Fuel 236 (2019) 30–42

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Full Length Article

Development of a new reduced diesel/natural gas mechanism for dual-fuel T

engine combustion and emission prediction
⁎
Haozhong Huang , Delin Lv, Jizhen Zhu, Zhaojun Zhu, Yingjie Chen, Yuping Pan, Mingzhang Pan
College of Mechanical Engineering, Guangxi University, Nanning 530004, China

A R T I C LE I N FO A B S T R A C T

Keywords: A diesel/natural gas (NG) dual-fuel engine is regarded as an appealing option to reduce emissions while
Dual-fuel engine maintaining high thermal efficiency. In this study, a reduced n-heptane–n-butylbenzene–NG–polycyclic aromatic
Reduced mechanism hydrocarbon (PAH) mechanism with 746 reactions and 143 species was developed for predicting the combustion
Diesel characteristics and emission in dual-fuel engines. A mixture of methane, ethane, and propane was used to model
Natural gas
the NG, and a mixture of n-heptane and n-butylbenzene was used to model the diesel. This mechanism was based
Emission
on a reduced n-heptane–PAH mechanism, and the detailed mechanisms of n-butylbenzene and NG were reduced
using the methods of directed relation graph with error propagation (DRGEP), rate of production (ROP), and
sensitivity analysis. The key kinetic parameters of the model were optimized and adjusted according to the
results of sensitivity analysis. The final optimized dual-fuel mechanism was verified against ignition delay, la-
minar flame speed, and homogenous charge compression ignition (HCCI) engine combustion, and a good pre-
diction was obtained. Finally, the present mechanism was coupled into the CFD software to simulate the
combustion characteristics and emission of a dual-fuel engine under four different NG substitution rates. The
simulation results are consistent with the experimental data of emissions and combustion characteristics, in-
dicating that the current mechanism can be applied to simulate practical diesel/NG dual-fuel engines.

1. Introduction advantages; for example, it is easy and inexpensive to convert tradi-

With the rapid development of automobile industry, increasing
demand for energy and decreasing rate of petroleum production, and
pollution problems caused by using fossil fuels, it is becoming in-
creasingly essential to explore alternative fuels that are renewable,
clean, and cost-effective [1,2]. Among various alternative fuels, natural
gas (NG) is a promising suitable alternative fuel to satisfy increasingly
stringent emission standards owing to its high fuel economy [3], clean
combustion characteristics [4], abundance, and relatively low cost [5].
In a diesel/NG dual-fuel engine, the main method of fuel delivery is
that a small amount of diesel is used as the ignition source and directly
injected into the cylinder, while NG is injected into the intake passage
to form a good premixed fuel-air. A diesel/NG dual-fuel engine can
produce a reliable output power and efficiency under different engine
loads. A diesel/NG dual-fuel engine not only has most of the advantages
of traditional diesel engines, but also has more advantages than tradi-
tional diesel engines [6–8]. In a diesel/NG dual-fuel engine, the NG
combustion mode is lean premixed combustion, refraining from the oil-
rich area produced by diesel spray producing soot. In addition, diesel/
NG dual-fuel engines have also attracted much interest because of other
tional diesel engines to dual-fuel engines in a flexible manner [4]. Dual-
fuel engines emit a low amount of CO2 emission because of the low C/H
ratio of NG [9,10] and produce a low amount of particulate matter (PM)
and NOX emissions compared with traditional diesel/gasoline engines
[6,11,12]. Further, an increase in the NG substitution rate can effec-
tively reduce the PM and NOX emissions of dual-fuel engines. However,
if the NG in the cylinder incompletely burns, the methane emission will
increase, thus increasing greenhouse gas emissions. This is because the
global warming potential of methane is 25 times higher than that of
total CO2 during a hundred year period [13]. If the emission of un-
burned hydrocarbon (UHC) can be minimized in dual-fuel engines, a
new method can be provided for implementing new greenhouse gas
emission regulations, thereby saving a lot of fuel cost. Therefore, for
researchers and engine producers, dual-fuel engine has become an
important research topic.
Extensive studies have been carried out to evaluate the emission and
combustion characteristics of dual-fuel engines. Zhang et al. [14] in-
vestigated the impact of injection timing on the emission and com-
bustion performance of dual-fuel engines, and found that the emissions
of carbon oxide (CO) and total hydrocarbon (THC) decreased while the

⁎
Corresponding author.
E-mail address: hhz421@gxu.edu.cn (H. Huang).

