and B20 soot at
different fuel injection pressures and engine load. Consistently for both diesel and B20
soot, the increase of fuel injection pressure slightly decreases the fractal dimension. In
addition, increase of engine load significantly increases the fractal dimension, which is
consistent with previous researches [170]. It is also consistent with previous research by
Soewono and Rogak [175] that biodiesel fueling does not have a significant impact on
soot fractal dimension.
133
2.8
30% load, diesel
2.6 60% load, diesel
30% load, B20
2.4
60% load, B20
Fractal dimension
2.2
2.0
1.8
1.6
1.4
1.2
1.0
50 75 100 125
Figure VII-4: Fractal dimensions for both diesel and B20 soot
Since the fractal dimension characterizes the structure of soot agglomerates,
which are formed through the collision of soot primary particles and smaller
agglomerates in the exhaust flow [237], a possible reason for the higher fractal dimension
at higher load and lower injection pressure could be the higher temperature in the exhaust
flow where the restructuring and collision of soot agglomerates happens. Mikhailov et al.
[238] investigated the impact of temperature on the restructuring of oil soot, and found
that at higher temperature the soot will restructure to less anisotropic structure, i.e.,
higher fractal dimension. Since biodiesel won’t significantly affect engine in-cylinder
temperature [44], no significant impact on the fractal dimension with biodiesel fueling
can be observed.
134
Figure VII-5 presents the comparison of number of primary particles from fractal
analysis. It can be observed that the number of primary particles increases with higher
engine load, but decreases with higher fuel injection pressure. This is consistent with
previous investigations [170]. In addition, no significant impact of biodiesel fueling can
be observed. Note although this analysis intends to investigate the change of number of
primary particles of exhaust gas, it was conducted based on localized TEM image. A
more accurate investigation of particle number should be conducted with SMPS, which
will be provided.
600 30% load, diesel

60% load, diesel
Number of primary particles (#)
500 30% load, B20

60% load, B20
400
300
200
100
50 75 100 125
Figure VII-5: Number of primary particles for both diesel and B20 soot
Figure VII – 6 presents the particle size distributions of diesel and B20 exhaust at
30% load, before the thermodenuder. These particle size distributions show as symmetric
135
peaks in a semi-log plot (linear y-axis scale and log x-axis scale). The peak at smaller
particle diameters is dominantly hydrocarbon droplets, i.e., the nucleation mode peak,
and the peak at larger particle diameters is predominantly exhaust particulate matter, i.e.,
the accumulation mode peak [63]. It can be observed that an increase of fuel injection
pressure can decrease the accumulation model peak while increasing the nucleation
model peak. This can be due to the better mixing from increased fuel injection pressure.
In addition, only slight decrease in the particle concentrations can be observed with
biodiesel fueling. These results are consistent with the trends of number of primary
particle sizes presented earlier.
8
1.0x10
Diesel 100 MPa
Diesel 75 MPa
Diesel 50 MPa
Particle concentration (#/cm3)
7
8.0x10 B20 100 MPa
B20 75 MPa
B20 50 MPa
7
6.0x10
7
4.0x10
7
2.0x10
0.0
10 100
Particle diamter (nm)
Figure VII-6: Particle size distribution of diesel and B20 combustion exhaust, 30%
load, before thermodenuder
136
Figure VII-7 presents the particle size distributions of diesel and B20 exhaust at
30% load, after the thermodenuder. The function of the thermodenuder is to remove the
hydrocarbon content in the exhaust gases. As shown in Figure VII -7, although not
completely, the nucleation mode peaks (dominantly hydrocarbon droplets) have been
significantly reduced and the accumulation mode peaks are clearer.
7
1.2x10
Diesel 100 MPa
Diesel 75 MPa
1.0x10
7 Diesel 50 MPa
B20 100 MPa

B20 75 MPa
6
B20 50 MPa
8.0x10
6
6.0x10
6
4.0x10
6
2.0x10
0.0
10 100
load, after thermodenuder
Figure VII-8 presents the particle size distributions of diesel and B20 exhaust at
60% load, before the thermodenuder. It can be observed that the nucleation mode peaks
dominated the particle size distribution. Consistent with lower load, an increase in the
nucleation mode peak can be observed with higher fuel injection pressure. Figure VII-9
shows the particle size distributions of diesel and B20 exhaust at 60% load, after
137
thermodenuder. It clearly shows that an increase of fuel injection pressure can decrease
the accumulation model peak. One can also observe that the particle number
concentration at higher engine load is lower than that at lower engine load. Intuitively
this seems to be contrary with the trend of number of primary particle presented earlier,
but it should be acknowledged that even the number concentration of particles in the
exhaust is lower for higher engine load. An increased flow rate at higher engine load can
promote an increase in the total particle number.
7
8.0x10
Diesel 125 MPa
Diesel 100 MPa
Diesel 75 MPa
B20 125 MPa

7
6.0x10 B20 100 MPa
B20 75 MPa
7
4.0x10
7
2.0x10
0.0
10 100
load, before thermodenuder
138
6
3.0x10
Diesel 125 MPa
Diesel 100 MPa
6 Diesel 75 MPa
2.5x10
B20 125 MPa

B20 100 MPa
6
B20 75 MPa
2.0x10
6
1.5x10
6
1.0x10
5
5.0x10
0.0
10 100
load, after thermodenuder
7.2 Analysis of soot oxidative reactivity
Figure VII – 10 presents the thermogravimetric analysis result of 30% load diesel
and B20 soot. It appears that the fuel injection pressure has a significant impact on the
soot oxidative reactivity for both diesel and B20 soot. The increase of fuel injection
pressure from 50 MPa to 75 MPa significantly increases the oxidative reactivity, and
further increasing the fuel injection pressure to 100 MPa increases the oxidative reactivity,
though not as significantly as the change from low to moderate fuel injection pressure.
This trend is consistent with the PM emissions investigated earlier, where the most
significant change happens when the fuel injection pressure is increased from low to
139
moderate. Figure VII-10 also confirms one of the hypotheses in this work, i.e., soot from
higher fuel injection pressure has higher oxidative reactivity.
In addition, consistent with previous investigations [75, 116, 117, 181], the
oxidative reactivity of B20 soot is higher than diesel soot, given the same fuel injection
pressure.
diesel 50 MPa
1.0
diesel 75 MPa
diesel 100 MPa
B20 50 MPa
0.8 B20 75 MPa
B20 100 MPa
0.6
m/m0
0.4
0.2
0.0
0 10 20 30 40 50 60
time (m)
Figure VII-10: Thermogravimetric analysis result of 30% load diesel and B20 soot,
m/m0 indicates the residual mass weight percentage.
Figure VII – 11 presents the thermogravimetric analysis result of 60% load diesel
and B20 soot. The general observations for the behavior of oxidative reactivity for 60%
load soot are quite similar to the 30% load: increase the fuel injection pressure can
increase soot oxidative reactivity, and at the same fuel injection pressure B20 soot has a
140
higher oxidative reactivity than diesel soot. Figure VII-11 further confirms the
hypotheses proposed in this work.
diesel 75 MPa
1.0
diesel 100 MPa
diesel 125 MPa
B20 75 MPa
0.8 B20 100 MPa
B20 125 MPa
0.6
m/m0
0.4
0.2
0.0
0 10 20 30 40 50 60
time (m)
Figure VII-11: Thermogravimetric analysis result of 60% load diesel and B20 soot,
m/m0 indicates the residual mass weight percentage.
Since the soot oxidation is a reaction between carbon and oxygen, a simple first
order kinetic model can be employed to describe the mass consumption rate [239-242]:
dm
= − kcCasCO2 ⋅ dm (24)
dt
where m is the mass of soot, kc is rate constant of carbon-oxygen reaction, CO2 is the
concentration of oxygen and Cas is the concentration of active surface sites. The active
surface site parameter is included in the model since it (characterized as active surface
area [243]) has been observed to affect the oxidation rate [184, 240, 244] due to its
141
significant role in oxygen diffusion rate [245, 246]. Given in TGA a stable flow of air
was fed into the environment, therefore the concentration of oxygen can be considered as
a constant. This reorganization reduces Equation (24) into a first order chemical kinetic
model, if one defines an apparent reactivity as:
k a = kcCas CO2 (25)
Then Equation (24) becomes:
dm
= − k a ⋅ dm (26)
dt
Re-arrange Equation (26) as:
d ln( m) d ln( m / m0 )
= = −ka (27)
dt dt
where m0 is the initial mass weight of the soot sample. Equation (27) indicates that in a
semi-log plot (linear scale in x-axis and log scale in y-axis) of time vs. m/m0, the slope of
the curve represent the apparent reactivity, in other words, the absolute value of the first
order derivative of a time vs. ln(m/m0) plot will be the reactivity. Based on this
consideration, the apparent oxidative reactivity was calculated, as shown in Figure VII -
12 (for 30% load soot) and Figure VII-13 (for 60% load soot).
142
0.08
diesel 50 MPa
diesel 75 MPa
0.07
diesel 100 MPa
B20 50 MPa
0.06 B20 75 MPa
reactivity (1/min)
B20 100 MPa
0.05
0.04
0.03
0.02
0.01
0.00
0 10 20 30 40 50 60
time (m)
Figure VII-12: Apparent oxidative reactivity of 30% load diesel and B20 soot
143
0.12
diesel 75 MPa
0.11 diesel 100 MPa
diesel 125 MPa
0.10
B20 75 MPa
0.09 B20 100 MPa
B20 125 MPa
0.08
reactivity (1/min)
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
0 10 20 30 40 50 60
time (m)
Figure VII-13: Apparent oxidative reactivity of 60% load diesel and B20 soot
Immediately one can observe that the apparent reactivity changes with burn-off,
which is consistent with various previous studies [242-244, 247]. Meanwhile, the change
of apparent reactivity varies for soot produced from different fuel injection pressures. The
diesel soot with the lowest fuel injection pressure appears to give an apparent reactivity
which is insignificantly affected by the burn-off. With increased fuel injection pressure,
the apparent reactivity tends to increase with burn-off. This increase in reactivity may be
due to the increase in the active surface area with burn-off, which have been observed
earlier [244, 247]. The consumption of carbon atoms during oxidation, which could
expand the pore size and consequently increase the area that can be bond with oxygen,
can contribute to the increase of active surface area [243]. It can also explain why the
apparent reactivity of soot from higher fuel injection pressure (or biodiesel fueling) has a
144
faster increase than the soot from lower fuel injection pressure (or diesel fueling) since
they possess higher oxidative reactivity, i.e., carbon consumption rate.
A very interesting observation is that for the highest fuel injection pressure, the
reactivity of soot shows an initial increase with burn-off while it decreases when the
burn-off is close to completion. This kind of reduction in reactivity at very high burn-off
has also been reported by Song et al. [116]. They concluded that the oxidation of high
reactivity soot particle follows an internal burning process, forming central hollow
structures. The residual carbon shell at high burn-off then tends to undergo a coalescence
process due to physical factors to minimize strain energy. This physical coalescence
process may behave as a decrease in active surface area, leading to a decrease in apparent
reactivity. In fact, this decrease in active surface area at high burn-off for certain soots
has also been reported by Cheng and Harriott [244].
However, in both work by Song et al. [116] and Cheng and Harriott [244], the
reduction of reactivity is less significant than the increase of reactivity at earlier burn-off.
In the work presented here, the reactivity at high burn-off was decreased to be even lower
than the initial reactivity or the reactivity of soot from lowest fuel injection pressure. To
investigate this observation, Raman results were presented. Figure VII-14 shows the
AD1
ratio with fuel injection pressure for both diesel and B20 soot, and Figure VII-15
Ag
shows the D1 FWHM values with fuel injection pressure for both diesel and B20 soot.
AD1
Although Seong [190] claimed that the ratio of and D1 FWHM can be correlated
Ag
with soot oxidative reactivity, this correlation has not been observed in the present work.
145
On the contrary, no unanimous trends from Raman analysis of the initial soot
nanostructure can be observed corresponding to the oxidative reactivity.
2.6
diesel,30%
2.5 B20, 30%
diesel, 60%
B20, 60%
2.4
2.3
AD1/Ag
2.2
2.1
2.0
1.9
50 75 100 125
Figure VII-14: AD1/Ag ratio with fuel injection pressure for both diesel and B20 soot
146
190
diesel,30%
B20, 30%
180
diesel, 60%
B20, 60%
D1 FWHM (cm-1)
170
160
150
140
130
50 75 100 125
Figure VII-15: D1 FWHM values with fuel injection pressure for both diesel and
B20 soot
The similar results from Raman analysis for soot at all fuel injection pressures
may be because the Raman spectrum essentially shows an averaged result of the whole
soot carbon structure. If the soot from high fuel injection pressure is composed of both a
highly disordered and a highly ordered portion, while soot from low fuel injection
pressure is composed more uniform and moderately ordered particles. This will make the
initial reactivity of soot from high fuel injection pressure higher, but the reactivity after
burn-off lower than soot from low fuel injection pressure. These kinds of structures may
not show significant differences in Raman analysis. The reason for this kind of structure
may be: at high fuel injection pressure there is longer soot inception time, leading to more
disordered initial structure. Meanwhile, the high in-cylinder temperature from high fuel
injection pressure promotes the self-organization process [194]. Although the time for
147
self-organization is shorter, the organized portion of carbon structure may approach to
more ordered state than that at lower temperature environment. For low fuel injection
pressure there is short soot inception time, leading to longer self-organization time, i.e.,
self-organization process dominates the final carbon structure. This can lead to a more
uniformed degree of disorder.
However, given the scope of this work, the investigation of detailed soot carbon
structure is not included. To fully verify the aforementioned hypothesis, active surface
area and soot carbon crystalline structure analysis with burn-off should be conducted.
They will be included in the recommendation for the future work.
The hypothesis that soot from high fuel injection pressure has higher oxidative
reactivity since the decreased soot precursor concentration can prolong the soot inception
time is validated.
148
Chapter VIII. Optimized heat release pattern to achieve
high engine efficiency
8.1 Model formulation
A zero-dimensional engine thermodynamic simulation was conducted in this
work. The assumptions used in the simulation included:
1. Fuel combustion efficiency is 100%, i.e., all energy of the input fuel is
converted to heat completely, and the combustion process is not included in
this model.
2. There is no heat loss due to fuel vaporization, and physical processes
including gas swirl and fuel spray formation do not contribute to any heat
exchange.
3. No radiation heat transfer in the system
4. No gas leakage in the system.
5. The gas in the combustion chamber is an ideal gas.
6. No heat exchange during the gas exchange cycle
The above assumptions show that the system created here is close to an idealized
system. Nevertheless, the interest in this work focuses on the design of an optimized heat
release pattern, for which these assumptions are sufficient. Sub-models for each
assumption can be implemented easily if more thorough investigations is needed.
The crank angle of one complete cycle is from 0º to 720º, i.e., the top dead center
is 360º. First, in one complete cycle the total indicated work is:
149
EVO
Wi = ∫ P (θ ) dV (θ ) (28)
IVC
where: P (θ ) : cylinder pressure
π DB2
V (θ ) : chamber (gas) volume, V = Vc + [ Lc + r − ( r cos θ + L2c − r 2 sin 2 θ )] ,
4
Vc is the minimum volume and Lc is the length of connection rod
θ : crank angle in radian. Conversion between radian and degree and be easily
180°
done as: degree = × radian
π
IVC and EVO: intake value close and exhaust valve open
For an engine system, the rate of total heat addition is:
dqt dW dE dqw dM e
= + + +h (29)
dθ dθ dθ dθ dθ
where: qt: total heat added
W: work done by the gas; dW = PdV
E: internal energy of gas. dE = Cv ( MdT + TdM )
M: mass of gas in the chamber. M = n ⋅ MWg , n is the gas mole in the chamber,
MWg is the gas molar mass. Since it is air, we will use 28.97 g/mol
qw: heat transfer to the wall.
h: enthalpy per unit of mass. h = C pT
Me: mass of gas in leakage
Cv and Cp are heat capacity for constant volume and constant pressure,
respectively.
One can define:
150
Cp
γ= (30)
Cv
Also for ideal gas
PV = MRuT (31)
C p = Ru + Cv (32)
where T is the temperature of gas, Ru is universal gas constant. It is also easy to see:
dM dM e
=− (33)
dθ dθ
Put Equations (30) – (32) in Equation (29), rewrite:
dqt dV dT dM dqw dM
=P + Cv M + CvT + − C pT
dθ dφ dθ dθ dθ dθ
dV dT dM dqw
=P + Cv M − RuT +
dθ dθ dθ dθ
dV d ( PV / Ru ) PV dM dqw
=P + Cv − + (34)
dθ dθ M dθ dθ
dV C dV Cv dP PV dM dqw
=P + vP + V − +
dθ R dθ R dθ M dθ dθ
γ dV 1 P PV dM dqw
= P + V − +
γ − 1 dθ γ − 1 dθ M dθ dθ
From assumption, dM = 0 . However, one can easily implement a leakage model
inside the simulation. For instance, Hoppie et al. [32] suggested a rate of leakage as a
function of cylinder pressure, temperature and size of leaking hole.
Then the rate of total heat addition becomes:
dqt γ dV 1 dP dqw
= P + V + (35)
dθ γ − 1 dθ γ − 1 dθ dθ
From Heywood [1], the apparent heat release (AHR) during firing is:
γ dV 1 dP
AHR = [ P + V ] firing (36)
γ − 1 dθ γ − 1 dθ
151
The value of γ can be estimated with a polynomial function of temperature [1]:
γ = 1 / (1 − 1 / ( a1 + a2 * T + a3 * T 2 + a4 * T 3 + a5 * T 4 )) (37)
a1 = 3.6359
where
a2 = −1.33736 * 0.001
a3 = 3.29421* 0.000001
a4 = −1.91142 * 0.000000001
a5 = 0.275462 * 0.000000000001
The total energy input is:
4π
dqt dqt
qt = ∫ ( dθ )dθ = ∫
0
firing
(
dθ
) dθ (38)
The overall heat transfer of the engine includes both convection and radiation.
However, since we are assuming complete combustion, the radiation term can be
neglected [1]. Hence, the heat transfer within the engine can be calculated as [248]:
dqw
= qɺw = hw [ Apis (T − Tpis ) + Acyw (T − Tcyw ) + Acyh (T − Tcyh )] (39)
dt
where: hw : convective heat transfer coefficient
Apis , Tpis : area and temperature of piston top
Acyw , Tcyw : area and temperature of cylinder wall
Acyh , Tcyh : area and temperature of cylinder head
The convective heat transfer coefficient hw can be predicted with the correlation
proposed by Woschni [249]:
hw = 3.26 ⋅ DB1− m ⋅ P m ⋅ T 0.75 −1.62 m wm (40)
152
where: DB : engine bore size
w : average cylinder gas velocity
m: an exponent used for correlation. Heywood [1] suggested a value of 0.8 in the
correlation. A later investigation by Han et al. [250] suggested that m possesses a
value between 0.7 to 0.8, which varies with engine.
For a four stroke direction injection compression ignition engine the average
cylinder gas velocity can be calculated as:
Vd Tr
w = Cw1S p + Cw 2 ( P − Pm ) (41)
PV
r r
where Vd : displacement volume
Tr , Pr and Vr : temperature, pressure and chamber volume at reference point, e.g.,
intake value closing
S p : mean piston speed, S p = 2 LN , L is piston stroke, N is engine speed
For gas exchange period: Cw1 = 6.18, C w2 = 0
For the compression period: Cw1 = 6.18, Cw2 = 0
For the combustion and expansion period: Cw1 = 6.18, Cw2 = 3.24*10-3
Finally, the temperatures of engine component can be estimated as [251]:
Tx = Tref + ax ( ρair S p )0.8 + bx (1000 ⋅ Y ) (42)
where Tref : reference temperature, here coolant temperature is used
ρ air : air density in the intake manifold
mf
Y : fraction of fuel in the gas mixture, Y =
ma + m f
153
When x = piston: ax = 6.9 , bx = 2.66
= Cylinder wall: ax = 1.8 , bx = 0.7
= Cylinder head: a x = 2.1 , bx = 0.91
Up to this point, it will be able to calculate the indicated thermal efficiency as:
Wi
ηi = (43)
qt
The brake specific thermal efficiency requires calculation of friction work. The
general friction work can be calculated as:
4π
W friction = ∫T f (θ , P ) dθ (44)
0
The friction torque of a diesel engine includes piston assembly friction and
crankcase assembly friction. Rakopoulos et al. [252] developed a comprehensive model
for friction torque in a direct injection diesel engine, which includes six components, and
the total friction torque is the summation of fraction torque of all components:
Tf = Tf 1 + Tf 2 + Tf 3 + Tf 4 + Tf 5 + Tf 6 (45)
c
T f 1 = c1 µ ⋅ S p ⋅ wo ⋅ ( P + Pe ) ⋅ ( no + 0.4nc ) ⋅ DB ⋅ r ⋅ , friction torque for ring
rω
hydrodynamic lubrication
c
T f 2 = c2 ⋅ π ⋅ DB ⋅ rc ⋅ wc ⋅ ( P + Pe ) ⋅ (1 − sin θ ) ⋅ r ⋅ , friction torque for ring mixed
rω
lubrication
c c
T f 3 = c3 ⋅ µ ⋅ ⋅ DB ⋅ l p ⋅ r ⋅ , friction torque for piston skirt
h rω
154
nv ⋅ N s ⋅ r c
T f 4 = c4 ⋅ ⋅ , friction for valve train
ω rω
T f 5 = c5 ⋅ µ ⋅ ω , friction for auxiliaries and unloaded bearing
π DB2 rjb ⋅ P ⋅ cosθ