https://doi.org/10.1016/j.fuel.2018.08.161
Received 1 June 2018; Received in revised form 27 August 2018; Accepted 31 August 2018
Available online 06 September 2018
0016-2361/ © 2018 Elsevier Ltd. All rights reserved.
H. Huang et al.
Nomenclature
A1 benzene
A2 naphthalene
A3 phenanthrene
A4 pyrene
CA50 crank angle corresponding to 50% of the total heat release
3D CFD three-dimensional computational fluid dynamic
HCCI homogeneous charge compression ignition
PRF primary reference fuel
TRF toluene reference fuel
THC total hydrocarbon
DICI direct injection compression ignition
NTC negative temperature coefficient
Pin initial pressure
PM particular matter
Cn the number of carbon atoms in a fuel molecule
NOX emission increased with increasing injection timing. Zhou et al.
[15] evaluated the impact of injection timing and quantity of pilot
diesel fuel on the combustion performance of a dual-fuel engine. The
results of the study show that the ignition delay increased with in-
creasing injection timing and decreased with increasing amount of pilot
diesel fuel. Imran et al. [16] carried out related experiments for ex-
amining the impact of pilot diesel fuel on the combustion performance
of a dual-fuel engine. It was found that the maximum cylinder pressure
increased, and the ignition delay decreased with the increase in pilot
diesel amount. At high loads, the thermal efficiency was the same or
slightly higher than conventional diesel engines; however, at low loads,
the thermal efficiency was slightly lower than conventional diesel en-
gines. Mittal et al. [17] carried out the experiments to study the char-
acteristics of emissions such as soot, PM, HC, CO, NOX and CO2 on a six-
cylinder conventional diesel engine and dual-fuel engine under dif-
ferent conditions. Karim et al. [18,19] evaluated the effect of methane
fraction, engine load, and quantity of pilot diesel fuel on methane and
CO emissions on a single-cylinder diesel/NG dual-fuel engine.
However, it is time-consuming and expensive to gain in-depth
knowledge and optimize the emission and combustion characteristics of
dual-fuel engines by experiments only. Numerical simulation is a fast
and economical way to better understand the complex combustion,
spray, pollutant formation processes, and evaporation in dual-fuel en-
gines in detail, for which it is difficult to obtain these information
through experimental studies. Singh et al. [20] proposed a dual-fuel
chemical kinetic mechanism and coupled it into KIVA3V software to
model the emission and combustion characteristics of a dual-fuel en-
gine, and found the limitations of this kinetic mechanism for simulating
premixed combustion of NG and air. Aggarwal et al. [21], Maghbouli
et al. [22], and Zhao et al. [23] developed a reduced chemical me-
chanism for dual-fuel engines respectively, where n-heptane was used
as the representative of diesel and methane was used as the re-
presentative of NG. These reduced mechanisms were coupled into CFD
software for multidimensional numerical simulation of engine perfor-
mance [22–24]. Yousefi et al. [24] and Maghbouli et al. [22] only va-
lidated the modeled results of in-cylinder heat release rate (HRR) and
pressure but neglected the laminar flame speed and ignition delay. In
contrast, Zhao et al. [23] simulated the in-cylinder pressure, ignition
delay, HRR, NOX and CO emissions for validation while neglecting the
laminar flame speed, soot and methane emissions. NG is a mixture of
fuel, mainly consisting of methane, ethane, and propane; the related
chemical kinetic mechanisms should include these three substances as
much as possible. Hocktt et al. [25] proposed a reduced chemical me-
chanism for a dual-fuel engine, where n-heptane was used as the re-
presentative of diesel and methane, ethane and propane were used as
31
Fuel 236 (2019) 30–42
ATDC after top dead center
CA crank angle
CO carbon oxide
C2H2 acetylene
CA90 crank angle corresponding to 90% of the total heat release
HACA hydrogen abstraction acetylene addition
PAH polycyclic aromatic hydrocarbon
SOI start of injection
EGR exhaust gas recirculation
HRR heat release rate
NG natural gas
Tin initial temperature
HC hydrocarbon
UHC unburned hydrocarbon
C4H4 tetrahedrane
C4H2 Diacetylene
the representatives of NG. Moreover, they coupled the mechanism into
CFD software for multidimensional numerical simulation engine and
calculated the ignition delay, in-cylinder pressure, HRR, and laminar
flame speed. Based on the numerical simulation results, the developed
reduced mechanism well reproduces the characteristics of detailed
mechanism; however, the emission properties including methane, NOX,
soot and CO emissions were not validated.
In summary, in the simplified dual-fuel kinetic models mentioned
above, most of the kinetic models used a single component of n-heptane
as the representative of diesel and a single component of methane as the
representative of NG. However, diesel is a mixture of more than 200
components. The main components of diesel include alkanes (straight-
chain and branched-chain alkanes), naphthenes, and aromatic hydro-
carbons [26,27]. NG is also a mixed fuel mainly containing methane,
ethane, and propane. It is difficult to establish a dual-fuel chemical
kinetic model containing all the components of diesel and NG. There-
fore, the established dual-fuel engine model should include important
components of diesel and NG as much as possible. Qian et al. [28]
performed engine bench tests on toluene/diesel and n-butylbenzene/
diesel mixtures with the same ratios. The results showed that n-bu-
tylbenzene was more appropriate as the representative of aromatic
hydrocarbons. In the existing reduced chemical kinetic mechanisms of
dual fuels, two or more components have been rarely used as the re-
presentative of diesel (n-butylbenzene was used as the representative of
aromatic hydrocarbons), and a mixture of methane, ethane and pro-
pane has been rarely used as the representative of NG.
Based on the abovementioned reasons, the objective of this study
was to develop a reduced chemical kinetic mechanism of n-heptane–n-
butylbenzene–NG–polycyclic aromatic hydrocarbon (PAH) (denoted as
GXU mechanism) for simulating the combustion characteristics and
emission of a dual-fuel engine. First, the detailed mechanisms for n-
butylbenzene and NG developed by Nakamura et al. [29] and Healy
et al. [30], respectively, were reduced. The reduced mechanisms for NG
and n-butylbenzene were then coupled into the n-heptane–PAH me-
chanism developed by Wang et al. [31,32], thus developing the GXU
mechanism. The GXU mechanism was adjusted by sensitivity analysis
combined with the optimization method proposed by Ra and Reitz
[33,34]; the mechanism was extensively verified by evaluating laminar
flame speeds, ignition delays, and homogenous charge compression
ignition (HCCI) engine combustion. Finally, the GXU mechanism was
coupled into CFD software to verify the emission and combustion
characteristics of a six-cylinder dual-fuel engine against experimental
results under four different NG substitution rates (0, 25%, 50%, and
75%) through numerical simulations.
H. Huang et al.
2. Mechanism development
2.1. Base n-heptane mechanism
The reduced mechanism of primary reference fuel (PRF), which was
developed by Wang et al. [31], including 296 reactions and 73 species,
was comprehensively verified by evaluating the ignition delays, com-
ponent concentration, flame speeds, and the HRR and in-cylinder
pressure profiles in direct injection compression ignition (DICI) engine
and HCCI engine. It was shown that the modeled results were consistent
with experimental values, indicating that PRF mechanism can be em-
ployed to simulate diesel engine combustion characteristics and emis-
sion. Because this mechanism contains isooctane and toluene-related
reactions and components, the relevant reactions and components as-
sociated with isooctane and toluene were first removed from the PRF
mechanism to obtain the mechanism of n-heptane. The reduced me-
chanism of n-heptane was hierarchically formulated from the decom-
position reactions of C7 (7 is the number of carbon atoms in a fuel
molecule) fuel, the intermediate reactions of C2-C4, and the C0-C1 me-
chanisms. In this reduced mechanism, the core mechanism of C3 was
derived from the mechanism of AramcoMech 1.3 [35]. In this study, the
reduced mechanism of n-heptane was employed as the base mechanism
for the reduced GXU mechanism.
2.2. n-Butylbenzene and NG sub-mechanisms
In the present mechanism, n-butylbenzene was taken as the re-
presentative of characterize aromatic hydrocarbons. The detailed n-
butylbenzene mechanism, containing 960 species and 4330 reaction
was mainly obtained by combining the sub-mechanism of alkyl-aro-
matics proposed by Darcy et al. [36] and the mechanism of Aramco-
Mech 1.3 [35] with the toluene mechanism proposed by Metcalfe et al
[37]. Furthermore, the mechanism of aromatic species related to indene
and naphthalene developed by Narayanaswamy et al. [38] has also
been included. The detailed NG mechanism, including 293 species and
1588 reactions was developed by Healy et al. [30]. The detailed NG
mechanism was extensively validated by comparing with the experi-
mental data of ignition delays in the shock tube and rapid compression
machine under pressures of 8–30 bar, equivalence ratios of 0.5, 1.0, and
2.0, and temperatures of 630–1550 K.
The sub-mechanisms of n-butylbenzene and NG were developed by
using the methods of directed relation graph with error propagation
(DRGEP), rate of production (ROP), and sensitivity analysis. During the
simplification of mechanism, first, the method of DRGEP was used to
automatically remove the species and reactions that are not important
for the mechanism. Then, sensitivity analysis, ROP analysis, and
manual simplification were used to further simplify the mechanism. By
constantly applying the above methods repeatedly, the sub-mechanisms
of n-butylbenzene and NG were simplified to the target size.
2.3. PAH sub-mechanism
So far, PAH is widely considered to be the soot precursor. To ac-
curately simulate soot generation, there is necessary to accurately de-
scribe the generation of PAH. Therefore, this work also incorporated a
reduced mechanism of PAH developed by Wang et al. [32]. This re-
duced the detailed mechanism of PAH proposed by Slavinskaya et al.
[39] for ethylene, ethane and methane flames. The PAH sub-mechanism
considers the main production pathways of both C3 + C3, benzene (A1),
and C4 + C2. In addition to relying on the hydrogen abstraction acet-
ylene addition (HACA) mechanism, larger PAHs (such as pyrene (A4),
phenanthrene (A3), and naphthalene (A2)), are mainly generated
through the interactions between small intermediate molecules (such as
diacetylene (C4H2) and tetrahedrane (C4H4)) and aromatic hydrocarbon
groups. The mechanism was also verified against experimental data
from shock tubes, jet-agitated reactors, and DICI engines.
32
Fuel 236 (2019) 30–42
2.4. Formation of reduced GXU mechanism
By merging the sub-mechanisms aforementioned, the reduced GXU
mechanism was established. To ensure that no reactions were elimi-
nated from the base n-heptane mechanism, only the repeating reactions
in the sub-mechanisms of n-butylbenzene, NG and PAH were eliminated
from the combined mechanism. To predict the NOX emission in the
engine, it is necessary to accurately describe the generation processes of
NOX; therefore, a simplified NOX mechanism with 12 reactions and 4
species was also incorporated into the present mechanism for de-
scribing the generation processes of NOX [40]. The pathways of N2O-
intermediate mechanism and thermal mechanism are still retained in
the simplified NOX mechanism. The approach for the formation of re-
duced GXU mechanism is given in Fig. 1.
After coupling the mechanisms together, the preliminary verification of
kinetic model showed that the predicted values of mechanism for n-bu-
tylbenzene and NG in the ignition delays clearly deviated from the experi-
mental data. Therefore, it is necessary to analyze, make the corresponding
adjustments, and optimize the coupled mechanism. To determine the key
reactions of ignition delays, a sensitivity analysis of ignition delays for NG
and n-butylbenzene was performed, as shown in Figs. 2 and 3. A positive
sensitivity coefficient demonstrates that the reaction is promoted, and a
negative sensitivity coefficient demonstrates that the reaction is inhibited.
As shown in Fig. 2, the reactions CH4 + OH = CH3 + H2O and
CH3O2 + CH2O = CH3O2H + HCO are the most sensitive to the ignition
delays of NG. Fig. 3 shows the sensitivity of n-butylbenzene to ignition
delays at 750 K, 1000 K, and 1400 K. As shown in Fig. 3, at 700 K and
1000 K, C6H5C4H8QJD = C6H5C4H7-3 + HO2, C6H5C4H9 + OH =
DC6H5C4H8 + H2O, C6H5C4H9 + HO2 = DC6H5C4H8 + H2O2, and
C6H5CH3 + O2 = C6H5CH2 + HO2 are most sensitive to the ignition delays
of n-butylbenzene. Finally, the optimization method proposed by Ra and
Reitz [34] was applied to optimize the reactions for ignition delays. The
adjustment results are shown in Table 1. The GXU mechanism with
CHEMKIN format is given in Supplemental material.
3. Mechanism validation
3.1. Ignition delays
Ignition delay is a very important parameter because it determines
the fuel mixing time and also affects the distribution, combustion, and
emission characteristics of fuel. Ignition delay is calculated using a
closed, homogenous, and constant reactor model. Shen et al. [41] and
Herzler et al. [42] performed rapid compression machine and shock
tube tests, respectively, to measure the ignition delays of n-heptane.
Fig. 4 displays the comparison between the experimental ignition de-
lays of n-heptane at different equivalent ratios and simulation results.
The present mechanism models the ignition delays of n-heptane well
and especially reflects the negative temperature coefficient (NTC)
phenomenon of n-heptane at low temperatures. Fig. 5 presents a com-
parison of the measured ignition delays of n-butylbenzene under dif-
ferent equivalent ratios and pressures with the corresponding simula-
tion data; the experimental data were obtained from the shock tube test
results tested by Darcy et al. [43] in the medium- and high-temperature
range and rapid compression machine test results measured by Naka-
mura et al. [29] in the low-temperature range. The prediction results
measured with the GXU mechanism fit well with the abovementioned
experimental data, especially in the high-temperature range. Despite
certain deviations between the simulation results and experimental
data, the NTC of n-butylbenzene in the low-temperature range can be
well predicted. Thus, it can be concluded that the reduced n-bu-
tylbenzene mechanism has overall good consistency with the detailed n-
butylbenzene mechanism. Because the related experimental data of NG
in these conditions are still rare, the prediction results of ignition delays
measured with both simplified and detailed mechanisms were com-
pared, as shown in Fig. 6. The 3.3/6.7/90 mixture is composed of 3.3%
H. Huang et al. Fuel 236 (2019) 30–42