T f 6 = c6 ⋅ , friction for loaded bearing
4 ω
where µ is lubricant dynamic viscosity, wo is width of oil ring, wc is width of compression
ring, r is crank case radius, Pe is elastic ring pressure force, no is number of oil rings, nc is
number of compression rings, ω is radial speed, rjb is radius of journal bearing, lp is the
length of piston skirt, h is the oil film thickness, Ns is the preloading force of the valve
spring, and c is the instantaneous piston velocity:
r
) cosθ(
Lc
c = rω{sin θ ⋅ [1 + ]}
r 2 2 0.5
(1 − ( ) sin θ )
Lc
From [252] the multiplier at each friction torque component, assuming a middle
speed of 2500 rpm for the engine worked here, are: c1 = 21.69 , c2 = 0.20 , c3 = 0.45 ,
c4 = 0.27 , c5 = 0.65 , c6 = 0.19 . Table VIII-1 summarizes the basic parameters used by
the friction model
Table VIII-1: Engine basic parameters used by friction model.
Lc 17.6 cm
wo* 3 mm
wc* 2 mm
Pe* 15050 N/m2
no 1
nc 2
lp 52.5 mm
h* 1.5 mm
µ 0.0855 kg/m.s
Ns* 200 N
155
rjb* 15 mm
*: data based on Ref. [252], but scaled to fit a lighter duty diesel engine.
Note the values from Table VIII-1 may not be accurate for the real engine used in
the earlier tasks of this work. Nevertheless, it can still provide qualitative understanding
of the engine behavior. If we assumed a negligible pumping loss (e.g., a high intake
pressure from turbocharger), then the brake specific thermal efficiency will be:
Wi − W friction
ηi = (46)
qt
8.2 Computational procedure
8.2.1 Motor cycle calculation
The motor cycle can be calculated based on Equation (35), with defining dqt = 0 ,
or:
γ dV 1 dp dqw
P+ V + =0 (47)
γ − 1 dθ γ − 1 dθ dθ
Calculation of Equation (47) follows a numerical procedure. With a known
pressure P0 at the crank angle θ0 and chamber volume V0 , the pressure P after a change
of crank angle of ∆θ has the relation:
γ 1
P ∆V + V ( P − P0 ) + ∆qw = 0 (48)
γ −1 γ −1
The term ∆qw represent the heat transfer during the period when crank angle
changes ∆θ , so
∆θ
∆qw = qɺw ∆t = qɺw (49)
ω
156
Note as mentioned earlier, ω is the angular velocity. Combining Equation (48)
with Equation (49) and re-arranging, the pressure after ∆θ is:
∆θ
PV − qɺw (γ − 1)
P=
0
ω (50)
γ (V − V0 ) + V
qɺ w can be calculated from Equation (39). In addition, one must note that γ is a
function of temperature, which will be affected by the pressure value. Therefore, one
needs to perform an iteration procedure to re-calculate γ value with the new P value,
until P became stable ( δ P < 10−7 between each iteration).
The first point of this numerical procedure ( P0,i , V0,i ) should choose the crank
angle location at which the intake value closes. The subscript i means the initial value for
the procedure. The real engine used in Tasks 1 – 4 has IVC at 200º and EVO at 520º,
which will be used in this model. The initial pressure P0,i will use the intake air pressure
measured from real engine data.
Finally, the obtained motor pressure trace needs to be synchronized with volume
to eliminate the thermodynamic lag [253], i.e., the maximum motor pressure peak timing
should be synchronized with piston top dead center. The pressure trace data obtained
from real engine in this work were synchronized by timing and clocking signal which
were fed into the data acquisition system.
To compare the model with real engine data, the pressure trace of an engine
running with diesel at 1500 rpm and 30% load, single injection timing at 1º before top
dead center were used. The pressure trace data was obtained from Task 4. At this
injection timing the start of combustion is after top dead center so that the pressure trace
157
before top dead center can be used as the motor pressure. The value of P0,i used the intake
air pressure of this condition.
The comparison of simulation and real engine pressure trace also requires
adjustment of the value m in Equation (40). The initial value of m used 0.8, and with
adjustment it was found that a best match can be obtained when m = 0.761.
Figure VIII-1 presents a comparison of simulated motor pressure trace and real
engine data with combustion existed. A very good match can be observed between the
simulation and engine pressure trace. This serves as a validation of the employed model.
100
motor pressure (simulation)
90 Engine pressure trace
80
70
60
Pressure (bar)
50
40
30
20
10
0
0 100 200 300 400 500 600 700
CA (degree)
Figure VIII-1: comparison of simulated motor pressure trace with engine data,
running at 1500 rpm, 30% load, single injection timing at 1º before top dead center,
100 MPa fuel injection pressure
158
8.2.2 Apparent heat release representation
A typical apparent heat release profile of diesel combustion includes a premixed
flame peak and a diffusion flame peak. Figure VIII-2 shows that the apparent heat
release profile can be represented with a convolution of one Gaussian peak and one
ExpConvExp peak. The engine AHR data was chosen from a condition of 1500 rpm, 30%
load and 100 MPa fuel injection pressure, which is one case in Task 4. Note there’s no
negative value before the start of combustion given it was assumed no heat exchange due
to fuel vaporization. The convolution was conducted with Igor pro.

Difference (J/degree)
8
4
0
-4
150 Engine AHR

AHR (J/degree)
Fit curve
100
50
Artificial peaks (J/degree)
0
0
120
80
40
1
0
350 360 370 380 390 400
CA (degree)
Figure VIII-2: Representation of apparent heat release (AHR) at 1500 rpm, 30%
load, 1º before top dead center injection, 100 MPa fuel injection pressure with a
convolution of two peaks: one Gaussian peak and one ExpConvExp peak. Top:
159
difference between the engine data and fit curve; middle: comparison of engine data
and fit curve; bottom: deconvolution of artificial peaks.
Figure VIII-3 presents another example of the peak convolution. The engine
apparent heat release profile was obtained from the operating conditions: 1500 rpm, 60%
load, single injection timing at 1º before top dead center, 100 MPa fuel injection pressure.
Although the match is not as good as lower load, the convoluted peak can still represent
the major characteristic of the real heat release. Nevertheless, in the work presented here
we will only investigate the 30% load case.

Difference (J/degree)
10
0
-10
140
120 Engine AHR
AHR (J/degree)
100 Fit curve

80
60
40
20
Artificial peaks (J/degree)
0
0
120
80
40 1
0
350 360 370 380 390 400
CA (degree)
Figure VIII-3: Representation of apparent heat release (AHR) at 1500 rpm, 60%
load, 1º before top dead center injection, 100 MPa fuel injection pressure with a
convolution of two peaks: one Gaussian peak and one ExpConvExp peak. Top:
difference between the engine data and fit curve; middle: comparison of engine data
and fit curve; bottom: deconvolution of artificial peaks.
160
Figures VIII-2 and VIII-3 show that the apparent heat release can be artificially
generated by a convolution of two peaks. Through adjusting the two peaks’ locations,
heights and widths, one is able to test the effect of different apparent heat release patterns.
In addition, the close representation of the artificial peak to the real engine AHR profile
suggests that it should be able to be realized in engine combustion.
8.2.3 Firing cycle calculation
Once the apparent heat release is generated, the cylinder pressure during firing
cycle can be easily calculated. From Equation (36), using a similar numerical procedure
as 8.2.1:
γ ∆V 1 P − P0
AHR = P + V (51)
γ − 1 ∆θ γ − 1 ∆θ
Then the cylinder pressure after ∆θ from a previous pressure P0 is
AHR (γ − 1) ∆θ + P0V
P= (52)
γ (V − V0 ) + V
Similarly, an iteration procedure is needed since γ needs to be consistent with the
temperature. The initial point P0 should choose the motor pressure at start of combustion,
i.e., when AHR > 0 for the first time with increased crank angle. Note Equation (52) is
also only valid during firing, i.e., as long as AHR > 0. Meanwhile the rate of energy input
from fuel will be:
dqt ∆θ
= AHR + qɺw (53)
dθ ω
161
Since the total energy input should compensate both the pressure trace and heat
transfer during the firing. This should be also calculated only during firing. With a known
pressure trace during firing from Equation (52), qɺ w can be readily calculated. The total
energy input from fuel is:
dqt
qt = ∑ dθ ∆θ
AHR
(54)
After the firing crank angle (AHR = 0 again), Equation (50) from Section 8.2.1
should be used to calculate the pressure trace.
Figure VIII-4 compares simulated pressure trace and real engine data, running at
1500 rpm, 30% load, single injection timing at 1º before top dead center, 100 MPa fuel
injection pressure. A good match can be observed (R2 = 0.9648).
162
100
Simulated pressure trace
90 Engine pressure trace
80
70
60
Pressure (bar)
50
40
30
20
10
0
0 100 200 300 400 500 600 700
CA (degree)
Figure VIII-4: comparison of simulated pressure trace with engine data, running at
1500 rpm, 30% load, single injection timing at 1º before top dead center, 100 MPa
fuel injection pressure
After the pressure trace is obtained, it is straightforward to calculate friction work
and energy thermal efficiency. With an apparent heat release identical to the 30% load
engine data, the model gives an indicated efficiency of 49.7% and a brake specific
efficiency of 32.4%. These values are reasonable for this engine [from around 32% (low
load) to around 37% (high load), see Figures IV-11 to IV-13].
8.3 Optimized heat release pattern probing
163
As shown earlier (Figures VIII-2 and VIII-8), the artificially generated apparent
heat release can be deconvoluted to two peaks:
AHR = pG ( h1 , w1 , l1 ) + pE ( h2 , l2 , k1 , k 2 ) (55)
θ − l1
pG ( h1 , w1 , l1 ) = h1 exp[ −( )2 ]
w1

0,if θ<l
 2
pE ( h2 , l2 , k1 , k2 ) =  h2 (θ − l2 ) k 22 exp[ − k2 (θ − l2 )],if k1 =k 2
h kk
 2 1 2 {exp[ − k1 (θ − l2 )] − exp[( − k 2 (θ − l2 )]},otherwise
 k 2 − k1
Where, h1 and h2 are the peak height, w1 is the peak width, l1 and l2 are the
locations of the peaks (determining the start of combustion), and k1, k2 are two
parameters that control the shape of ExpConvExp peak, pE. It is noticed that pE has four
parameters, which introduces high complexity for adjustment. In this work, the
independent adjustment will be conducted primarily on the Gaussian peak pG through h1,
w1 and l1. The h2 of pE will be adjusted based on the principle that the total energy of
apparent heat release is constant, i.e.,
∫ AHR ⋅ dθ = constant (56)
The baseline of the AHR was the default engine output (Figure VIII-2), which
will be indicated by parameters h1b , w1b , l1b , h2b , l2b . To investigate the effect of start of
combustion (SOC), l1 is varied, and l2 is varied with l1 at the principle that the distance
between l1 and l2 is constant, i.e., l2-l1 = constant. Later the impact of distance between l1
and l2 on engine thermal efficiency will also be investigated.
164
8.3.1 The effect of peak height
Figure VIII-5 shows three apparent heat release profiles. These patterns have
different heights in the Gaussian peak. The height of ExpConvExp peak was adjusted
along with the height of the Gaussian peak, so the total area under AHR was constant. A
significant peak height of maximum heat release rate can be observed.
Figure VIII-6 investigates the indicated thermal efficiency of the heat release
profiles in Figure VIII-5. Each curve in Figure VIII-6 is generated corresponding to one
heat release rate pattern. The apparent heat release pattern was literally shifted around the
top dead center. At each location, both the start of combustion (which is noted as the first
location the apparent heat release is higher than 0.1 J/degree) and engine indicated
thermal efficiency were recorded. Figure VIII-6 then summarizes the change of
indicated efficiency with start of combustion. The first observation is that the maximum
indicated efficiencies are all around the start of combustion of 354º. With decreased peak
height, the SOC timing for the maximum indicated efficiency advances slightly. In
addition, the maximum indicated efficiency increases slightly with decreased peak height.
165
200
h1/h1b = 0.2
180 h1/h1b = 0.6
h1/h1b = 1
160
140
AHR (J/degree)
120
100
80
60
40
20
0
350 355 360 365 370 375 380 385 390 395 400
CA (degree)
Figure VIII-5: Examples of apparent heat release with: h1/h1b = 0.2, h2/h2b=1.29;
h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1
166
0.51
h1/h1b = 0.2
h1/h1b = 0.6
0.505 h1/h1b = 1
0.5
indicated efficiency
0.495
0.49
0.485
0.48
0.475
350 352 354 356 358 360 362 364 366
Start of Combustion (degree)
Figure VIII-6: Effect of start of combustion on indicated thermal efficiency, h1/h1b

= 0.2, h2/h2b=1.29; h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1
Figure VIII-7 presents the effect of SOC on brake thermal efficiency. Similar to
indicated efficiency observation, the maximum brake efficiency also requires an early
SOC. However, decrease of peak height seems to be able to increase the maximum brake
thermal efficiency for the particular friction model employed here. This could be due to
decreased friction from lower cylinder pressure, when a low height is used for the
apparent heat release peak.
167
0.34
h1/h1b = 0.2
h1/h1b = 0.6
0.335
h1/h1b = 1
0.33
0.325
Brake efficiency
0.32
0.315
0.31
0.305
0.3
350 352 354 356 358 360 362 364 366 368
State of Combustion (degree)
Figure VIII-7: Effect of start of combustion on brake thermal efficiency, h1/h1b =