Detailed n-butylbenzene Detailed NG mechanism PRF mechanism,

mechanism, 960 species 293 species and 1588 73 species and 296
and 4330 reaction reactions reactions

The 24 species and 75

DRG, DRGEP, ROP
reactions related to
Sensitivity analysis
iso-octane were removed

N-butylbenzene NG Based n-heptane

sub-mechanism sub-mechanism mechanism

A reduced PAH A reduced NOx

mechanism mechanism

Reduced diesel/NG mechanism mechanism, 143 species and 746 reactions

Fig. 1. The approach for the formation of reduced GXU mechanism.

C6H5C4H8QJD=C6H5C4H7-3+HO2
Natural gas C3H6OOH1-2=C3H6+HO2
C6H5C4H9+OH=CC6H5C4H8+H2O
T=850K P=30bar CH3O2+CH2O=CH3O2H+HCO 1400 K
C6H5CH3+O2=C6H5CH2+HO2
1000 K
Phi=0.1 NC3H7O2=C3H6OOH1-2
C6H5C4H9+H=BC6H5C4H8+H2 750 K
IC3H7O2=C3H6+HO2
n-Butylbenzene C6H5C4H9(M)=C6H5CH2+NC3H7(M)
IC3H7O2=C3H6OOH2-1 Phi=1.0
C6H5CH2+HO2=C6H5CH2O+OH
30 atm
CH3O2+C3H8=CH3O2H+IC3H7 C6H5C4H9+HO2=DC6H5C4H8+H2O2

C3H8+HO2=NC3H7+H2O2 BBZQDQJB=BBZODQB+OH

C3H8+HO2=IC3H7+H2O2 C6H5C4H8QJD=BBZQDRB

CH4+OH=CH3+H2O C6H5C4H9+OH=DC6H5C4H8+H2O

H2O2(+M)=2OH(+M) -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6

Sensitivity coefficient
-0.8 -0.4 0.0 0.4 0.8
Fig. 3. Sensitivity analysis of ignition delays for n-butylbenzene at Φ = 1.0,
Sensitivity Cofficient
pressure of 30 atm, and temperature of 750 K, 1000 K, and 1400 K.
Fig. 2. Sensitivity analysis of ignition delays for NG for Φ = 1.0, pressure of
30 bar, and temperature of 850 K.
Table 1
Most sensitive reactions for ignition delays in the n-butylbenzene and NG sub-
propane, 6.7% ethane, and 90% methane, and the 10/20/70 mixture is mechanisms.
composed of 10% propane, 20% ethane, and 70% methane. Fig. 6 Reaction Pre-exponential Pre-exponential
clearly shows the simulated ignition delays modeled with the reduced factor before factor after
mechanism are consistent with those modeled with the detailed me- optimization optimization
chanism under different temperatures, equivalent ratios, and pressures,
C6H5C4H9(+M) = C6H5CH2 + NC3H7(+M) 2.304E + 22 4.608E+22
indicating that the reduced NG mechanism can be employed to model
C6H5C4H9 + HO2 = DC6H5C4H8 + H2O2 1.02E+04 5.10E+03
ignition delays. C6H5C4H9 + H = BC6H5C4H8 + H2 1.30E+06 2.06E+07
BBZQDQJB = BBZODQB + OH 1.25E+10 1.00E+10
C6H5CH3 + O2 = HO2 + C6H5CH2 2.18E+07 1.09E+07
3.2. Laminar flame speeds C6H5O = C5H5 + CO 2.50E+11 5.00E+12
C6H5CH2 + HO2 = OH + C6H5CH2O 1.00E+13 4.00E+13
Laminar flame speed is also a vital combustion parameter for fuel C6H5OH + OH = C6H5O + H2O 2.95E+06 1.475E+06
CH3O2 + CH2O = HCO + CH3O2H 1.99E+12 0.8E+12
combustion; therefore, the current mechanism was also used to verify IC3H7O2 = CH3H6OOH2-1 1.8E+12 1.5E+12
the laminar flame speed. The laminar flame speeds of n-heptane/air CH4 + OH = CH3 + H2O 5.83E+4 6.199E+4

33
H. Huang et al. Fuel 236 (2019) 30–42

mixture were measured by Lipzig et al. [44] under a temperature of

Fig. 4. Comparisons of experimental [41,42] and modeled ignition delays of n-
heptane.
298 K and pressure of 1 bar. Mehl et al. [45] used the heat flux method
to test the premixed flame speeds of n-butylbenzene under different
conditions (1) bar, 358 K/398 K). The prediction results of laminar
flame speeds of n-butylbenzene and n-heptane were also compared with
the experimental values; the results are shown in Fig. 7. Overall, the
experimental data are consistent with the simulation data. Rozenchan
et al. [46] tested the laminar flame speeds of methane in a constant-
volume chamber under different equivalent ratios and pressures
(40 atm and 60 atm). Specifically, a mixture of helium and oxygen with
a volume ratio of 83:17 was used as the oxidant under 40 atm, and a
mixture of helium and oxygen with a volume ratio of 85:15 was used as
the oxidant under 60 atm. The temperature of unburned gas was
maintained at 298 K. Fig. 8 shows a comparison of the measured flame
velocities by Rozenchan et al. [46] with the simulation results. In the
lean-fuel region, the predicted flame speeds are consistent with the
experimental data; in the rich-fuel region, the predicted flame speeds
are clearly larger than the experimental results. However, these two
results show an overall identical trend. Actually, a strong turbulent
flame velocity in a dual-fuel diesel engine mainly depends upon the
turbulent intensity, whereas laminar flame speed has a slight effect.

Fig. 5. Comparisons of experimental [29,43] and modeled ignition delays of n-butylbenzene. Symbols: experimental data. Solid lines: reduced GXU mechanism;
dashed lines: detailed n-butylbenzene mechanism.

34
H. Huang et al. Fuel 236 (2019) 30–42

1000 10000
3.3/6.7/90 NG in air 70/20/10 NG in air
ĭ 1.0 ĭ 0.2
1000
100

Igniyion Delay (ms)

Igniyion Delay (ms)

100
10
10
1
NUIG,10bar NUIG,10bar
1
NUIG,30bar NUIG,30bar
NUIG,80bar NUIG,80bar
0.1
GXU,10bar 0.1 GXU,10bar
GXU,30bar GXU,30bar
GXU,80bar GXU,80bar
0.01 0.01
0.7 0.8 0.9 1.0 1.1 1.2 1.3 0.7 0.8 0.9 1.0 1.1 1.2 1.3
1000/T (1/K) 1000/T (1/K)
(a) (b)
10000 10000
3.3/6.7/90 NG in air 10/20/70 NG in air
ĭ 0. ĭ 0.
1000 1000
Igniyion Delay (ms)

Igniyion Delay (ms)

100 100

10 10

NUIG,10bar NUIG,10bar
1 1
NUIG,30bar NUIG,30bar
NUIG,80bar NUIG,80bar
0.1 GXU,10bar 0.1 GXU,10bar
GXU,30bar GXU,30bar
GXU,80bar GXU,80bar
0.01 0.01
0.7 0.8 0.9 1.0 1.1 1.2 1.3 0.7 0.8 0.9 1.0 1.1 1.2 1.3
1000/T (1/K) 1000/T (1/K)

(c) (d)
10000 1000
3.3/6.7/90 NG in air 10/20/70 NG in air
ĭ 0.2 ij 1
1000
100
Igniyion Delay (ms)
Igniyion Delay (ms)

100
10
10
1
NUIG,10bar NUIG,10bar
1
NUIG,30bar NUIG,30bar
NUIG,80bar NUIG,80bar
0.1
0.1 GXU,10bar GXU,10bar
GXU,30bar GXU,30bar
GXU,80bar GXU,80bar
0.01 0.01
0.7 0.8 0.9 1.0 1.1 1.2 1.3 0.7 0.8 0.9 1.0 1.1 1.2 1.3
1000/T (1/K) 1000/T (1/K)

(e) (f)
Fig. 6. Comparisons of the reduced and detailed mechanisms of NG. Dashed lines: reduced GXU mechanism; solid lines: detailed NG mechanism (NUIG).