0.2, h2/h2b=1.29; h1/h1b = 0.6, h2/h2b=1.17; h1/h1b = 1, h2/h2b=1
Figures VIII-8 to VIII-10 tested the impact of increasing the Gaussian peak
height. Consequently, with an increased Gaussian peak area, the ExpConvExp peak will
decrease. For the high Gaussian peak case, the ExpConvExp almost disappears, resulting
of a symmetry heat release rate. A consistent observation with Figure VIII-6 is that the
timing of SOC for the maximum indicated thermal efficiency retards with increased peak
height. However, a further increase of the peak height will slightly increase the maximum
indicated efficiency.
168
600
h1/h1b = 1
h1/h1b = 2
500 h1/h1b = 3
h1/h1b = 4
400
AHR (J/degree)
300
200
100
0
350 355 360 365 370 375 380 385 390 395 400
CA (degree)
Figure VIII-8: Examples of apparent heat release with: h1/h1b = 1, h2/h2b=1;

h1/h1b = 2, h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4, h2/h2b=0.02
169
0.51
0.505
0.5
Indicated efficiency
0.495
0.49
0.485
h1/h1b = 1
h1/h1b = 2
0.48 h1/h1b = 3
h1/h1b = 4
0.475
350 352 354 356 358 360 362 364 366
Figure VIII-9: Effect of start of combustion on indicated thermal efficiency, h1/h1b

= 1, h2/h2b=1; h1/h1b = 2, h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4,
h2/h2b=0.02
Figure VIII-10 shows the brake thermal efficiency with an increase of Gaussian
peak height. The timing of SOC for the maximum heat release retards with the increased
peak height. Meanwhile, the maximum brake efficiency does not show a monotonic trend.
With increase of peak height, the maximum brake efficiency decreases first then
increases.
170
0.34
0.335
0.33
Brake efficiency
0.325
0.32
0.315
h1/h1b = 1
h1/h1b = 2
0.31 h1/h1b = 3
h1/h1b = 4
0.305
350 352 354 356 358 360 362 364 366
Figure VIII-10: Effect of start of combustion on brake thermal efficiency, h1/h1b =

1, h2/h2b=1; h1/h1b = 2, h2/h2b=0.66; h1/h1b = 3, h2/h2b=0.34; h1/h1b = 4,
h2/h2b=0.02
8.3.2 The effect of peak width
Figures VIII-11 to VIII-13 investigated the effect of Gaussian peak width (this
may represent the duration of premixed combustion). By decreasing the peak width, the
ExpConvExp peak size can be slightly increased. The behaviors of indicated thermal
efficiencies are quite similar for decreased Gaussian peak. For brake thermal efficiency, a
decrease in Gaussian peak seems to be able to increase the maximum brake thermal
efficiency. Both the indicated and brake thermal efficiency requires early SOC to reach
the maximum value.
171
250
w1/w1b = 0.6
w1/w1b = 0.8
w1/w1b = 1
200
AHR (J/degree)
150
100
50
0
350 355 360 365 370 375 380 385 390 395 400
CA (degree)
Figure VIII-11: Examples of apparent heat release with: w1/w1b = 0.6, h2/h2b=1.17;
w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1
172
0.51
w1/w1b = 0.6
w1/w1b = 0.8
0.505
w1/w1b = 1
0.5
0.495
0.49
0.485
0.48
0.475
352 354 356 358 360 362 364 366 368
Figure VIII-12: Effect of start of combustion on indicated thermal efficiency,

w1/w1b = 0.6, h2/h2b=1.17; w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1
173
0.34
w1/w1b = 0.6
w1/w1b = 0.8
0.335 w1/w1b = 1
0.33
Brake efficiency
0.325
0.32
0.315
0.31
0.305
352 354 356 358 360 362 364 366 368
Start of Combustion
Figure VIII-13: Effect of start of combustion on brake thermal efficiency, w1/w1b =

0.6, h2/h2b=1.17; w1/w1b = 0.8, h2/h2b=1.11; w1/w1b = 1, h2/h2b=1
Figure VIII-14 to VIII-16 also investigated the effect of peak width. For this
investigation, the Gaussian peak width is increased, which at the same time forced a
decrease for the peak height. With increased peak width, the timing of SOC for the
maximum indicated thermal efficiency advances, and the maximum value can also be
able to increase by about 0.7%. Similar observations can be found for brake thermal
efficiency. With increased peak width, the brake efficiency can be increased by 1%.
174
200
w1/w1b = 1
180 w1/w1b = 2
w1/w1b = 3
160 w1/w1b = 4
140
AHR (J/degree)
120
100
80
60
40
20
0
350 355 360 365 370 375 380 385 390 395 400
CA (degree)
Figure VIII-14: Examples of apparent heat release with: w1/w1b = 1, h2/h2b=1;

w1/w1b = 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b = 4, h2/h2b=0.02
175
0.515
w1/w1b = 1
w1/w1b = 2
0.51 w1/w1b = 3
w1/w1b = 4
0.505
0.5
0.495
0.49
0.485
0.48
340 345 350 355 360 365
Figure VIII-15: Effect of start of combustion on indicated thermal efficiency,

w1/w1b = 1, h2/h2b=1; w1/w1b = 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b
= 4, h2/h2b=0.02
176
0.345
0.34
0.335
0.33
Brake efficiency
0.325
0.32
0.315
w1/w1b = 1
w1/w1b = 2
0.31 w1/w1b = 3
w1/w1b = 4
0.305
340 345 350 355 360 365
Figure VIII-16: Effect of start of combustion on brake thermal efficiency, w1/w1b =

1, h2/h2b=1; w1/w1b = 2, h2/h2b=0.66; w1/w1b = 3, h2/h2b=0.34; w1/w1b = 4,
h2/h2b=0.02
8.3.3 The effect of peak distance
Figures VIII-17 to VIII-19 investigates the effect of peak distance. The default
peak convolution was used for this purpose. It can be observed that this split-peak does
not have a significant impact on the maximum indicated and brake thermal efficiency.
177
200
∆L = 0 deg
180 ∆L = 2 deg
∆L = 4 deg
160
∆L = 6 deg
140
AHR (J/degree)
120
100
80
60
40
20
0
350 355 360 365 370 375 380 385 390 395 400
CA (degree)
Figure VIII-17: Examples of apparent heat release with different peak distances
178
0.51
0.505
0.5
0.495
0.49
0.485
0.48
0.475
∆L = 1 deg
0.47 ∆L = 2 deg
∆L = 3 deg
0.465
∆L = 4 deg
0.46
350 352 354 356 358 360 362 364 366
Figure VIII-18: Effect of start of combustion on indicated thermal efficiency with

different peak distance, ∆L = l2-l1
179
0.34
0.33
0.32
Brake efficiency
0.31
0.3
∆L = 1 deg
∆L = 2 deg
0.29
∆L = 3 deg
∆L = 4 deg
350 352 354 356 358 360 362 364 366

Start of Combustion (deg)
Figure VIII-19: Effect of start of combustion on brake thermal efficiency with

different peak distance
Finally, two extreme cases were investigated: a very sharp peak and a very wide
but low peak (Figure VIII-20). Significant differences can be observed. The timing of
SOC for the maximum indicated and brake thermal efficiency is significantly advanced
with a low and wide peak, and the indicated and brake thermal efficiency of low and
wide peak appears to be generally higher than that of a sharp peak (Figures VIII-21 and
VIII-22).
180
1000 w1/w1b=8
h1/h1b = 8
800
AHR (J/degree)
600
400
200
0
350 360 370 380 390 400
CA (degree)
Figure VIII-20: Examples of apparent heat release profiles for sharp peak (h1/h1b =
8) versus low and wide peak (w1/w1b = 8)
181
0.52 w1/w1b=8
h1/h1b = 8
0.51
0.50
0.49
0.48
0.47
325 330 335 340 345 350 355 360 365 370
Figure VIII-21: Indicated thermal efficiency for sharp peak (h1/h1b = 8) versus low
and wide peak (w1/w1b = 8)
182
0.36
w1/w1b=8
h1/h1b = 8
0.35
brake efficiency
0.34
0.33
0.32
0.31
325 330 335 340 345 350 355 360 365 370
Figure VIII-22: Brake thermal efficiency for sharp peak (h1/h1b = 8) versus load
and wide peak (w1/w1b = 8)
It is of interest to investigate the origin of engine thermal efficiency differences
from the change of apparent heat release pattern. Figure VIII-23 shows the heat transfer
loss of the two extreme heat release patterns (Figure VIII-20). It can be observed that the
difference of heat transfer loss between the two cases is not significant. However, the
cylinder pressure traces (Figure VIII-24) are significantly different between the two heat
release patterns. The “sharp” heat release gives a much higher maximum cylinder
pressure and narrower pressure peak, while the “wide and low” heat release pattern gives
a lower maximum cylinder pressure.
183
240
h1/h1b=8
w1/w1b=8
220
200
Heat loss (Joule)
180
160
140
120
100
80
325 330 335 340 345 350 355 360 365 370
Figure VIII-23: Heat transfer loss for sharp peak (h1/h1b = 8) versus load and wide
peak (w1/w1b = 8)
184
160
h1/h1b=8
140 w1/w1b=8
120
100
Pressure (bar)
80
60
40
20
300 320 340 360 380 400 420 440

CA (degree)
Figure VIII-24: Comparison of cylinder pressure for sharp peak (h1/h1b = 8) versus
load and wide peak (w1/w1b = 8) apparent heat release
The significant difference in the pressure trace between these two heat release
patterns leads to a significant difference in the indicated work. As shown in Figure VIII-
25, the “wide and low” apparent heat release pattern gives generally higher indicated
work, i.e., given similar energy input, the “wide and low” heat release pattern can output
higher work. This explains the higher engine thermal efficiency from the “wide and low”
apparent heat release pattern.
185
650
h1/h1b=8
w1/w1b=8
640
Indicated work (Joule)
630
620
610
600
590
580
570
325 330 335 340 345 350 355 360 365 370
Figure VIII-25: Indicated work for sharp peak (h1/h1b = 8) versus low and wide
peak (w1/w1b = 8)
Figure VIII-26 illustrates an example of comparison of P-V diagram between the
two extreme cases. Although the “sharp” heat release profile appears to benefit from the
higher pressure at low volume, the total area of P-V diagram from “low and wide” heat
release profile actually is higher given that the pressure from the “low and wide” heat
release profile is generally higher than that from the “sharp” heat release profile for a
significant range of volume. This illustrates the differences between indicated work from
the two extreme cases.
186
140 h1/h1b=8
w1/w1b=8
120
100
Pressure (Bar)
80
60
40
20
0
0 200 400 600 800
Volume (cm3)
Figure VIII-26: P-V diagram for sharp peak (h1/h1b = 8) versus low and wide peak
(w1/w1b = 8)
It must be acknowledged that the higher engine thermal efficiency from a heat
release profile that is significantly different from a “sharp” shape, i.e., the P-V diagram is
significantly different from the air standard Otto cycle, is counter intuitive. However,
Hoppie et al. [32] in their thermodynamic engine simulation concluded that the Otto
cycle is only an optimized cycle if all loss functions (friction, heat transfer, mass leakage,
etc.) are independent of pressure if maximum cylinder pressure is not constrained. The
model employed in this work indeed calculated the heat transfer loss, for which cases
Hoppie et al. [32] suggested neither Otto nor Diesel cycle is optimum. In fact, Lyn [254]
in his thermodynamic engine simulation concluded that for a given heat release period,
the cycle efficiency is a single valued function of peak pressure, and there appears to be
187
an optimized peak pressure beyond which no significant efficiency benefit can be
observed. For instance, Figure VIII-27 illustrates two kinds of heat release profiles:
AHR1 is more close to the “sharp” shape while AHR2 is more close to the “low and wide”
shape.
200 AHR 1
AHR 2
150
AHR (J/degree)
100
50
300 330 360 390 420 450

CA (degree)
Figure VIII-27: Comparison of two apparent heat release (AHR) profiles with the
same start of combustion timing at 362.2º (2.2º after top dead center), indicated
efficiency of AHR1 = 49.74%, indicated efficiency of AHR2 = 47.23%
With the model employed in this work, the engine cylinder pressure traces of the
two heat release profiles can be easily calculated, as shown in Figure VIII-28. It can be
observed that the peak pressure from the more “low and wide” heat release profile is
significantly lower than that from the more “sharp” heat release profile, and the indicated
thermal efficiency from the more “low and wide” heat release profile is also lower (47.23%
188
to 49.74%). The observation from Figure VIII-28 is entirely consistent with the
observation by Lyn [254], which indicates that if the majority of heat release occurs
during the expansion cycle, a “sharp” heat release profile is preferred since it can
generate higher peak pressure.
160
AHR 1
140
AHR 2
120
Pressure (bar)
100
80
60
40
20
0
300 320 340 360 380 400
CA (degree)
Figure VIII-28: Comparison of cylinder pressure traces of two apparent heat

release (AHR) profiles with the same start of combustion timing at 362.2º (2.2º after
top dead center), indicated efficiency of AHR1 = 49.74%, indicated efficiency of
AHR2 = 47.23%
At the same time, one can tend to advance the start of combustion. As shown in
Figure VIII-29, the exact same shapes of heat release profiles in Figure VIII-27 can be
shifted to an earlier start of combustion timing at 350º (10º before top dead center). The
cylinder pressure trances generated from these two heat release profiles can be found in
Figure VIII-30. It can be observed that the peak pressures of both heat release profiles
189
increase compared to the case with later start of combustion. This explains the increase of
thermal efficiency of both cases. However, the increase of peak pressure for the more
“low and wide” heat release profile has a more significant improvement on the cycle
efficiency. In addition, the extreme high cylinder pressure during the compression cycle
for the more “sharp” heat release profile is actually generating a negative impact on the
cycle efficiency, while its increased peak pressure has a less significant impact on the
cycle efficiency given the peak pressure from the later start of combustion is considerably
high already. These combined effects lead to a higher cycle efficiency for the more “low
and wide” heat release profile (50.84% to 50.02%).
200
AHR 1
AHR 2
150
AHR (J/degree)
100
50
300 330 360 390 420 450

CA (degree)
Figure VIII-29: Comparison of two apparent heat release (AHR) profiles with the
same start of combustion timing at 350º (10º before top dead center), indicated
efficiency of AHR1 = 50.02%, indicated efficiency of AHR2 = 50.84%
190
160
AHR 1
140
AHR 2
120
Pressure (bar)
100
80
60
40
20
0
300 320 340 360 380 400
CA (degree)
Figure VIII-30: Comparison of cylinder pressure traces of two apparent heat

release (AHR) profiles with the same start of combustion timing at 350º (10º before
top dead center), indicated efficiency of AHR1 = 50.02%, indicated efficiency of
AHR2 = 50.84%
The observations from Figures VIII 27 to VIII-30 suggest an explanation of the
unexpected high efficiency from the “low and wide” heat release profile: the “low and
wide” heat release profile with an early start of combustion can effectively make use of
the piston compression movement to reach a high peak cylinder pressure while maintain
a low negative work generated before the expansion cycle. This advantage is usually
difficult to be observed for the “sharp” heat release profile since it tends to either
generate high negative work if it has an early start of combustion or a lower peak
pressure if it has a late start of combustion.
191
In summary, this zero-dimensional engine efficiency simulation presents the
surprising result that an optimized apparent heat release pattern should possess the feature
of a low peak value and a wide profile. This suggests that the one hypothesis that a sharp
peak gives higher engine efficiency is not correct. The reason for this surprising result is
that a lower heat release peak can make use of piston compression to reach a high peak
cylinder pressure without introducing significant negative work.
It can be observed that it is extremely difficult to improve diesel engine efficiency
(even with a dramatic change in apparent heat release pattern as Figure VIII-20, the
maximum improvement in engine efficiency is around 2%). The chief observation from
this simulation work is that the most important consideration in the attempt to improve
engine thermal efficiency is to decrease the heat loss. A “wide and low” apparent heat
release pattern will certainly benefit due to increase of indicated work output. This kind
of “wide and low” apparent heat release can be actually approached from advanced
combustion, e.g., lean-fuel combustion [255]. It will be also interesting to investigate
other factors with this model, such as the maximum cylinder pressure, friction with
advanced lubricating oil and turbo-compounding. However, such considerations were not
included in the scope of this work.
192
Chapter IX. Conclusions and recommendations for future
work
From this work, the fully validated hypotheses include:
• The impact on NOx emission from engine fuel injection strategy control and
biodiesel fueling can be explained by the equivalence ratio at autoignition
zone near the flame lift-off length.
• Soot from high injection pressure has higher oxidative reactivity since the
decreased soot precursor concentration can prolong the soot inception time.
The conditionally validated hypotheses (only valid at certain conditions) are:
• For conventional diesel combustion at high efficiency (fuel injection timing
earlier that 5º before top dead center), biodiesel NOx penalty can be
eliminated by engine fuel injection control. With a decrease fuel injection
pressure, the biodiesel NOx emissions can be reduced to the same level as
petroleum diesel, while maintaining the same level of particulate emissions.
This approach is less effective for combustion with late fuel injection timing.
• Biodiesel fueling can introduce lubricating oil dilution with post injection
strategy. This statement depends on the timing of post injection, and only a
post injection timing later than 45º after top dead center can introduce
lubricating oil dilution.
193
The invalidated hypothesis is:
• Optimized heat release design is a sharp peak at top dead center. On the
contrary, a wide and low peak is the optimized heat release pattern to improve
engine thermal efficiency since it can decrease the heat loss during
combustion.
Other conclusions from this work include:
In a compression ignition direct injection engine, the increase of injection quantity
when biodiesel is used can be obtained by two means: (1) increase of injection pressure
and (2) extension of injection duration. Experiments of investigating both these two
factors have been conducted and the NOx increase is observed for both cases when
biodiesel is used, which suggests that both types of change in injection strategy (injection
pressure or duration) have similar impacts on NOx emissions. Nevertheless, a change of
injection pressure seems to have a higher impact on NOx emissions than a change of
injection duration.
Meanwhile, the impacts of injection strategy and biodiesel fueling on PM
emissions strongly depend on the engine load condition. At low load condition, both the
increase of fuel injection pressure and biodiesel fueling can significantly reduce PM
emissions. However, at moderate to high load conditions, the biodiesel fueling has a less
significant impact on PM emissions.
No significant difference in brake thermal efficiency is observed with biodiesel
fueling or change of fuel injection pressure. However, it is observed that retarding the
start of injection can reduce the brake thermal efficiency. The highest brake thermal
194
efficiency in this work has been observed at SOI between 9˚ to 5˚ before top dead center.
In addition, the NOx-PM emission trade-off of SOI timing at 9˚ BTDC shows that
decreasing the fuel injection pressure is an effective way to counter the biodiesel NOx
effect while maintaining a similar or even lower level of PM emissions than ULSD.
Therefore, for an engine which intends to run at a high brake thermal efficiency, a
decrease of fuel injection pressure is an applicable approach to solve the biodiesel NOx
penalty problem.
The heat release analysis shows that, for all load conditions, retarding the start of
injection after top dead center can increase the fraction of premixed combustion. At the
same SOI, higher fuel injection pressure leads to higher apparent heat release rate and a
slightly earlier start of combustion (SOC) due to the better mixing of fuel and air. At low
load conditions, the biodiesel fueling can advance the SOC and decrease the premixed
heat release. At moderate to high loads, however, the biodiesel fueling does not have a
significant impact on the heat release profile. Therefore in this study, combustion phasing
is not a primary factor that contributes to the biodiesel NOx effect for higher load.
It is confirmed that the key factor that determines the NOx emissions is the
closeness to stoichiometric conditions of the oxygen equivalence ratio of the fuel-air
mixture at the autoignition zone near the lift-off length, for all load conditions and a
range of start of injection from 9˚ BTDC to 3˚ ATDC. On the basis of the fuel spray
model by Pickett, Siebers and co-workers, the average oxygen equivalence ratios at lift-
off length are observed to be linearly correlated with brake specific NOx emissions at the
same SOI for all load conditions. In addition, the slopes and Y-axis intercepts (at NOx
emissions equal to 0) of the correlations are found to be exponentially correlated with
195
SOI, based on which general equations to predict NOx emissions are proposed. These
correlations are applicable for both petroleum diesel and biodiesel. In addition, given
their significantly lower requirement for CPU power than complicated computational
fluid dynamic models, these correlations may be implemented into the electronic control
unit of a modern engine to perform post (or even real time) evaluation of the engine NOx
emissions
Multi-injection strategy including post injection is an effective way to decrease
NOx emissions and increase exhaust temperature for diesel engine. Depending on the
timing and quantity of post injection, the engine showed a significant change in exhaust
temperature, brake specific fuel consumption, UHC and CO emissions. Between the post
injection timings from 45º to 70º ATDC, the consumption of post injected fuel was found
to primarily depend on the injection timing.
Post injection at 45º ATDC with the quantity of 5mg/stk had significant
combustion at late crank angle (LCA), i.e., later than 40º after top dead center.
Consequently it generated power – reducing the energy requirement for the pilot and
main injections – and produced a significantly higher exhaust temperature without an
apparent increase in BSFC. The UHC and CO emissions only slightly increased
compared with the test without post injection.
Post injection at 60º ATDC (5mg/stk), 60º ATDC (8mg/stk) and 70º ATDC
(5mg/stk) did not have apparent heat release at LCA. Compared with 45º ATDC
(5mg/stk), they barely generated any power, decreased the exhaust temperature and
significantly increased BSFC, UHC and CO emissions.
196
A model was employed to calculate the ignition delay of post injected fuel. The
temperature at which the ignition was supposed to happen was confirmed to coordinate
with the quality of combustion of post injected fuel at late crank angle. A bulk cylinder
temperature of 1500K was confirmed to be the lowest acceptable temperature for which
high temperature combustion of the post injected fuel could occur in the engine.
When the fuel was injected into an environment with temperature lower than
1500K, only slow fuel oxidation occurs. Hence, the fuel escaped the cylinder at a much
higher concentration than during high temperature combustion. GC-MS analysis showed
that the primary hydrocarbons in the exhaust gases were vaporized diesel and biodiesel
components regardless of the post injection strategy, and those hydrocarbons became the
primary source for lubricant dilution when they were dissolved into the oil film on the
cylinder wall. Additionally, in this experiment only alkanes with carbon number larger
than 14 from the diesel fuel and biodiesel were observed to dilute into the lubricant
because of the high temperature of the cylinder wall. The comparative mass fraction of
fatty acid methyl esters shifted towards more saturated and higher boiling point
components when biodiesel was diluted into the lubricant.
Fuel injection pressure can significantly affect the morphology of soot aggregates.
Increase of fuel injection pressure can decrease the fractal dimension of the aggregates
and decrease the number of primary particle. Increase of fuel injection pressure can also
increase the nucleation mode peak and decrease the accumulation mode peak in the
particle size distribution.
At the same fuel injection pressure, B20 soot has higher oxidative reactivity.
Meanwhile, the reactivity of soot particle is varies with burn-off. It often increases with
197
higher burn-off. However, for soot generated at very high fuel injection pressure, the
oxidative reactivity tends to firstly increase with burn-off and decreased when the burn-
off is close to completion.
The recommended future work includes:
• Investigate the kinetic role of start of injection timing in engine NOx
emissions. As shown in Chapter V, the start of injection timing serves as a
parameter similar to the time in a kinetic model.
• Investigate lubricating oil dilution in an advanced combustion mode, such as
low temperature combustion. The investigation of post injection in this work
(Chapter VI) is still conventional combustion. A significant increase of EGR
can promote advanced combustion, which may introduce more serious oil
dilution.
• Investigate the evolution of active surface area and carbon crystalline structure
of diesel and biodiesel soot from different fuel injection pressures. The
information of active surface area and carbon nanostructure during burn-off
can be critical in the explanation of evolution of instantaneous reactivity of
soot particles.
• Investigate the impact of maximum cylinder pressure, friction with advanced
lubricating oil and turbo compounding on engine efficiency. This can be
implemented to investigate the pathways of high engine thermal efficiency.
198
References
[1] J. Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill, Inc, (1988).
[2] T.V. Johnson, Review of Diesel Emissions and Control, International Journal of
Engine Research, 10 (2009) 275-285.
[3] DieselNet, United States New Engine and Vehicle Emissions,
http://www.dieselnet.com/standards/, (2011).
[4] H. Lund, Renewable Energy Strategies for Sustainable Development, Energy, 32
(2007) 912-919.
[5] ASTM, Standard Specification for Biodiesel Fuel Blend Stock (B100) for Middle
Distillate Fuels, in, ASTM International Specification ASTM D 6751, 2008.
[6] M. Graboski, R. McCormick, Combustion of Fat and Vegetable Oil Derived Fuels in
Diesel Engines, Progree in Energy and Combustion Science, 24 (1998) 125-164.
[7] G. Knothe, "Designer" Biodiesel: Optimizing Fatty Ester (Composition to Improve
Fuel Properties, Energy & Fuels, 22 (2008) 1358-1364.
[8] R.L. McCormick, The Impact of Biodiesel on Pollutant Emissions and Public Health,
Inhalation Toxicology, 19 (2007) 1033-1039.
[9] J. Yanowitz, R.L. Mccormick, Effect of Biodiesel Blends on North American Heavy-
Duty Diesel Engine Emissions, European Journal of Lipid Science and Technology, 111
(2009) 763-772.
[10] M. Graboski, J.D.Ross, R.L.McCormick, Transient Emissions from No. 2 Diesel and
Biodiesel Blends in a DDC Series 60 Engine., SAE paper 961166 (1996).
199
[11] S. Murillo, J. Miguez, J. Porteiro, E. Granada, J. Moran, Performance and Exhaust
Emissions in the Use of Biodiesel in Outboard Diesel Engines, Fuel, 86 (2007) 1765-
1771.
[12] T.L.Allenman, R.L.Mccormick, Results of the 2007 B100 Quality Survey, NREL
Technical Report NREL/TP 540-42787, (2008).
[13] Y. Zhang, A.L. Boehman, Impact of Biodiesel on NOx Emissions in a Common Rail
Direct Injection Diesel Engine, Energy & Fuels, 21 (2007) 2003-2012.
[14] J.P. Szybist, A.L. Boehman, J.D. Taylor, R.L. Mccormick, Evaluation of
Formulation Strategies to Eliminate the Biodiesel NOx Effect, Fuel Processing
Technology, 86 (2005) 1109-1126.
[15] M. Lapuerta, O. Armas, J. Rodriguez-Fernandez, Effect of Biodiesel Fuels on Diesel
Engine Emissions, Progress in Energy and Combustion Science, 34 (2008) 198-223.
[16] G.A. Ban-Weiss, J.Y. Chen, B.A. Buchholz, R.W. Dibble, A Numerical
Investigation into the Anomalous Slight NOx Increase When Burning Biodiesel: a New
(Old) Theory, Fuel Processing Technology, 88 (2007) 659-667.
[17] C.Y. Choi, R.D. Reitz, A Numerical Analysis of the Emissions Characteristics of
Biodiesel Blended Fuels, Journal of Engineering for Gas Turbines and Power,
Transactions of the ASME, 121 (1999) 31-37.
[18] C.J. Mueller, A.L. Boehman, G. Martin, An Experimental Investigation of the Origin
of Increased Nox Emissions When Fueling a Heavy-Duty Compression-Ignition Engine
with Soy Biodiesel, SAE paper 2009-01-1792, (2009).
200
[19] K. Yamamoto, K. Takada, J. Kusaka, Y. Kanno, M. Nagata, Influence of Diesel Post
Injection Timing on Hc Emissions and Catalytic Oxidation Performance, SAE paper
2006-01-3442, (2006).
[20] M.M. Andreae, H.L. Fang, K. Bhandary, Biodiesel on Fuel Dilution of Engine Oil,
SAE paper 2007-01-4036, (2007).
[21] P.J. Tennison, R. Reitz, An Experimental Investigation of the Effects of Common-
Rail Injection System Parameters on Emissions and Performance in a High-Speed Direct-
Injection Diesel Engine, Journal of Engineering for Gas Turbines and Power,
Transactions of the ASME, 123 (2001) 167-174.
[22] P.K. Karra, M.K. Veltman, S.C. Kong, Characteristics of Engine Emissions Using
Biodiesel Blends in Low-Temperature Combustion Regimes, Energy & Fuels, 22 (2008)
3763-3770.
[23] A.L. Boehman, D. Morris, J. Szybist, E. Esen, The Impact of the Bulk Modulus of
Diesel Fuels on Fuel Injection Timing, Energy & Fuels, 18 (2004) 1877-1882.
[24] G. Knothe, Dependence of Biodiesel Fuel Properties on the Structure of Fatty Acid
Alkyl Esters, Fuel Processing Technology, 86 (2005) 1059-1070.
[25] K. Yehliu, A.L. Boehman, O. Armas, Emissions from Different Alternative Diesel
Fuels Operating with Single and Split Fuel Injection, Fuel, 89 (2010) 423-437.
[26] D.T. Hountalas, V. Lamaris, E. Pariotis, H. Ofner, Parametric Study Based on a
Phenomenological Model to Investigate the Effect of Post Fuel Injection on HDDI Diesel
Engine Performance and Emissions - Model Validation Using Experimental Data, SAE
paper 2008-01-0641, (2008).
201
[27] W.D. Ojeda, P. Zoldak, R. Espinosa, R. Kumar, Development of a Fuel Injection
Strategy for Partially Premixed Compression Ignition Combustion, SAE paper 2009-01-
1527, (2009).
[28] H.H. Yun, R.D. Reitz, An Experimental Investigation on the Effect of Post-Injection
Strategies on Combustion and Emissions in the Low-Temperature Diesel Combustion
Regime, Journal of Engineering for Gas Turbines and Power, Transactions of the ASME,
129 (2007) 279-286.
[29] M. Yao, H. Wang, Z. Zheng, Y. Yue, Experimental Study of Multiple Injections and
Coupling Effects of Multi-Injection and EGR in a HD Diesel Engine, SAE paper 2009-
01-2807, (2009).
[30] A. Vanegas, H. Won, C. Felsch, M. Gauding, N. Peters, Experimental Investigation
of the Effect of Multiple Injections on Pollutant Formation in a Common-Rail Di Diesel
Engine, SAE paper 2008-01-1191, (2008).
[31] K. Anand, R.P. Sharma, P.S. Mehta, Experimental Investigations on Combustion,
Performance, and Emissions Characteristics of a Neat Biodiesel-Fueled, Turbocharged,
Direct Injection Diesel Engine, Journal of Automobile Engineering, 224 (2009) 661-679.
[32] L.O. Hoppie, Y.K. Min, N. Srinivasan, Optimum Heat Release for a Reciprocating
Internal Combustion Engine, SAE paper 870572 presented in International Congree and
Exposition in Detroit, MI, (1987).
[33] G. Knothe, Biodiesel and Renewable Diesel: A Comparison, Progress in Energy and
Combustion Science, 36 (2010) 364-373.
[34] A.K. Agarwal, Biofuels (Alcohols and Biodiesel) Applications as Fuels for Internal
Combustion Engines, Progress in Energy and Combustion Science, 33 (2007) 233-271.
202
[35] A. Demirbas, Progress and Recent Trends in Biofuels, Progress in Energy and
[36] G. Knothe, Historical Perspectives on Vegetable Oil-Based Diesel Fuels, INFORM,
12 (2001) 1103-1107.
[37] A. Chavanne, Fuels from Vegetable Oils. (Procede De Transformation D'huiles
Vegetables En Vue De Leur Utilisation Comme Carburants), Belgian Patent BE422,877,
(1937).
[38] A. Srivastava, R. Prasad, Triglycerides-Based Diesel Fuels, Renewable &
Sustainable Energy Reviews, 4 (2000) 111-133.
[39] B.R. Moser, Biodiesel Production, Properties, and Feedstocks, In Vitro Cell DEV-
PL, 45 (2009) 229-266.
[40] F.D. Funstone, J.L. Harwood, A.J. Dijkstra, The Lipid Handbook, 3rd Ed., Roca
Raton, 2007.
[41] G. Knothe, K.R. Steidley, Kinematic Viscosity of Biodiesel Fuel Components and
Related Compounds. Influence of Compound Structure and Comparison to Petrodiesel
Fuel Components, Fuel, 84 (2005) 1059-1065.
[42] G. Knothe, K.R. Steidley, Lubricity of Components of Biodiesel and Petrodiesel.
The Origin of Biodiesel Lubricity, Energy & Fuels, 19 (2005) 1192-1200.
[43] A. Schonborn, N. Ladommatos, J. Williams, R. Allan, J. Rogerson, The Influence of
Molecular Structure of Fatty Acid Monoalkyl Esters on Diesel Combustion, Combustion
and Flame, 156 (2009) 1396-1412.
[44] J. Sun, J.A. Caton, T.J. Jacobs, Oxides of Nitrogen Emissions from Biodiesel-Fueled
Diesel Engines, Progress in Energy and Combustion Science, 36 (2010) 677-695.
203
[45] Y. Ra, R. Reitz, J. Mcfarlane, C. Daw, Effects of Fuel Physical Properties on Diesel
Engne Combustion Using Diesel and Biodiesel Fuels, SAE paper 2008-01-1379 SAE
Trans - Journal of Fuels and Lubricants, (2008).