Moreover, because NG is mainly burned in the lean-fuel region of a
dual-fuel engine, the flame speeds in this region can be well predicted.
Thus, the established reduced mechanism can be used to predict the
flame speeds of NG in a dual-fuel engine.
3.3. HCCI combustion
In previous studies, the ignition delays were measured under a
constant volume for validation; the initial pressure or temperature was
also fixed. However, the temperature and pressure in the engine

35
H. Huang et al. Fuel 236 (2019) 30–42

45
n-heptane
P=1.0 bar
Sim 60 n-butylbenzene/air 398K
40 Exp P=1.0 bar 358K
Flame Speed (cm/s) Tin=298K
35 50

Flame Speed (cm/s)

30
40

25
30
20
20
15

10 10
0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Equivalence Ratio Equivalence Ratio
(a) (b)
Fig. 7. Comparisons of experimental [44,45] and modeled laminar flame speeds of n-heptane and n-butylbenzene.

25
Experimental,60 atm,15% O2
50 n-heptane/air Current 360K
Simulation,60 atm,15% O2 Current 390K
Phi=0.8
Experimental,40 atm,17% O2 20 Current 420K
Pini=1.2 bar
TRF 360K
40 Simulation,40 atm,17% O2
Flame speed (cm/s)

Pressure (MPa) TRF 390K

TRF 420K
15
30

10
20

5
10

0
-40 -20 0 20 40
0
0.8 0.9 1.0 1.1 1.2 1.3 1.4 Crank angle (deg)
Equivalence Ratio
Fig. 9. Comparisons of TRF mechanism and current mechanism for HCCI en-
Fig. 8. Comparisons of experimental [46] and modeled laminar flame speeds of gine simulation.
methane.
10 show the simulation validation results of n-heptane and NG in the
Table 2 HCCI engine simulator. The mechanism of toluene reference fuel (TRF)
Engine Operating Conditions and Specifications. [32] that exhibits good prediction results for the HCCI engine was used
Engine type 6MK375
in this study to validate the established mechanism. The simulation
Stroke × bore 145 mm × 123 mm results of NG measured by the established mechanism were compared
Connecting rod length 224 mm with those measured by the detailed mechanism of NG. Regarding NG,
Compression ratio 17.5:1 the data modeled by the reduced mechanism are clearly consistent with
Displacement 10.338 L
those modeled by the detailed mechanism at a high initial temperature;
Number of nozzle 8
Intake valve close timing −156 CA ATDC however, at a low initial temperature, the results simulated by the re-
Exhaust valve open timing 144 CA ATDC duced mechanism significantly deviate from those simulated by the
Speed 1220 rpm detailed mechanism. Regarding n-heptane mechanism, the results si-
Intake air pressure 1.8 bar
mulated with the reduced mechanism are highly consistent with those
Intake air temperature 310 K
SOI −9 CA ATDC
obtained with TRF mechanism. Overall, the reduced mechanism ex-
EGR 0 hibits the desired prediction performance, indicating that the reduced
mechanism is very reliable and can be used to predict the combustion
performance in the HCCI engine.
cylinder are variable during the piston motion. Then, the measurements
were performed under varying temperature and pressure conditions to
further validate the reliability of established GXU dual-fuel mechanism. 4. Engine application
HCCI combustion is mainly subjected to the effects of chemical kinetic
mechanism while unaffected by the evaporation and spray of physical A 6MK375 dual-fuel engine was used in the experiment. Fig. 11
processes. Therefore, the established GXU dual-fuel mechanism was shows the schematic diagram of engine setup used in this study. The
validated under zero-dimensional adiabatic HCCI engine simulation experimental conditions of diesel/NG dual-fuel engine are given in
conditions. Table 2 shows the engine parameters in detail. Figs. 9 and Table 2. The exhaust gas recirculation (EGR) was maintained at 0. The
start of injection (SOI) was maintained at −9 CA ATDC. The intake

36
H. Huang et al. Fuel 236 (2019) 30–42

20 20
3.3/6.7/90 NUIG,400K 10/20/70 NUIG,400K
18 NUIG,420K 18 ĭ=0.5 NUIG,420K
ĭ=0.5
NUIG,440K P=1.7 bar NUIG,440K
16 P=1.7 bar NUIG,460K
16 NUIG,460K
14 GXU,400K 14 GXU,400K

Pressure (MPa)
GXU,420K GXU,420K
Pressure (MPa)

12 GXU,440K 12 GXU,440K
GXU,460K GXU,460K
10 10
8 8
6 6
4 4
2 2
0 0
-20 -10 0 10 20 30 -20 -10 0 10 20 30
Crank Angle (deg) Crank Angle (deg)
(a) (b)
Fig. 10. Comparisons of detailed mechanism of NG (NUIG) and present mechanism (GXU) for HCCI engine simulation.