[46] M.E. Tat, J.H. Van Gerpen, Speed of Sound and Isentropic Bulk Modulus of Alkyl
Monoesters at Elevated Temperatures and Pressures, Journal of American Oil Chemistry
Society, 80 (2003) 1249-1256.
[47] H.S. Hess, The Impact of Oxygenated Fuels on Diesel Combustion and Emissions,
in: Department of Energy and Geo-Environmental Engineering, The Pennsylvania State
University, University Park, PA, 2002.
[48] C. Allen, K. Watts, Comparative Analysis of the Atomization Characteristics of
Fifteen Biodiesel Fuel Types, Transactions of the ASAE, 43 (2000) 207-211.
[49] S. Takahashi, K. Wakimoto, N. Iida, D. Nikolic, Effects of Aroamatics Content and
90% Distillation Temperature of Diesel Fuels on Flame Temperature and Soot Formation,
SAE paper 2001-01-1950 SAE Trans. - Journal of Fuels and Lubricants, (2001).
[50] W.A. Eckerle, E.J. Lyford-Pike, D. Stanton, J. Wall, L. Lapointe, S. Whitacre,
Effects of Methyl-Ester Biodiesel Blends on NOx Production, SAE paper 2008-01-0078,
(2008).
[51] EPA, Assessment and Standards Division (Office of Transportation and Air Quality
of the Us Environmental Protection Agency). A Comprehensive Analysis of Biodiesel
Impacts on Exhaust Emissions, EPA 420-P-02-001, (2002).
[52] FEV Engine Technology, Emissiosn and Performance Characteristics of the Navistar
T444E DI Engine Fueled with Blends of Biodiesel and Low Sulphur Diesel, Final report
to National Biodiesel Board, (1994).
204
[53] W. Marshall, L. Schumacher, S. Howell, Engine Exhaust Emissions Evaluation of a
Cummins L10E When Fueled with a Biodiesel Blend, SAE paper 952363, (1995).
[54] J.P. Szybist, J.H. Song, M. Alam, A.L. Boehman, Biodiesel Combustion, Emissions
and Emission Control, Fuel Processing Technology, 88 (2007) 679-691.
[55] P. Ye, A.L. Boehman, Investigation of the Impact of Engine Injection Strategy on
the Biodiesel NOx Effect with a Common-Rail Turbocharged Direct Injection Diesel
Engine, Energy & Fuels, 24 (2010) 4215-4225.
[56] A. Serdari, K. Fragioudakis, C. Teas, F. Zannikos, S. Stournas, E. Lois, Effect of
Biodiesel Addition to Diesel Fuel on Engine Performance and Emissions, Journal of
Propulsion and Power, 15 (1999) 224-231.
[57] K Hamasaki, E Kinoshita, H Tajima, K Takasaki, D Morita, Combustion
Characteristics of Diesel Engines with Waste Vegetable Oil Methyl Ester, The 5th
international symposium on diagnostics and modeling of combustion in internal
combustion engines (COMODIA), (2001).
[58] F. Staat, P. Gateau, The Effects of Rapeseed Oil Methyl Ester on Diesel Engine
Performance, Exhaust Emissiosn and Long Term Behaviour - a Summary of Three Years
of Experimentation, SAE paper 950053, (1995).
[59] M.E. Tat, Investigation of Oxides of Nitrogen Emissions from Biodiesel-Fueled
Engines, in: Mechanical Engineering, Iowa State University, 2003.
[60] R.L. McCormick, T.L. Alleman Impact of Biodiesel on Pollutant Emissions from
Diesel Engines, National Renewable Energy Laboratory, (2005).
[61] R.L. McCormick, Effects of Biodiesel on Nox Emissions, National Renewable
Energy Laboratory, (2005).
205
[62] X. Li, O.L. Gulder, Influence of Diesel Fuel Cetane Number and Aromatic Content
on Engine Exhaust Emissions, Journal of Canadian Petroleum Technology, 37 (1998) 56-
60.
[63] O. Armas, K. Yehliu, A.L. Boehman, Effect of Alternative Fuels on Exhaust
Emissions During Diesel Engine Operation with Matched Combustion Phasing, Fuel, 89
(2010) 438-456.
[64] T. Durbin, J. Collins, J. Norbeck, M. Smith, Effects of Biodiesel, Biodiesel Blends,
and a Synthetic Diesel on Emissions from Light Heavy-Duty Diesel Vehicles,
Environmental Science & Technology, 34 (2000) 349-355.
[65] S. Mandpe, S. Kadlaskar, W. Degen, S. Keppeler, On Road Testing of Advanced
Common Rail Diesel Vehicles with Biodiesel from the Jatropha Curcas Plant, SAE paper
2005-26-356 (2005 ).
[66] M.N. Nabi, M.S. Akhter, M.M.Z. Shahadat, Improvement of Engine Emissions with
Conventional Diesel Fuel and Diesel-Biodiesel Blends, Bioresource Technology, 97
(2006) 372-378.
[67] M.S. Graboski, R.L. McCormick, T.L. Alleman, A.M. Herring, The Effect of
Biodiesel Composition on Engine Emissions from a DDC Series 60 Diesel Engine,
National Renewable Energy Laboratory; NREL/SR-510-31461, (2003).
[68] M. Lapuerta, O. Armas, R. Ballesteros, J. Fernandez, Diesel Emissions from
Biofuels Derived from Spanish Potential Vegetable Oils, Fuel, 84 (2005) 773-780.
[69] M. Lapuerta, O. Armas, R. Ballesteros, Diesel Particulate Emissions from Biofuels
Derived from Spanish Vegetable Oils, SAE paper 2002-01-1657 (2002).
206
[70] M.P. Dorado, E. Ballesteros, J.M. Arnal, J. Gomez, F.J. Lopez, Exhaust Emissions
from a Diesel Engine Fueled with Transesterified Waste Olive Oil, Fuel, 82 (2003) 1311-
1315.
[71] W.G. Wang, D.W. Lyons, N.N. Clark, M. Gautam, P.M. Norton, Emissions from
Nine Heavy Trucks Fueled by Diesel and Biodiesel Blend without Engine Modification,
Environmental Science & Technology, 34 (2000) 933-939.
[72] M. Cardone, M. Prati, V. Rocco, M. Seggiani, A. Senatore, S. Vitolo, Brassica
Carinata as an Alternative Oil Crop for the Production of Biodiesel in Italy: Engine
Performance and Regulated and Unregulated Exhaust Emissions, Environmental Science
& Technology, 36 (2002) 4656-4662.
[73] J.F. Mcdonald, S.T. Bagley, L.D. Gratz, J.H. Johnson, Effects of an Oxidation
Catalytic Converter and a Biodiesel Fuel on the Chemical, Mutagenic, and Particle Size
Characteristics of Emissions from a Diesel Engine, Environmental Science &
Technology, 32 (1998) 1183-1191.
[74] K.F. Hansen, M.G. Jensen, Chemical and Biological Characteristics of Exhaust
Emissions from a Di Diesel Engine Fuelled with Rapeseed Oil Methyl Ester (Rme), SAE
paper 971689, (1997).
[75] A. Boehman, J. Song, M. Alam, Impact of Biodiesel Blending on Diesel Soot and
the Regeneration of Particulate Filters, Energy & Fuels, 19 (2005) 1857-1864.
[76] J.H. Van Gerpen, M. Canakci, Comparison of Engine Performance and Emissions
for Petroleum Diesel Fuel, Yellow Grease Biodiesel, and Soybean Oil Biodiesel,
Transactions of the ASAE, 46 (2003) 937-944.
207
[77] L. Schumacher, W. Marshall, J. Krahl, W. Wetherell, M.S. Graboski, Biodiesel
Emissions Data from Series 60 DDC Engines, Transactions of the ASAE 44 (2006) 1465-
1468.
[78] M. Haas, K. Scott, T. Alleman, R. Mccormick, Engine Performance of Biodiesel
Fuel Prepared from Soybean Soapstock: A High Quality Renewable Fuel Produced from
a Waste Feedstock, Energy & Fuels, 15 (2001) 1207-1212.
[79] M. Lapuerta, O. Armas, R. Ballesteros, M. Carmona, Fuel Formulation Effects on
Passenger Car Diesel Engine Particulate Emissions and Composition, SAE paper 2000-
01-1850 (2000).
[80] O. Armas, J.J. Hernandez, M.D. Cardenas, Reduction of Diesel Smoke Opacity from
Vegetable Oil Methyl Esters During Transient Operation, Fuel, 85 (2006) 2427-2438.
[81] R. Last, M. Kruger, M. Durnholz, Emissions and Performance Characteristics of a 4-
Stroke, Direct Injected Diesel Engine Fueled with Blends of Biodiesel and Low Sulfur
Diesel Fuel, SAE paper 950054, (1995).
[82] J. Krahl, A. Munack, M. Bahadir, L. Schumacher, N. Elser, Review: Utilization of
Rapeseed Oil, Rapeseed Oil Methyl Ester or Diesel Fuel: Exhaust Gas Emissions and
Estimation of Environmental Effects, SAE paper 962096, (1996).
[83] T. Durbin, J. Norbeck, Effects of Biodiesel Blends and Arco Ec-Diesel on Emissions
from Light Heavy-Duty Diesel Vehicles, Environmental Science & Technology, 36
(2002) 1686-1691.
[84] D.Y.C. Leung, Y. Luo, T.L. Chan, Optimization of Exhaust Emissions of a Diesel
Engine Fuelled with Biodiesel, Energy & Fuels, 20 (2006) 1015-1023.
208
[85] C. Peterson, D. Reece, Emissions Testing with Blends of Esters of Rapeseed Oil
Fuel with and without a Catalytical Converter, SAE paper 961114, (1996).
[86] J.A. Miller, C.T. Bowman, Mechanism and Modeling of Nitrogen Chemistry in
Combustion, Progress in Energy and Combustion Science, 15 (1989) 287-338.
[87] S.R. Turns, Understanding Nox Formation in Nonpremixed Flames: Experiments
and Modeling, Progress in Energy and Combustion Science, 21 (1995) 361-385.
[88] K.K. Kuo, Principles of Combustion, John Wiley & Sons, Inc, U.S.A., (1986).
[89] W. Yuan, A.C. Hansen, M.E. Tat, J.H. Van Gerpen, Z. Tan, Spray, Ignition, and
Combustion Modeling of Biodiesel Fuels for Investigating Nox Emissions, Transactions
of the ASAE, 48 (2005) 933-939.
[90] J. Szybist, D. Morris, A. Boehman, E. Esen, Diesel Fuel Formation Effects on
Injection Timing and Emissions, paper presented in American Chemistry. Society, 48
(2003) 428-429.
[91] J.P. Szybist, S.R. Kirby, A.L. Boehman, NOx Emissions of Alternative Diesel Fuels:
A Comparative Analysis of Biodiesel and Ft Diesel, Energy & Fuels, 19 (2005) 1484-
1492.
[92] C. Allen, K. Watts, Comparative Analysis of the Atomization Characteristics of
Fiften Biodiesel Fuel Types, Transactions of the ASAE 2000, 43 (2000) 207-211.
[93] T.L. Allenman, R.L. Mccormick, Fischer-Tropsch Diesel Fuels - Properties and
Exhaust Emissions: A Literature Review, SAE paper 2003-01-0763, (2003).
[94] Y. Icingur, D. Altiparmak, Effect of Fuel Cetane Number and Injection Pressure on a
DI Diesel Engine Performance and Emissions, Energy Conversion and Management, 44
(2003) 389-397.
209
[95] C.F.Taylor, Internal Combustion Engines, International Textbook Company,
Seranton, Pennsylvania, 1989.
[96] J.H. Song, K. Cheenkachorn, J.G. Wang, J. Perez, A.L. Boehman, P.J. Young, F.J.
Waller, Effect of Oxygenated Fuel on Combustion and Emissions in a Light-Duty Turbo
Diesel Engine, Energy & Fuels, 16 (2002) 294-301.
[97] J. Song, V. Zello, A.L. Boehman, F.J. Waller, Comparison of the Impact of Intake
Oxygen Enrichment and Fuel Oxygenation on Diesel Combustion and Emissions, Energy
& Fuels, 18 (2004) 1282-1290.
[98] S. Garner, R. Sivaramakrishnan, K. Brezinsky, The High-Pressure Pyrolysis of
Saturated and Unsaturated C-7 Hydrocarbons, Proceedings of the Combustion Institute,
32 (2009) 461-467.
[99] N. Nabi, Z. Shahadat, S. Rhaman, B. Alam, Behavior of Diesel Combustion and
Exhaust Emission with Neat Diesel Fuel and Diesel-Biodiesel Blends, SAE paper 2004-
01-3034, (2004).
[100] Y. Huang, S. Wang, L. Zhou, Effects of Fischer-Tropsch Diesel Fuel on
Combustion and Emissions of Direct Injection Diesel Engine, Frontiers of Energy and
Power Engineering in China, 2 (2008) 261-267.
[101] N. Miyamoto, H. Ogawa, M. Shibuya, K. Arai, O. Esmilaire, Influence of the
Molecular Structure of Hydrocarbon Fuels on Diesel Exhaust Emissions, SAE paper
940676, (1994).
[102] A. Monyem, J.H. Van Gerpen, M. Canakci, The Effect of Timing and Oxidation on
Emissions from Biodiesel-Fueled Engines, Transactions of the ASAE, 44 (2001) 35-42.
210
[103] A. Cheng, A. Upatnieks, C.J. Mueller, Investigation of the Impact of Biodiesel
Fueling on Nox Emissions Using an Optical Direct Injection Diesel Engine, International
Journal of Engine Research, (2006).
[104] W.A. Sirignano, Fuel Droplet Vaporization and Spray Combustion Theory,
Progress in Energy and Combustion Science, 9 (1983) 291-322.
[105] D.L. Siebers, Scaling Liquid-Phase Fuel Penetration in Diesel Sprays Based on
Mixing-Limited Vaporization, SAE paper 1999-01-0528, 108 (1999).
[106] J.D. Naber, D.L. Siebers, Effects of Gas Density and Vaporization on Penetration
and Dispersion of Diesel Sprays, SAE paper 960034, (1996).
[107] C.D. Rakopoulos, D.T. Hountalas, E.G. Giakoumis, E.C. Andritsakis. Performance
and emissions of bus engine using blends of diesel fuel with bio-diesel of sunflower or
cottonseed oils derived from Greek feedstocks. Fuel, (2007)
[108] P.J.M. Frijters, R.S.G. Baert, Oxygenated Fuels for Clean Heavy-Duty Diesel
Engines, International Journal of Vehicle Design, 41 (2006) 242-255.
[109] N. Ladommatos, K. Sison, H.W. Song, H. Zhao, Soot Generation of Diesel Fuels
with Substantial Amounts of Oxygen-Bearing Compounds Added, Fuel, 86 (2007) 345-
352.
[110] P. Flynn, R. Durrett, G. Hunter, A. Loye, O. Akinyemi, J.E. Dec, Diesel
Combustions: An Integrated View Combining Laser Diagnostics, Chemical Kinetics, and
Emprical Validation, SAE paper 1999-01-0509, (1999).
[111] J.E. Dec, A Conceptual Model of DI Diesel Combustion Based on Laser-Sheet
Imaging, SAE paper 970873, (1997).
211
[112] D. Chang, J.H. Van Gerpen, Fuel Properties and Engine Performance for Biodiesel
Prepared from Modified Feedstockes, SAE paper 971684, (1997).
[113] K. Schmidt, J.H. Van Gerpen, The Effect of Biodiesel Fuel Composition on Diesel
Combustion and Emissions, SAE paper 961086 (1996).
[114] G. Knothe, C. Sharp, T. Ryan, Exhaust Emissions of Biodiesel, Petrodiesel, Neat
Methyl Esters, and Alkanes in a New Technology Engine, Energy & Fuels, 20 (2006)
403-408.
[115] C.Y. Choi, G.R. Bower, R.D. Reitz, Effects of Biodiesel Blended Fuels and
Multiple Injections on D.I .Diesel Engine, SAE paper 970218, (1997).
[116] J. Song, M. Alam, A. Boehman, U. Kim, Examination of the Oxidation Behavior of
Biodiesel Soot, Combustion and Flame, 146 (2006) 589-604.
[117] K. Yehliu, R.L. Vander Wal, A.L. Boehman, Development of an HRTEM Image
Analysis Method to Quantify Carbon Nanostructure, Combustion and Flame, In Press,
Corrected Proof (2011).
[118] R.E. Franklin, Crystallite Growth in Graphitizing and Non-Graphitizing Carbons,
Proceedings of the Royal Society of London Series a-Mathematical and Physical
Sciences, 209 (1951).
[119] H. Marsh, Introduction to Carbon Science, Butterworths, London, (1989) 107-152.
[120] D.M. Smith, A.R. Chughtai, The Surface-Structure and Reactivity of Black Carbon,
Colloids and Surfaces A, Physicochemical and Engineering Aspects, 105 (1995) 47-77.
[121] A. Duran, J.M. Monteagudo, O. Armas, J.J. Hernandez, Scrubbing Effect on Diesel
Particulate Matter from Transesterified Waste Oils Blends, Fuel, 85 (2006) 923-928.
212
[122] U. Tl, S. Kb, M. Rl, Effects of Cetane Number, Cetane Improver, Aromatics and
Oxygenates on 1994 Heavy-Duty Diesel Engine Emissions, SAE paper 941020 (1994).
[123] M. Bunce, D. Snyder, G. Adi, C. Hall, J. Koehler, B. Davila, S. Kumar, P.
Garimella, D. Stanton, G. Shaver, Stock and Optimized Performance and Emissions with
5 and 20% Soy Biodiesel Blends in a Modern Common Rail Turbo-Diesel Engine,
Energy & Fuels, 24 (2010) 928-939.
[124] M. Bunce, D. Snyder, G. Adi, C. Hall, J. Koehler, B. Davila, S. Kumar, P.
Garimella, D. Stanton, G. Shaver, Optimization of Soy-Biodiesel Combustion in a
Modern Diesel Engine, Fuel, 90 (2011) 2560-2570.
[125] E. Mancaruso, S.S. Merola, B.M. Vaglieco, Study of the Multi-Injection
Combustion Process in a Transparent Direct Injection Common Rail Diesel Engine by
Means of Optical Techniques, International Journal of Engine Research, 9 (2008) 483-
498.
[126] W.L. Hardy, R.D. Reitz, An Experimental Investigation of Partially Premixed
Combustion Strategies Using Multiple Injections in a Heavy-Duty Diesel Engine, SAE
paper 2006-01-0917, (2006).
[127] M. Zheng, R. Kumar, Implementation of Multiple-Pulse Injection Strategies to
Enhance the Homogeneity for Simultaneous Low-Nox and -Soot Diesel Combustion,
International Journal of Thermal Sciences, 48 (2009) 1829-1841.
[128] Z. Han, A. Uludogan, G.J. Hampson, R.D. Reitz, Mechanism of Soot and Nox
Emission Reduction Using Multiple-Injection in a Diesel Engine, SAE paper 960633,
(1996).
213
[129] Y. Wang, C. Zhang, J. Jiang, A Numerical Study of the Effect of Multi-Injection
Strategy on Nox Reduction in DI Diesel Engines, HT2005: Proceedings of the ASME
Summer Heat Transfer Conference 2005, Vol 1, (2005) 759-765.
[130] X. Lu, Y. Ge, S. Wu, X. Han, An Experimental Investigation on Combustion and
Emissions Characteristics of Turbocharged DI Engines Fueled with Blends of Biodiesel,
SAE paper 2005-01-2199, presented at the SAE Fuels and Lubricants Meeting, Rio de
Janeiro, Brazil, (2005).
[131] V. Pradeep, R.P. Sharma, Evaluation of Performance, Emissions and Combustion
Parameters of a CI Engine Fuelled with Bio-Diesel from Rubber Seed Oil and Its Blends,
SAE paper 2005-26-353, (2005).
[132] D. Taylor, B. Walsham, Combustion Processes in a Medium-Speed Diesel Engine,
Proceedings of the Institution of Mechanical Engineers 1847-1996, 184 (1969) 67-76.
[133] R. Burt, K. Troth, Penetration and Vaporization of Diesel Fuel Sprays, Proceedings
of the Institution of Mechanical Engineers 1847-1996 184 (1969) 147-170.
[134] I.V. Roisman, L. Araneo, C. Tropea, Effect of Ambient Pressure on Penetration of
a Diesel Spray, International Journal of Multiphase Flow, 33 (2007) 904-920.
[135] Sauter, Die Grössenbestimmung Der in Gemischnebeln Von
Verbrennungskraftmaschinen Vorhandenen Brennstoffteilchen, VDI-Forschungsheft 279
(1926).
[136] H.K. Suh, H.G. Roh, C.S. Lee, Spray and Combustion Characteristics of
Biodiesel/Diesel Blended Fuel in a Direct Injection Common-Rail Diesel Engine, Journal
of Engineering for Gas Turbines and Power, Transactions of the ASME, 130 (2008) -.
214
[137] B.Y. Guo, D.F. Fletcher, T.a.G. Langrish, Simulation of the Agglomeration in a
Spray Using Lagrangian Particle Tracking, Applied Mathematics Modelling, 28 (2004)
273-290.
[138] T. Takagi, Y.F. Ching, T. Kamimoto, T. Okamoto, Numerical-Simulation of
Evaporation, Ignition and Combustion of Transient Sprays, Combustion Science &
Technology, 75 (1991) 1-12.
[139] G. Gouesbet, A. Berlemont, Eulerian and Lagrangian Approaches for Predicting
the Behaviour of Discrete Particles in Turbulent Flows, Progress in Energy and
[140] J. Reveillon, L. Vervisch, Spray Vaporization in Nonpremixed Turbulent
Combustion Modeling: A Single Droplet Model, Combustion and Flame, 121 (2000) 75-
90.
[141] X. Jiang, G.A. Siamas, K. Jagus, T.G. Karayiannis, Physical Modelling and
Advanced Simulations of Gas-Liquid Two-Phase Jet Flows in Atomization and Sprays,
Progress in Energy and Combustion Science, 36 (2010) 131-167.
[142] L.M. Pickett, D.L. Siebers, Soot Formation in Diesel Fuel Jets near the Lift-Off
Length, International Journal of Engine Research, 7 (2006) 103-130.
[143] L.M. Pickett, D.L. Siebers, An Investigation of Diesel Soot Formation Processes
Using Micro-Orifices, Proceedings of the Combustion Institute, 29 (2002) 655-662.
[144] L.M. Pickett, D.L. Siebers, Soot in Diesel Fuel Jets: Effects of Ambient
Temperature, Ambient Density, and Injection Pressure, Combustion and Flame, 138
(2004) 114-135.
215
[145] D.L. Siebers, B.S. Higgins, L.M. Pickett, Flame Lift-Off on Direct-Injection Diesel
Fuel Jets: Oxygen Concentration Effects, SAE paper 2002-01-0890, 111 (2002).
[146] W.A. Sirignano, Fluid-Dynamics of Sprays - 1992 Freeman Scholar Lecture,
Journal of Fluids Engineering, Transactions of the ASME, 115 (1993) 345-378.
[147] X. Lu, D. Han, Z. Huang, Fuel Design and Management for the Control of
Advanced Compression-Ignition Combustion Modes, Progress in Energy and
Combustion Science, In Press, Corrected Proof.
[148] Y. Takeda, K. Nakagome, K. Niimura, Emission Characteristics of Premixed Len
Diesel Combustion with Extremely Early Staged Fuel Injection, SAE paper 961163,
(1996).
[149] K. Nakagome, N. Shimazak, K. Niimura, Combustion and Emission Characteristics
of Premixed Lean Diesel Combustion Engine, SAE paper 970898, (1997).
[150] T. Miyamoto, H. Akagawa, K. Tsujimura, Combustion and Emission
Characteristics of Multiple-Stage Diesel Combustion, SAE paper 980505, (1998).
[151] Y. Nishijima, Y. Asaumi, Y. Aoyagi, Premixed Lean Diesel Combustion (Predic)
Using Impingenment Spray System, SAE paper 2001-01-1892, (2001).
[152] G.K. Lilik, J.M. Herreros, A.L. Boehman, Advanced Combustion Operation in a