Fig. 11. Schematic diagram of engine setup.

pressure and temperature were maintained at 1.8 bar and 310 K, re- Table 3
spectively. The dual-fuel engine speed was maintained at 1220 rpm. Sub-models used in the simulation.
The NOX emission was measured using a HORIBA MEXA-710DEGR. Description Model
Four different NG substitution rates were used in the experiment: 0%
NG, 25% NG, 50% NG, and 75% NG. The substitution rate of NG was Droplet collision Walljet model [47]
calculated using the following equations: Evaporation Multi-component model [48]
Spray breakup KH-RT model [49]
Total energy of fuel:LHVDiesel × mDiesel + LHVnatural gas × mnatural gas (1) Turbulence Renormalized group (RNG) k-ε model [50]

LHVnatural gas × mnatural gas

NG substitution rate:
LHVDiesel × mDiesel + LHVnatural gas × mnatural gas
× 100%
where LHVnatural gas and LHVDiesel denote the low heating values of NG
and diesel, respectively; mnatural gas and mDiesel denote the quality of NG
and diesel, respectively.
In this study, to make the power tendency of engine consistent with
energy efficiency, the total energy of fuel was kept the same.
To further validate the reliability of reduced dual-fuel mechanism,
in this study, the GXU mechanism was coupled into 3D CFD software
for numerical simulation. Then, the modeled results were compared
with the experimental values. Table 3 shows the used physical models
in this simulation. Using the renormalized group (RNG) k-ε model, the
turbulence model was used to simulate the turbulence characteristics
(2)
[47]. The spray atomization and droplet breakup processes were pre-
dicted by the Kelvin–Helmholtz and Rayleigh–Taylor (KH-RT) models,
respectively [48]. First, the KH model was used for simulating the in-
itial breakup of droplets caused by the unstable waves generated by the
velocity difference between the liquid phase and gas phase. For the KH
model, the droplet size depends on the fastest growing wavelength or
the wavelength of the most probable unstable surface wave. Then, the
RT model was employed to simulate smaller droplets of the second
breakup process of droplets. A multi-component model was employed
for predicting the evaporation of mixed fuel droplets [49]. The model
offers a possibility to predict the spray combustion of multi-component

37
H. Huang et al.
Fig. 12. 45° sector mesh calculation network.
Table 4
Mass of NG and diesel consumed under each experimental condition.
Case tested NG mass (mg/ Diesel mass (mg/ Substitution rate of NG (%)
cycle) cycle)
Case 1 0 94.5 0
Case 2 23.4 69.1 25
Case 3 40.5 50.4 50
Case 4 70.67 25.2 75
surrogate model that is able to accurately reflect the physical and
chemical properties of each component. Droplet collision plays an im-
portant role in small-bore diesel engines. Therefore, Walljet model was
employed to predict the interaction of droplet with the cylinder wall
[50]. The combustion process was simulated by combining the current
mechanism with the SAGE combustion model [51]. In the simulation
process, a mixture of n-heptane and n-butylbenzene was used as the
representatives of diesel for modeling diesel combustion, while a mix-
ture of methane, ethane, and propane was used as the representatives of
NG for modeling NG combustion. Because an 8-hole fuel injector was
used in this study, a 45° sector mesh computing network as shown in
Fig. 12 was used in the simulation for saving time. The adaptive mesh
refinement method was applied to refine the meshes of high tempera-
ture and velocity gradient regions with a minimum cell size of 0.25 mm
in all direction. In all direction, the size of base cell was 2 mm. The
adaptive mesh refinement and base cell size were used based on the
mesh sensitivity research. The temperature and velocity of sub-grid
criterion were set as 5 K and 2 m/s, respectively. The simulation lasted
from the intake valve closing (−156 CA ATDC) to the exhaust valve
opening (144 CA ATDC); the preset initial conditions were identical to
the engine’s operating conditions shown in Table 2. Table 4 lists the
mass flow rate of NG and diesel under each experimental condition.
Fig. 13 shows a comparison of simulated in-cylinder HRR and
pressure with the experimental data measured under four different NG
substitution rates (0, 25%, 45%, and 75%). Overall, the modeled results
of in-cylinder pressure are consistent with the experimental results, but
the simulation results of HRR are slightly higher than the experimental
data under the four air conditions. The primary reasons are as follows:
First, only two different components of surrogate fuels were used in this
simulation as the representative of diesel. Second, the impact of tur-
bulence on chemical kinetics was inadequately considered in the si-
mulation, and fuel injection rules may differ from the experimental
results. Finally, the measured data showed some error in the experi-
ments. Although certain errors exist in the simulation, the established
mechanism can be employed to predict the combustion characteristics
of a dual-fuel engine under different substitution rates of NG.
In-cylinder temperature, one of the key parameters, plays an im-
portant role in controlling the auto-ignition process. As shown in
Fig. 14, the in-cylinder mean temperature simulated with the present
mechanism decreased with increasing substitution rate of NG. As shown
in Fig. 16, this trend indicates the spatial distribution of temperature at
38
Fuel 236 (2019) 30–42
CA50. At a smaller substitution rate of NG, more diesel was injected
into the cylinder, the corresponding spray penetration was longer, and
many fire sources for igniting NG were formed. This also exhibited a
wider distribution. Therefore, the combustion rate of fuel in the cy-
linder was larger, and the in-cylinder temperature was higher. In con-
trast, at a higher substitution rate of NG, less diesel was used for igni-
tion, the spray penetration was shorter, and a smaller number of fire
sources for igniting NG were formed. This also exhibited a smaller
distribution. Additionally, when the concentration of NG was higher,
the oxygen concentration was lower. This inhibited the collision be-
tween pilot diesel and oxygen molecules, thereby decreasing the che-
mical reaction rate, delaying the combustion of diesel, and lowering the
maximum in-cylinder temperature.
Fig. 15 shows a comparison of the modeled results of soot, HC, NOX
and CO emissions with the experimental data. Both the HC and CO
emissions from a dual-fuel engine were higher than those obtained from
a traditional diesel engine, but the NOX and soot emissions obtained
from a dual-fuel engine were less than those obtained from a traditional
diesel engine. Despite certain deviations, the simulation results and
experimental data exhibit similar trends, indicating that the reduced
mechanism is able to reasonably predict the trends of soot, NOX, HC and
CO emissions. Furthermore, based on the abovementioned observa-
tions, the GXU mechanism was validated.
The experimental data are consistent with the simulated emission of
NOX; this further confirms the sub-mechanism of NOX in the developed
mechanism. High temperature, sufficient oxygen, and long duration
time are the preconditions of NOX production. According to the ex-
tended Zeldovich mechanism [52], NO is the main component in NOX.
The following three chemical reactions are involved in the production
of NO:
N2 + O< = > N+ NO (3)
N+ O2 < = > O+ NO (4)
N+ OH < = > H+ NO (5)
As the substitution rate of NG was increased, less NOX was emitted.
As shown in Fig. 16, such a phenomenon could be confirmed by the
spatial NOX concentration contour at CA90. This is because less fuel was
injected with the increase in the substitution rate of NG. As a result,
both the number and range of diesel pilot fire nuclei decreased, in-cy-
linder combustion rate slowed down, combustion became less in-
tensive, and maximum in-cylinder temperature also decreased, thus
reducing the NO and NOX production.
Fig. 15 shows that the soot emission also decreased with the in-
crease in the substitution rate of NG. Such a trend is also observed from
the spatial distribution pattern of soot at CA90 as shown in Fig. 16. Soot
is mainly produced under low-oxygen-concentration and high-com-
bustion-temperature conditions, where the fuel molecules are prone to
pyrolysis. With the increase in the substitution rate of NG, less pilot
diesel was injected, the ignition delay period was prolonged, and the
local over-thick regions in the cylinder shrank. Accordingly, both the
number and range of pilot fire nuclei decreased, and the ratio of dif-
fusive combustion and in-cylinder combustion temperature decreased,
thus inhibiting soot production and reducing soot emission.
As also shown in Fig. 15, as the substitution rate of NG increased,
both the CO and HC emissions (where methane occupied the main
proportion) increased. This clear trend is shown in Fig. 16 exhibiting
the spatial CO and methane concentration contours at CA90. The
generation of CO is mainly affected by the combustion temperature and
oxygen concentration inside the cylinder. When the combustion tem-
perature in the cylinder is high and the oxygen concentration is large,
CO oxidizes to generate CO2, thus reducing the CO emission. With the
increase in the substitution rate of NG, both the excess air coefficient
and in-cylinder temperature decreased, and less CO was oxidized, thus
increasing the CO emission.
The amount of UHC emission in exhaust gas mainly depends upon
H. Huang et al. Fuel 236 (2019) 30–42