Compression Ignition Engine, Energy & Fuels, 23 (2009) 143-150.
[153] K. Akihama, Y. Takatori, K. Inagaki, S. Sasaki, A. Dean, Mechanism of the
Smokeless Rich Diesel Combustion by Reducing Temperature, SAE paper 2001-01-0655,
(2001).
216
[154] P. Mark, B. Musculus, J.E. Dec, L.M. Pickett, C.A. Idicheria, In-Cylinder Imaging
of Conventional and Advanced, Low-Temperature Diesel Combustion, 11th diesel engine
emissions reduction conference, Chicago IL (2005).
[155] T. Fang, R. Coverdill, C. Lee, R. White, Low Temperature Combustion within a
Small Diesel Engine Bore High Speed Direct Injection (Hsdi), SAE paper 2005-01-0919,
(2005).
[156] G.C. Martin, C.J. Mueller, D.M. Milam, M.S. Radovanovic, C.R. Gehrke, Early
Direct-Injection, Low-Temperature Combustion of Diesel Fuel in an Optical Engine
Utilizing a 15-Hole Dual-Row, Anrrow-Included-Angle Nozzle, SAE paper 2008-01-
2400, (2008).
[157] T.G. Fang, R.E. Coverdill, C.F.F. Lee, R.A. While, Low-Sooting Combustion in a
Small-Bore High-Speed Direct-Injection Diesel Engine Using Narrow-Angle Injectors,
Proceedings of the Institution of Mechanical Engineers Part D-Journal of Automobile
Engineering, 222 (2008) 1927-1937.
[158] D. Choi, P.C. Miles, H. Yun, R.D. Reitz, A Parametric Study of Low-Temperature,
Late-Injection Combustion in a Hsdi Diesel Engine, JSME International Journal Series B-
Fluids and Thermal Engineering, 48 (2005) 656-664.
[159] T. Fang, C.-F.F. Lee, Low Sooting Combustion of Narrow-Angle Wall-Guided
Sprays in an Hsdi Diesel Engine with Retarded Injection Timings, Fuel, 90 (2011) 1449-
1456.
[160] T.G. Fang, C.F.F. Lee, Bio-Diesel Effects on Combustion Processes in an Hsdi
Diesel Engine Using Advanced Injection Strategies, Proceedings of the Combustion
Institute, 32 (2009) 2785-2792.
217
[161] R. Oinuma, S. Takuma, T. Koyano, M. Takiguchi, Effects of Post-Injection on
Piston Lubrication in a Common-Rail, Small-Bore Diesel Engine, SAE Trans. 2005-01-
2166, (2005).
[162] T. Fang, Y.C. Lin, T.M. Foong, C.F. Lee, Biodiesel Combustion in an Optical
HSDI Diesel Engine under Low Load Premixed Combustion Conditions, Fuel, 88 (2009)
2154-2162.
[163] M.J. Thornton, T.L. Alleman, J. Luecke, R.L. Mccormick, Impacts of Biodiesel
Fuel Blends Oil Dilution on Light-Duty Diesel Engine Operation, SAE paper 2009-01-
1790, (2009).
[164] M. Morcos, G. Parsons, W. Hartgers, M. Boons, S. Roby, F. Lauterwasser,
Detection Methods for Accurate Measurements of the Fame Biodiesel Content in Used
Crankcase Engine Oil, SAE paper 2009-01-2661, (2009).
[165] H.L. Fang, S.D. Whitacre, E.S. Yamaguchi, M. Boons, Biodiesel Impact on Wear
Protection of Engine Oils, SAE paper 2007-01-4141, (2007).
[166] K.M. Richard, S. Mctavish, Impact of Biodiesel on Lubricant Corrosion
Performance, SAE paper 2009-01-2660, (2009).
[167] K.O. Lee, J. Zhu, S. Ciatti, A. Yozgatligil, M.Y. Choi, Sizes, Graphitic Structures
and Fractal Geometry of Light-Duty Diesel Engine Particulates, SAE paper 2003-01-
3169, (2003).
[168] M. Lapuerta, R. Ballesteros, F.J. Martos, A Method to Determine the Fractal
Dimension of Diesel Soot Agglomerates, Journal of Colloid and Interface Science, 303
(2006) 149-158.
[169] B. Mandelbrot, The Fractal Geometry of Nature, W.H.Freeman and Co., 1982.
218
[170] M. Lapuerta, R. Ballesteros, F.J. Martos, The Effect of Diesel Engine Conditions
on the Size and Morphology of Soot Particles, International Journal of Vehicle Design,
50 (2009) 91-106.
[171] U.O. Koylu, G.M. Faeth, T.L. Farias, M.G. Carvalho, Fractal and Projected
Structure Properties of Soot Aggregates, Combustion and Flame, 100 (1995) 621-633.
[172] J.O. Muller, D.S. Su, R.E. Jentoft, J. Krohnert, F.C. Jentoft, R. Schlogl,
Morphology-Controlled Reactivity of Carbonaceous Materials Towards Oxidation,
Catalysis Today, 102 (2005) 259-265.
[173] A. Tsolakis, Effects on Particle Size Distribution from the Diesel Engine Operating
on Rme-Biodiesel with EGR, Energy and Fuels, 20 (2006) 1418-1424.
[174] A. Neer, U.O. Koylu, Effect of Operating Conditions on the Size, Morphology, and
Concentration of Submicrometer Particulates Emitted from a Diesel Engine, Combustion
and Flame, 146 (2006) 142-154.
[175] A. Soewono, S. Rogak, Morphology and Raman Spectra of Engine-Emitted
Particulates, Aerosol Science and Technology, 45 (2011) 1206-1216.
[176] U. Mathis, M. Mohr, R. Kaegi, A. Bertola, K. Boulouchos, Influence of Diesel
Engine Combustion Parametes on Primary Soot Particle Diameter, Environmental
Science & Technology, 39 (2005) 1887-1892.
[177] M. Lapuerta, F.J. Martos, J.M. Herreros, Effect of Engine Operating Conditions on
the Size of Primary Particles Composing Diesel Soot Agglomerates, Journal of Aerosol
Science, 38 (2007) 455-466.
219
[178] A. Smekens, R.H.M. Godoi, P. Berghmans, R.E. Van Grieken, Characterisation of
Soot Emitted by Domestic Heating, Aircraft and Cars Using Diesel or Biodiesel, Journal
of Atmospheric Chemistry, 52 (2005) 45-62.
[179] J. Zhu, K.O. Lee, A. Yozgatligil, M.Y. Choi, Effects of Engine Operating
Conditions on Morphology, Microstructure, and Fractal Geometry of Light-Duty Diesel
Engine Particulates, Proceedings of the Combustion Institute, 30 (2005) 2781-2789.
[180] M.M. Maricq, Physical and Chemical Comparison of Soot in Hydrocarbon and
Biodiesel Fuel Diffusion Flames: A Study of Model and Commercial Fuels, Combustion
and Flame, 158 (2011) 105-116.
[181] J.H. Song, M. Alam, A.L. Boehman, Impact of Alternative Fuels on Soot
Properties and DPF Regeneration, Combustion Science & Technology, 179 (2007) 1991-
2037.
[182] R.L. Vander Wal, C.J. Mueller, Initial Investigation of Effects of Fuel Oxygenation
on Nanostructure of Soot from a Direct-Injection Diesel Engine, Energy & Fuels, 20
(2006) 2364-2369.
[183] A. Braun, F.E. Huggins, N. Shah, Y. Chen, S. Wirick, S.B. Mun, C. Jacobsen, G.P.
Huffman, Advantages of Soft X-Ray Absorption over TEM-EELS for Solid Carbon
Studies - a Comparative Study on Diesel Soot with EELS and NEXAFS, Carbon, 43
(2005) 117-124.
[184] K. Al-Qurashi, A.L. Boehman, Impact of Exhaust Gas Recirculation (EGR) on the
Oxidative Reactivity of Diesel Engine Soot, Combustion and Flame, 155 (2008) 675-695.
[185] K. Yehliu, Impact of Fuel Formulation and Engine Operating Parameters on the
Nanostructure and Reactivity of Diesel Soot, in, The Pennsylvania State University, 2010.
220
[186] Z. Li, C.L. Song, J.O. Song, G. Lv, S.R. Dong, Z. Zhao, Evolution of the
Nanostructure, Fractal Dimension and Size of in-Cylinder Soot During Diesel
Combustion Process, Combustion and Flame, 158 (2011) 1624-1630.
[187] R.L. Vander Wal, A.J. Tomasek, Soot Nanostructure: Dependence Upon Synthesis
Conditions, Combustion and Flame, 136 (2004) 129-140.
[188] R.L. Vander Wal, A.J. Tomasek, Soot Oxidation: Dependence Upon Initial
Nanostructure, Combustion and Flame, 134 (2003) 1-9.
[189] H.J. Seong, A.L. Boehman, Impact of Intake Oxygen Enrichment on Oxidative
Reactivity and Properties of Diesel Soot, Energy & Fuels, 25 (2011) 602-616.
[190] H.J. Seong, Impact of Oxygen Enrichment on Soot Properties and Soot Oxidative
Reactivity, Ph.D Dissertation, The Pennsylvania State University, 2010.
[191] Ö.L. Gülder, Soot Formation in Laminar Diffusion Flames at Elevated
Temperatures, Combustion and Flame, 88 (1992) 75-82.
[192] B.M. Kumfer, S.A. Skeen, R. Chen, R.L. Axelbaum, Measurement and Analysis of
Soot Inception Limits of Oxygen-Enriched Coflow Flames, Combustion and Flame, 147
(2006) 233-242.
[193] D.X. Du, R.L. Axelbaum, C.K. Law, Soot Formation in Strained Diffusion Flames
with Gaseous Additives, Combustion and Flame, 102 (1995) 11-20.
[194] R.H. Hurt, G.P. Crawford, H.S. Shim, Equilibrium Nanostructure of Primary Soot
Particles, Proceedings of the Combustion Institute, 28 (2000) 2539-2546.
[195] H.X. Chen, R.A. Dobbins, Crystallogenesis of Particles Formed in Hydrocarbon
Combustion, Combustion Science & Technology, 159 (2000) 109-128.
221
[196] R.L. Vander Wal, A TEM Methodology for the Study of Soot Particle Structure,
Combustion Science & Technology, 126 (1997) 333-357.
[197] A. Senatore, M. Cardone, V. Rocco, M.V. Prati, A Comparative Analysis of
Combustion Process in DI Diesel Engine Fueled with Biodiesel and Diesel Fuel, SAE
paper 2000-01-0691, (2000).
[198] A. Shaheed, E. Swain, Combustion Analysis of Coconut Oil and Its Methyl Esters
in a Diesel Engine, IMECHE conference transactions, (1998).
[199] M. Canakci, Performance and Emissions Characteristics of Biodiesel from Soybean
Oil, IMECHE conference transactions, (2005).
[200] M. Lapuerta, J.R. Fernandez, J.R. Agudelo, Diesel Particulate Emissions from
Used Cooking Oil Biodiesel, Bioresource Technology, 99 (2008) 731-740.
[201] A. Monyem, G. Van, Jh, The Effect of Biodiesel Oxidation on Engine Performance
and Emissions, Biomass Bioenerg, 20 (2001) 317-325.
[202] M.N. Nabi, Theoretical Investigation of Engine Thermal Efficiency, Adiabatic
Flame Temperature, Nox Emission and Combustion-Related Parameters for Different
Oxygenated Fuels, Applied Thermal Engineering, 30 (2010) 839-844.
[203] C. Kaplan, R. Arslan, A.Surmen, Performance Characteristics of Sunflower Methyl
Esters as Biodiesel, Energy Sources, Part A 28 (2006) 751-755.
[204] T.J. Jacobs, J.A. Bittle, J.K. Younger, Biodiesel Effects on Influencing Parameters
of Brake Fuel Conversion Efficiency in a Medium Duty Diesel Engine, Journal of
Engineering for Gas Turbines and Power, Transactions of the ASME, 132 (2010).
222
[205] L. Gvidonas, S. Stasys, The Effect of Rapeseed Oil Methyl Easter on Direct
Injection Diesel Engine Performance and Exhaust Emissions., Energy Convers Manage,
47 (2006) 1954-1967.
[206] A. Kowalewicz, Eco-Diesel Engine Fuelled with Rapeseed Oil Methyl Ester and
Ethanol. Part 2: Comparison of Emissions and Efficiency for Two Base Fuels: Diesel
Fuel and Ester, Proceedings of the Institution of Mechanical Engineers Part D, Journal of
Automobile Engineering, 220 (2006) 1275-1282.
[207] Y. Aoyagi, E. Kunishima, Y. Asaumi, Y. Aihara, M. Odaka, Y. Goto, Diesel
Combustion and Emission Using High Boost and High Injection Pressure in a Single
Cylinder Engine - (Effects of Boost Pressure and Timing Retardation on Thermal
Efficiency and Exhaust Emissions), JSME International Journal Series B, Fluids and
Thermal Engineering, 48 (2005) 648-655.
[208] A. Ramesh, J.N. Reddy, Parametric Studies for Improving the Performance of a
Jatropha Oil-Fuelled Compression Ignition Engine, Renewable Energy, 31 (2006) 1994-
2016.
[209] T. Shudo, Y. Nakajima, T. Futakuchi, Thermal Efficiency Analysis in a Hydrogen
Premixed Combustion Engine, JSAE Review, 21 (2000) 177-182.
[210] K. Nagase, Y. Fukushima, A Fuel-Injection System for Diesel-Engines by Injection
Pressure Control .1. Theories, Influence Upon Theoretical Thermal Efficiency and the
Result of Basic Experiments, Bulletin of the Japan Society of Mechanical Engineers, 29
(1986) 1788-1794.
223
[211] M. Zheng, U. Asad, G.T. Reader, Y.Y. Tan, M.P. Wang, Energy Efficiency
Improvement Strategies for a Diesel Engine in Low-Temperature Combustion,
International Journal of Energy Resourse, 33 (2009) 8-28.
[212] T. Fang, Y.C. Lin, T.M. Foong, C.F. Lee, Reducing Nox Emissions from a
Biodiesel-Fueled Engine by Use of Low-Temperature Combustion, Environmental
Science & Technology, 42 (2008) 8865-8870.
[213] H. Hiroyasu, T. Kadota, M. Arai, Development and Use of a Spray Combustion
Modeling to Predict Diesel-Engine Efficiency and Pollutant Emissions .1. Combustion
Modeling, Bulletin of the Japan Society of Mechanical Engineers, 26 (1983) 569-575.
[214] H. Hiroyasu, T. Kadota, M. Arai, Development and Use of a Spray Combustion
Modeling to Predict Diesel-Engine Efficiency and Pollutant Emissions .2.
Computational-Procedure and Parametric Study, Bulletin of the Japan Society of
Mechanical Engineers, 26 (1983) 576-583.
[215] H. Hiroyasu, O. Furukawa, M. Arai, S. Iida, H. Motonaga, Development and Use
of a Spray Combustion Modeling to Predict Diesel-Engine Efficiency and Pollutant
Emissions .3. An Analysis by the Method of System-Model Transformation, Bulletin of
the Japan Society of Mechanical Engineers, 26 (1983) 584-591.
[216] T.J. Williams, N.D. Whitehouse, Investigation into Some Aspects of the
Computation of Diesel Engine Combustion, Proceedings of the institution of Mechanical
Engineers 190 (1976) 467 - 476.
[217] NIST, Chemistry Web Boook, http://webbook.nist.gov/chemistry, (2011).
[218] Y. Zhang, Y. Yang, A.L. Boehman, Premixed Ignition Behavior of C-9 Fatty Acid
Esters: A Motored Engine Study, Combustion and Flame, 156 (2009) 1202-1213.
224
[219] A.C. Ferrari, J. Robertson, Interpretation of Raman Spectra of Disordered and
Amorphous Carbon, Physical Review B, 61 (2000) 14095-14107.
[220] T. Jawhari, A. Roid, J. Casado, Raman-Spectroscopic Characterization of Some
Commercially Available Carbon-Black Materials, Carbon, 33 (1995) 1561-1565.
[221] C.S. Lee, S.W. Park, An Experimental and Numerical Study on Fuel Atomization
Characteristics of High-Pressure Diesel Injection Sprays, Fuel, 81 (2002) 2417-2423.
[222] C.W. Yu, S. Bari, A. Ameen, A Comparison of Combustion Characteristics of
Waste Cooking Oil with Diesel as Fuel in a Direct Injection Diesel Engine, Proceedings
of the Institution of Mechanical Engineers Part D, Journal of Automobile Engineering,
216 (2002) 237-243.
[223] M. Munoz, F. Moreno, J. Morea, Emissions of an Automobile Diesel Engine
Fueled with Sunflower Methylester, Transactions of the ASAE, 47 (2004) 5-11.
[224] U.S. DOE EERE, Properties of Fuels, online resources,
http://www.afdc.energy.gov/afdc/pdfs/fueltable.pdf, (2009).
[225] Argonne National Laboratory, Greet Transportation Fuel Cycle Analysis Model,
Greet 1.8, http://www.transportation.anl.gov/modeling_simulation/GREET/index.html,
(2008).
[226] C.A. Idicheria, L.M. Pickett, Quantitative Mixing Measurements in a Vaporizing
Diesel Spray by Rayleigh Imaging, SAE paper 2007-01-0647, (2007).
[227] D.L. Siebers, B.S. Higgins, Flame Lift-Off on Direct Injection Diesel under
Quiescent Conditions, SAE paper 2001-01-0530, (2001).
[228] H.J. Curran, P. Gaffuri, W.J. Pitz, C.K. Westbrook, A Comprehensive Modeling
Study of N-Heptane Oxidation, Combustion and Flame, 114 (1998) 149-177.
225
[229] J.P. Szybist, A.L. Boehman, D.C. Haworth, H. Koga, Premixed Ignition Behavior
of Alternative Diesel Fuel-Relevant Compounds in a Motored Engine Experiment,
Combustion and Flame, 149 (2007) 112-128.
[230] D. Assanis, Z. Filipi, S. Fiveland, M. Syrimis, A Predictive Ignition Delay
Correlation under Steady-State and Transient Operation of a Direct Injection Diesel
Engine, Journal of Engineering for Gas Turbines and Power, transactions of the ASME,
125 (2003) 450-457.
[231] M. Sjoberg, J.E. Dec, An Investigation into Lowest Acceptable Combustion
Temperatures for Hydrocarbon Fuels in HCCI Engines, in: 30th International
Symposium on Combustion, Chicago, IL, 2004, pp. 2719-2726.
[232] S.I. Sandler, Chemical, Biochemical, and Engineering Thermodynamics, 4e, John
Wiley & Sons, Inc, U.S.A., (2006) 319.
[233] W. Yuan, A.C. Hansen, Q. Zhang, Vapor Pressure and Normal Boiling Point
Predictions for Pure Methyl Esters and Biodiesel Fuels, Fuel, 84 (2005) 943-950.
[234] J.W. Goodrum, Volatility and Boiling Points of Biodiesel from Vegetable Oils and
Tallow, Biomass & Bioenergy, 22 (2002) 205-211.
[235] X.Y. Zou, J.M. Shaw, Phase Behavior of Hydrocarbon Mixtures, in: Encyclopedia
of Chemical Processing, Taylor & Francis, 2006, pp. 2067-2075.
[236] K.S. Sanvordenker, Engine Friction-Effect of Oil Viscosity and Break-in on
Cranking Resistance, SAE paper 841095, (1984).
[237] P. Meakin, B. Donn, G.W. Mulholland, Collisions between Point Masses and
Fractal Aggregates, Langmuir, 5 (1989) 510-518.
226
[238] E.F. Mikhailov, S.S. Vlasenko, A.A. Kiselev, T.I. Ryshkevitch, Restructuring of
Soot Particles: Experimental Study, Journal of Aerosol Science, 27 (1996) 771-712.
[239] J.H. Song, Effect of Fuel Formulation on Soot Properties and Regeneration of
Diesel Particulate Filters, in: Fuel Science, The Pennsylvania State University, 2005.
[240] L.R. Radovic, P.L. Walker, R.G. Jenkins, Importance of Carbon Active-Sites in the
Gasification of Coal Chars, Fuel, 62 (1983) 849-856.
[241] R.G. Jenkins, S.P. Nandi, P.L. Walker, Reactivity of Heat-Treated Coals in Air at
500 Degrees C, Fuel, 52 (1973) 288-293.
[242] K. Al-Qurashi, A.D. Lueking, A.L. Boehman, The Deconvolution of the Thermal,
Dilution, and Chemical Effects of Exhaust Gas Recirculation (EGR) on the Reactivity of
Engine and Flame Soot, Combustion and Flame, 158 (2011) 1696-1704.
[243] A. Arenillas, F. Rubiera, C. Pevida, C.O. Ania, J.J. Pis, Relationship between
Structure and Reactivity of Carbonaceous Materials, Journal of Thermal Analysis and
Calorimetry, 76 (2004) 593-602.
[244] A. Cheng, P. Harriott, Kinetics of Oxidation and Chemisorption of Oxygen for
Porous Carbons with High Surface-Area, Carbon, 24 (1986) 143-150.
[245] P. Gilot, F. Bonnefoy, F. Marcuccilli, G. Prado, Determination of Kinetic Data for
Soot Oxidation - Modeling of Competition between Oxygen Diffusion and Reaction
During Thermogravimetric Analysis, Combustion and Flame, 95 (1993) 87-100.
[246] A.L. Boehman, J. Song, C.H. Jeon, Impacts of Oxygen Diffusion on the
Combustion Rate of in-Bed Soot Particles, Energy & Fuels, 24 (2010) 2418-2428.
[247] N.R. Laine, F.J. Vastola, P.L. Walker, The Importance of Active Surface Area in
Carbon Oxygen Reaction, Journal of Physical Chemistry, 67 (1963) 1461-1475.
227
[248] A. Agudelo, J.R. Agudelo, P. Benjumea, Diagnóstico De La Combustión De
Biocombustibles En Motores (Combustion Diagnosis of Biofuels in Engines), Imprenta
Universidad de Antioquia, 2007.
[249] G. Woschni, Universally Applicable Equation for the Instantaneous Heat Transfer
Coefficient in the Internal Combustion Engine, SAE paper 670930, (1967).
[250] B.S. Han, Y.J. Chung, Y.J. Kwon, S. Lee, Empirical Forrmula for Instantaneous
Heat Transfer Coefficient in Spark Ignition Engines, SAE paper 972995, (1997).
[251] O. Armas, Diagnóstico Experimental Del Proceso De Combustión En Motores
Diesel De Inyección Directa., in, Universidad Politécnica de Valencia, España, 1988.
[252] C.D. Rakopoulos, D.T. Hountalas, A.P. Koutroubousis, T.C. Zannis, Application
and Evaluation of Detailed Friction Model on a Di Diesel Engine with Extreme High
Peak Combustion Pressure, SAE paper 2002-01-0068, (2002).
[253] M. Lapuerta, O. Armas, S. Molina, Study of the Compression Cycle of a
Reciprocating Engine through the Polytropic Coefficient, Applied Thermal Engineering,
23 (2003) 313-323.
[254] [254] W.T. Lyn, Calculations of the Effect of Rate of Heat Release on the Shape of
Cylinder-pressure Diagram and Cycle Efficiency, Proceedings of the Institution of
Mechanical Engineers, (1960).
[255] S.H. Yoon, C.S. Lee, Lean Combustion and Emission Characteristics of Bioethanol
and Its Blends in a Spark Ignition (SI) Engine, Energy & Fuels, 25 (2011) 3484-3492.
228
Appendix A: Matlab code
Oxygen equivalence ratio field calculation, LLP.m
clear all;
nfC = 16.15; % fuel carbon atom number, averaged, diesel 14.48, B40 16.15
**************************************************Biodiesel
nfH = 30.38; % fuel hydrogen atom number, averaged, diesel 26.23; B40
30.38***************************************************Biodiesel
nfO = 0.8;% fuel oxygen atom number, averaged,diesel 0, B40
0.8;***************************************************Biodiesel
MWf = nfC*12+nfH*1+nfO*16; % averaged molecular weight (g/mol) of fuel

MWa = 28.97; % averaged molecular weight (g/mol) of air
Tg = 1172; % cylinder gas temperature at injection (K) 355 CAD

***********************************************
Pg = 112; % cylinder pressure at injection (Bar) 355 CAD
**************************************
VI = 49.8; % cylinder volume (cm3) 51.05 at injection 355 CAD, 52.93 at 353, 55.43 at 351,
49.80 at 3B, 49.17 at 1B ***************************,
Ru = 8.314472; % gas constant (J/K/mol)

rhoa_1 = 1.14; % density of air (kg/m3) at 1 bar at ~40 C
Pg_180 = 2.3; % boost pressure (bar), or cylinder pressure at 180 CAD
Pf = input('input Pf = '); % fuel injection pressure (bar)

***************************************************
Tbst = 60; % boost temperature (C)

rhoa_180 = Pg_180*rhoa_1*(40+273)/(Tbst+273); % density (kg/m3)of air at 180 CAD
Vb = 844.34; % volume (cm3) at bottom dead center
rhoa_I = rhoa_180*Vb/VI; % density of air (kg/m3) at injection
rhog = rhoa_I; % assume the density by fuel injection is negelectable, density of environmental
gas at injection
rhof = 855; %$ density of fuel (kg/m3), ULSD is 842, B100 is 877, B40 is 855
******************************************Biodiesel
dn = 0.111; % diameter of orifice (mm)

Yf_1 = 1; % fuel concentration at injector
XO2_2 = 0.21; % oxygen volume concentration at injector, 0.21 is the concentration of oxygen in
air
YO2_2 = 0.232; % oxygen mass concentration at injector
nf_st = 1; % fuel at stoichiometrix for nf_st*fuel + mO2_st*O2 = 0
mO2_st = nf_st*(nfC+nfH/4-nfO/2); % oxygen number at stoichiometric
mu = mO2_st*32/nf_st/MWf; % mu*Yf + YO2 = 0, reaction parameter
mf_st = nf_st*MWf; % mass of fuel at stoichiometric (g)
ma_st = mO2_st*32/YO2_2; % mass of air at stoichiemetric (g)
AFst = ma_st/mf_st; % stoichiometric air/fuel ratio
Zst = (1+mu*Yf_1/YO2_2)^-1; % stoichiometric mixture fraction
Ca = 0.91+(0.86-0.91)/(1380-720)*(Pf-720); % area contraction coeffecient Ca =0.82 for dn =
0.18, Ca =0.86 for dn = 0.1, 0.071, 0.05 mm @ 1400bar injection
229
Cd = 0.80; % discharged coefficient
Cv = Cd/Ca; % Bernolli's equation coefficient
Uf = Cv*sqrt(2*(Pf-Pg)*10^5/rhof); % velocity of injected fuel at orifice (m/s), Bernoulli's
equation, pressure is transferred to Pa, SI unit
c = 0.255; % speading angle coefficient interpolated from Siebers 1999 SAE
a = 0.75; % Constant from Siebers 1996 SAE
Tan_half_theta = c*((rhog/rhof)^0.19-0.0043*sqrt(rhof/rhog));% spray half angle
x_plus = sqrt(rhof/rhog)*dn*sqrt(Ca)/a/Tan_half_theta; % penetration length (mm), same unit as
dn
H = 7.04*10^8*Tg^-3.74*rhog^-0.85*(dn*10^3)^0.34*Uf*Zst; % calculate lift-off length, also SI
unit, except from that unit of dn is "um"
x = [x_plus:2:50 1000*H]; % buliding matrix for x
phi_x = 2.*AFst./(sqrt(1+16.*(x./x_plus).^2)-1);% equivalence ratio as f(x)
phi_omega_f_B40 = (2*nfC+0.5*nfH)/nfO; % oxygen equavalence ratio of the fuel alone, only
for biodiesel, here is B40
phi_x_omega_B40 = phi_x./(1-(1-phi_x)/phi_omega_f_B40); % oxygen equivalence ratio matrix
for B40
phi_x_omega_ULSD = phi_x; % oxygen equivalence ratio matrix for ULSD
r = -3:0.5:3; % building spray diameter matrix (mm)
num_initial = 1; %define the initial index
num = num_initial; % initial the index
numr = size(r); % count the number of row & column of r
numx = size(x); % count the number of row & column of x
numrr = numr(1,2); % total element in r
numxx = numx(1,2); % total element in x
phi_omega_x_r_ULSD = zeros(numrr, numxx); % initialized the phi_omega_x_r_ULSD
phi_omega_x_r_B40 = zeros(numrr,numxx); % initialized the phi_omega_x_r_B40
while num<=numrr
phi_omega_x_sr = 1.3.*phi_x_omega_ULSD.*exp((r(1,num)./x).^2*log(0.08)/Tan_half_theta^2); %
building oxygen equivalence ratio matrix based on f(x,r)
phi_omega_x_r_ULSD(num,:) = phi_omega_x_sr; % replace row of num of phi_omega_x_r with
data
num = num +1; %index num
end
num = num_initial; % reinitialize num index
while num<=numrr
phi_omega_x_sr = 1.3.*phi_x_omega_B40.*exp((r(1,num)./x).^2*log(0.08)/Tan_half_theta^2); %
building oxygen equivalence ratio matrix based on f(x,r)
phi_omega_x_r_B40(num,:) = phi_omega_x_sr; % replace row of num of phi_omega_x_r with
data
num = num +1; %index num
end
disp 'lift-off length';
disp(H);
Efficiency simulation, efficiency.m
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% efficiency calculation for limited pressure cycle
% clear all;
% SOCtest = 0;
% C1 = 1; %control the ratio of premixed and diffusion burn
% C2 = 1;
230
gamma = 1.367;
alpha = [1:0.1:2]';
beta = 1:10;
rc = 17.2;
eta_ideal = alpha*beta; % initialize eta_otto matrix, efficiency in ideal cycle
for j = 1:size(alpha,1)
for i = 1:size(beta,2)
eta_ideal(j,i) = 1 - 1/rc^(gamma-1)*(alpha(j)*beta(i)^gamma-1)/(alpha(j)*gamma*(beta(i)-
1)+alpha(j)-1);
end
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% this model assumes no leakage, but it can be easily added
theta = 0:0.1:719.9; % crank angle

bore = 9.82; % cm
Stroke = 10.5; % cm
CR = 17.2; % compression ratio
Vc = 49.089340; %cm3
lcr = 17.6; % cm, connect rod length
rc = 5.25; % cm, crank radius
rpm = 1500; % speed
omega = rpm/60*360; % dtheta/dt, degree per second
ncy = 8; % number of cylinder
dtheta = 0.1; % delta theta
V = Vc + pi*bore^2/4*(lcr + rc - (rc.*cos(theta*pi/180)+(lcr^2-rc^2.*sin(theta*pi/180).^2).^0.5));
% chamber volume with theta

A = pi*bore*(lcr + rc - (rc.*cos(theta*pi/180)+(lcr^2-rc^2.*sin(theta*pi/180).^2).^0.5)); %
instantaneous cylinder area, cm2

Ru = 8.314472; % gas constant, J/mol/K
IVC = 200; % timing of inlet valve close
EVO = 520; % timing of exhaust valve open, CA
Y = 0.7; % 0.027 is the mf/(mf+ma) value, 0.7 seems to give better match, but no sense in
temperature
pintake = 2.52835; % intake pressure,

bar %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
pexhaust = 1.5; % exhaust pressure, here we first assume no pumping loss
Tintake = 40 + 273; % intake temperature, K
gamma = fgamma(Tintake).*theta./theta; % initialize a gamma

qw = 0.*theta;
231
Sp = 2*Stroke*0.01*rpm/60; % average piston velocity, m/s
wmotor = 0.*theta; % the Woschni velocity, initialize, m/s, only for motor
for i = 1:size(theta,2)
if theta(i) < IVC
wmotor(i) = 6.18*Sp;
elseif theta(i) == IVC
n_IVC = i; % record the index of inlet valve close
% elseif theta(i) <360
% wmotor(i) = 2.28*Sp;
elseif theta(i) < EVO
elseif theta(i) == EVO
n_EVO = i; % index of exhaust value close
else
end
end
wcomb = wmotor; % initialize a combution w factor
%corr = 1.18/2.3789; % a correction for the MAF (don't have justification

%yet), for 1801
corr = 0.0407/0.0805; % for IVC (0.0671)
nintake = pintake*10^5*V(n_IVC)*10^-6/Ru/Tintake*corr;
Mwair = 28.97; % g/mol
Tw = 90+273; % cylinder wall temperature, K, coolant temp
MAF = Mwair*nintake; % g/cycle
if i <= n_IVC
pmotor(i) = pintake;
Tmotor(i) = Tintake;
elseif i<= n_EVO
% pmotor(i) = pmotor(i-1)*V(i-1)^gamma(i)/V(i)^gamma(i);
% Tmotor(i) = pmotor(i)*10^5*V(i)*10^-6/Ru/nintake;
% gamma(i) = fgamma(Tmotor(i));
% while abs(pmotor(i-1)*V(i-1)^gamma(i)/V(i)^gamma(i) - pmotor(i)) > 10^-3
% pmotor(i) = pmotor(i-1)*V(i-1)^gamma(i)/V(i)^gamma(i);
% Tmotor(i) = pmotor(i)*10^5*V(i)*10^-6/Ru/nintake;
% gamma(i) = fgamma(Tmotor(i));
% end % (adiabatic) process
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
pinter = pmotor(i-1);
Tmotor(i) = pmotor(i-1)*10^5*V(i)*10^-6/Ru/nintake;
% Tmotor(i) = ftemp(pmotor(i-1),V(i),MAF);
gamma(i) = fgamma(Tmotor(i));
pmotor(i) = (V(i)*pmotor(i-1) - 10*dtheta/omega*(gamma(i)-1)*dtheta*fconvection2
(pinter,Tmotor(i),wmotor(i),A(i),0))/(gamma(i)*(V(i)-V(i-1))+V(i));
while abs(pmotor(i) - (V(i)*pmotor(i-1) - 10*dtheta/omega*(gamma(i)-1)
232
*dtheta*fconvection2(pinter,Tmotor(i),wmotor(i),A(i),0))/(gamma(i)*(V(i)-V(i-1))+V(i))) >10^-3
Tmotor(i) = pmotor(i)*10^5*V(i)*10^-6/Ru/nintake;
% Tmotor(i) = ftemp(pmotor(i),V(i),MAF);
pmotor(i) = (V(i)*pmotor(i-1) - 10*dtheta/omega*(gamma(i)-1)*dtheta*fconvection2
(pinter,Tmotor(i),wmotor(i),A(i),0))/(gamma(i)*(V(i)-V(i-1))+V(i));
pinter = pmotor(i);
end % with heat transfer
else
pmotor(i) = pmotor(n_EVO);
Tmotor(i) = Tmotor(n_EVO);
end
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%
% synchronize TDC with the maximum cylinder pressure
pmotor_old = pmotor; % record pmotor before a correction of TDC with maximum cylinder
pressure
Tmotor_old = Tmotor;
dpmotor = theta./theta; % initialized dqmotor
for i = 3400:3800
dpmotor(i) = (pmotor(i+1)-pmotor(i-1))/(theta(i+1)-theta(i-1));
end
for i = 3400:3800
if abs(dpmotor(i)) == min(abs(dpmotor))
n_TDCcorr = i; % the index the TDC correction
end
end
delta_n = 3601 - n_TDCcorr; % the correction
for i = delta_n:(size(theta,2) - delta_n)

pmotor(i) = pmotor_old(i-delta_n+1);
Tmotor(i) = Tmotor_old(i-delta_n+1);
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% arbitrary heat release
% for Gaussian peak
lgauss = 365.18 + SOCtest; % location, SOC

wgauss = 1.1683*1;
hgauss = 125.87*1;
% for ExpConvExp test

lece = 363.32 + SOCtest; % location, crank angle, SOC
hece = 619.69; % height, J/deg
233
k1 = 0.12822;
k2 = 0.77745;
ng = 1; % number of gaussian heat release

nece = 1; % recording the heat release period, number of ExpConvExp heat release
HRCAgauss = 0.*theta;
HRCAece = HRCAgauss;
HRCAgauss_index = HRCAgauss;
HRCAece_index = HRCAece;
fece(i) = hece/k2*fexpconvexp(theta(i) - lece ,k1,k2); % result of ExpConvExp
fgau(i) = hgauss*fgauss(theta(i),lgauss,wgauss); %result of gauss
if fgau(i) >= 0.1
HRCAgauss(i) = theta(i);
HRCAgauss_index(i) = 1;
ng = ng+1;
end
if fece(i) > 0.1
HRCAece(i) = theta(i);
HRCAece_index(i)=1;
nece = nece + 1;
end
end
AHR = C1*fgau + C2*fece; % total apparent heat release, J/Deg.

AHR_index = 1*HRCAgauss_index+C2*HRCAece_index;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%
% calculate teh combustion
pcomb = pmotor; % initialize a pressure with combustion

Tcomb = Tmotor; % initialize a temperature with combustion
if AHR_index(i) > 0
Tcomb(i) = pcomb(i-1)*10^5*V(i)*10^-6/Ru/nintake;
gamma(i) = fgamma(Tcomb(i));
pcomb(i) = (V(i)*pcomb(i-1) + 10*AHR(i)*(gamma(i)-1)*dtheta)/(gamma(i)*(V(i)-V(i-
1))+V
(i)); % multiply 10 for unit conversion

while abs(pcomb(i) - ((V(i)*pcomb(i-1) + 10*AHR(i)*(gamma(i)-
1)*dtheta)/(gamma(i)*(V(i)-V
(i-1))+V(i)))) >10^-5
234
pcomb(i) = (V(i)*pcomb(i-1) + 10*AHR(i)*(gamma(i)-1)*dtheta)/(gamma(i)*(V(i)-V(i-
1))+V(i));
Tcomb(i) = pcomb(i)*10^5*V(i)*10^-6/Ru/nintake;
end % to obtain the right gamma value
wcomb(i) = wmotor(i) + 3.24*10^-
3*Stroke*pi*bore^2/4/V(n_IVC)*Tintake/pintake*(pcomb(i)-
pmotor(i));
elseif theta(i) >= lgauss
if i<=n_EVO
wcomb(i) = wmotor(i) + 3.24*10^-
3*Stroke*pi*bore^2/4/V(n_IVC)*Tintake/pintake*(pcomb
(i-1)-pmotor(i));
pinter = pcomb(i-1);
Tcomb(i) = pcomb(i-1)*10^5*V(i)*10^-6/Ru/nintake;
pcomb(i) = (V(i)*pcomb(i-1) - 10*dtheta/omega*(gamma(i)-1)*dtheta*fconvection2
(pinter,Tcomb(i),wcomb(i),A(i),Y))/(gamma(i)*(V(i)-V(i-1))+V(i));
while abs(pcomb(i) - (V(i)*pcomb(i-1) - 10*dtheta/omega*(gamma(i)-1)
*dtheta*fconvection2(pinter,Tcomb(i),wcomb(i),A(i),Y))/(gamma(i)*(V(i)-V(i-1))+V(i))) >10^-3
Tcomb(i) = pmotor(i)*10^5*V(i)*10^-6/Ru/nintake;
wcomb(i) = wmotor(i) + 3.24*10^-3*Stroke*pi*bore^2/4/V(n_IVC)*Tintake/pintake*
(pcomb(i)-pmotor(i));
pcomb(i) = (V(i)*pcomb(i-1) - 10*dtheta/omega*(gamma(i)-1)*dtheta*fconvection2
(pinter,Tcomb(i),wcomb(i),A(i),Y))/(gamma(i)*(V(i)-V(i-1))+V(i));
pinter = pcomb(i);
end % with heat transfer
end
end
if i > n_EVO
pcomb(i) = pcomb(n_EVO);
Tcomb(i) = Tcomb(n_EVO-1);
end
end
qt = 0.*theta; % initialize a total energy addtion (essentially dQ/dtheta)

for i = 1: size(theta,2)
if AHR_index(i)>0
qt(i) = AHR(i) + dtheta/omega*fconvection2(pcomb(i),Tcomb(i),wcomb(i),A(i),0.027)/10; %
total energy addition, devide by 10 to convert the unit to J/deg.

end
end
CHR = 0; % cumulative heat release, J

tQ = 0; % total heat addition
235
if AHR_index(i)>0
CHR = CHR + (AHR(i)+AHR(i+1))/2*dtheta;
tQ = tQ + (qt(i)+qt(i+1))/2*dtheta;
end
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% indicated power
dV = 0.*V; % initialize a V, dV/dtheta

dV(1) = (V(2)-V(1))/dtheta;
dV(size(theta,2)) = (V(size(theta,2))-V(size(theta,2)-1))/dtheta;
for i = 2:size(theta,2)-1
dV(i) = (V(i+1)-V(i-1))/2/dtheta;
end
% pengine = xlsread('p_100MPa_25%'); % result engine pressure trace data, 25%, 1B, 100MPa,
CA and
pressure(bar)
indiQ = 38.6392; % a correction for the motor difference

indiQ_engine = 0;
for i = n_IVC:n_EVO
indiQ = indiQ + (pcomb(i)+pcomb(i-1))/2*(V(i)-V(i-1))*0.1;
% indiQ_engine = indiQ_engine + (pengine(i,2)+pengine(i-1,2))/2*(V(i)-V(i-1))*0.1;
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%
% calculate friction torque
% also friction power
frictiontorque2;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%
% efficiency
eta_indi = indiQ/tQ; % indicated efficiency
eta_brake = (indiQ - qf)/tQ; % brake efficiency
Appendix B: Repeatability of thermogravimetric analysis
236
This section concerns the repeatability of the isothermal oxidation tests. Since
each sample was conducted repeatability two times, here we only present one example of
the repeatability of isothermal test.
The presented example is an isothermal oxidation test of diesel soot from 30%
load and 75MPa fuel injection pressure. Both test followed exactly the same procedure:
• Gas environment is N2
• Initialize temperature to 30ºC
• Ramp at a rate of 10ºC/min to 500 ºC
• Isotherm 30 min
• Ramp at a rate of 5ºC/min to 550 ºC
• Switch gas to air
• Isotherm 60ºC
Figure B-1 actually only present the isothermal test in air (the last step), which
can be used for oxidative reactivity analysis. One can observe that the repeatability of
isothermal test is satisfying.
237
1.0
test 1
test 2
0.8
0.6
m/m0
0.4
0.2
0.0
0 10 20 30 40 50 60
time (m)
Figure B-1: Repeatability of thermogravimetric analysis for diesel soot, 30% load
and 75 MPa
238
Appendix C: Repeatability of Raman analysis
This section concerns the repeatability of Raman analysis. Raman spectrum was
obtained for random five locations of the sample. The 3L + 1G curve fitting techniques
were conducted individually on each spectrum. An Example of the result of AD1/AG and
D1 FWHM was reported here.
The example is B20 soot, 30% load, 50 MPa fuel injection pressure. It can be
observed that the repeatability of Raman analysis is satisfying.
Table C-1: Repeatability of Raman analysis for B20 soot, 30% load and 50 MPa fuel
injection pressure
Test number AD1/AG D1 FWHM (cm-1)

1 2.24 174.07
2 2.28 175.56
3 2.24 178.09
4 2.34 174.94
5 2.35 172.85
Average 2.29 175.10
Standard derivation 0.05 1.46
239
Appendix D: Repeatability of Morphology analysis
The morphology analysis requires more than 25 images of the transmission
electronic microscopy. The primary particle size and fractal dimension of each images
are measured and averaged value is reported. Again, one example (Table D-1) of
repeatability test is presented here.
Table D-1: Repeatability of number of primary particles and fractal dimension for
B20 soot, 60% load at 75 MPa fuel injection pressure
Image number Number of primary particles Fractal dimension

1 100.17 1.90
2 117.17 2.18
3 113.18 1.83
4 141.87 2.16
5 105.10 1.85
6 134.48 1.91
7 112.58 1.96
8 106.56 1.98
9 105.31 2.03
10 101.37 1.97
11 98.68 1.83
12 102.86 2.05
13 106.55 1.98
14 100.23 1.87
15 133.12 2.22
16 209.89 2.00
17 114.52 1.91
18 162.23 1.97
19 140.36 2.13
20 109.02 1.83
21 104.10 1.87
22 107.93 2.08
23 116.33 2.15
24 131.97 1.86
25 108.92 1.88
26 108.03 2.02
Average 118.94 1.98
Standard derivation 24.39 0.12
240
The repeatability of measurement of diameter of primary particles is demonstrated
with diesel soot, 60% load at 125 MPa fuel injection pressure, as show in Table D-2:
Table D-2: repeatability of measurement of diameter of primary particles of diesel

soot, 60% load at 125 MPa fuel injection pressure
Primary particle number Diameter of primary particle (nm)

1 15.10
2 26.21
3 25.71
4 21.90
5 28.98
6 19.86
7 26.15
8 34.84
9 22.05
10 24.20
Average 24.50
Standard derivation 5.07
241
Appendix E: Repeatability of particle size distribution (SMPS)
result
Three samplings were conducted with the scanning mobility particle sizer for
each condition. The reported result was an averaged of the three results. An example
(engine exhaust running with B20 at 30% load and 50 MPa fuel injection pressure) is
presented in the Figure E-1. The repeatability of the particle size distribution is satisfying.
7
1.0x10
Test1
Test2
8.0x10
6 Test3
6
6.0x10
6
4.0x10
6
2.0x10
0.0
10 100
Figure E-1: Repeatability of particle size distribution of exhaust gases running with
B20 at 30% load and 50 MPa fuel injection pressure
242
VITA
Peng Ye
Education
Aug. 2007 – Dec. 2011
Ph.D in Energy and Mineral Engineering, Fuel Science Option
Dissertation: Investigation of impact of fuel injection strategy and biodiesel
fueling on engine emissions and performance
Aug.2009 – Dec. 2011
M.S. in Energy and Mineral Engineering, Petroleum Engineering Option
Thesis: A pseudofunctionless density diffusivity equation approach to natural gas
reservoir engineering analysis
Sep. 2003 – Jul. 2007
B.S in Geology, Peking University, China
B.S. in Economics, Peking University, China

PengYe Dissertation 2011

Caricato da

Informazioni sul documento

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

PengYe Dissertation 2011

Caricato da

Copyright:

Formati disponibili

The Pennsylvania State University

The Graduate School

INVESTIGATION OF IMPACT OF FUEL INJECTION STRATEGY AND

BIODIESEL FUELING ON ENGINE EMISSIONS AND PERFORMANCE

Energy and Mineral Engineering

Submitted in Partial Fulfillment

for the Degree of

Randy L. Vander Wal

*Signatures are on file in the Graduate School

the biodiesel engine NOx emission penalty can be eliminated.

controlled by the oxygen equivalence ratio at the autoignition zone.

The impact of late in-cylinder (post) injection combustion with biodiesel on

efficiency was investigated with a zero-dimensional engine thermodynamic simulation.

and consequently decrease heat loss.

List of Figures ................................................................................................................... vii

List of Tables. .................................................................................................................. xiv

Acknowledgements ........................................................................................................ xviii

Chapter I. Introduction ................................................................................................... 1

Chapter II. Literature Review.......................................................................................... 7

2.1 Biodiesel background ............................................................................................ 7

3.1 Experiment apparatus.......................................................................................... 49

6.1 Exhaust Emissions ............................................................................................ 108

7.1 Morphology analysis......................................................................................... 129

8.1 Model formulation ............................................................................................ 149

References…. .................................................................................................................. 199

Appendix A: Matlab code ............................................................................................... 229

Appendix B: Repeatability of thermogravimetric analysis ............................................. 236

Appendix C: Repeatability of Raman analysis .............................................................. 239

Appendix D: Repeatability of Morphology analysis ...................................................... 240

Appendix E: Repeatability of particle size distribution (SMPS) result .......................... 242

Figure II-4: A sketch of limit-pressure cycle .................................................................... 46

Figure II-5: Indicated efficiency map with α and β .......................................................... 47

Figure III-2: Molecular structures[217], names, abbreviations, chemical formulas, and

Figure III-3: Thermophoretic sampling apparatus ............................................................ 57

Figure III-6: Demonstration of curve fitting for Raman spectrum ................................... 63

Figure VIII-9: Effect of start of combustion on indicated thermal efficiency, h1/h1b = 1,

Figure VIII-10: Effect of start of combustion on brake thermal efficiency, h1/h1b = 1,

Figure VIII-12: Effect of start of combustion on indicated thermal efficiency, w1/w1b =

Figure VIII-15: Effect of start of combustion on indicated thermal efficiency, w1/w1b = 1,

Figure VIII-16: Effect of start of combustion on brake thermal efficiency, w1/w1b = 1,

Figure VIII-18: Effect of start of combustion on indicated thermal efficiency with

Table III-1: Engine specifications ..................................................................................... 49

Table III-2: Engine operating parameters in Task 1 and 2 ............................................... 54

Table III-3: Engine operating parameters in Task 3 ......................................................... 55

Table III-4: Engine operating parameters in Task 4 ......................................................... 56

Table V-1: Input parameters during calculation ............................................................... 94

Table D-2: repeatability of measurement of diameter of primary particles of diesel soot,

AHR = apparent heat release

NOx = Nitrogen Oxides

IVC = inlet valve close

EVO = outlet valve open

LCA = late crank angle

LOL = Lift-Off Length

l1,l2 = locations of artificial heat release peaks

h1, h2 = height of artificial heat release peaks

w1 = width of artificial Gaussian heat release peaks

TDC = top dead center

TAN = total acid number

FAME = Fatty Acid Methyl Esters

SME = Soybean Methyl Esters

SMPS = scanning mobility particle sizer

SOI = Start of Injection

SOC = Star of Combustion

ULSD = Ultra Low Sulfur Diesel