the fuel combustion quality in the combustion chamber. In addition, the 2000
0% NG
total relative air/fuel ratio affects the in-cylinder combustion process
25% NG

In-cylinder mean temperature/K

and plays a very important role in the flame propagation mechanism 45% NG
under a dual-fuel engine [53]. The HC emission is also affected by the 1600
75% NG
geometry, such as squish volumes and crevices, and the combustion
temperature of the cylinder near cylinder walls. As shown in Fig. 16,
most of the unburned methane remained in the combustor wall and 1200
moved away from the pilot fuel plume. As the substitution rate of NG
increased, less air entered the cylinder, the air/fuel ratio increased, and
800
the in-cylinder temperature decreased, thus increasing the HC emission.

400
5. Conclusion

This study established a reduced GXU dual-fuel mechanism, mainly 0

including 143 species and 746 reactions. This mechanism was estab- -40 -20 0 20 40 60 80
lished by adding the reduced mechanisms of n-butylbenzene and NG Crank Angle /°CA
into the n-heptane–PAH mechanism. The detailed mechanisms of n-
butylbenzene and NG were first simplified using DRGEP, ROP, and Fig. 14. Simulation of in-cylinder mean temperature simulated by the present
sensitivity analysis. Then, the critical kinetic parameters of the estab- mechanism under different NG substitution rates (0, 25%, 45%, and 75%).
lished model were optimized. After the optimization, the reduced GXU
dual-fuel mechanism was extensively validated with ignition delay, 1. A reduced GXU mechanism for dual-fuel engine was established,
laminar flame speed, HCCI engine and dual-fuel engine combustion. mainly including 143 species and 746 reactions. A mixture of n-
The main conclusions are as follows: butylbenzene and n-heptane was used as the representative of diesel

800 800
120 0% NG Exp 120 25% NG Exp
Sim 700 Sim 700

100 600 100 600

HRR /(J(°CA)-1)
HRR /(J(°CA)-1)

Pressure /bar

80 500 500
Pressure /bar

80

400 400
60 60
300 300
40 40
200 200

20 100 20 100

0 0 0 0
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
Crank Angle /°CA Crank Angle /°CA

(a) (b)
120 800
1200 Exp
75% NG
120 45% NG Exp Sim 700
Sim 100
1000
100 600
80
HRR /(J(°CA)-1)

800
HRR /(J(°CA)-1)

Pressure /bar

500
80
Pressure /bar

60
400
600
60
40 300
400
40 200
20

20 200 100
0
0
0 0
-80 -60 -40 -20 0 20 40 60 80 -80 -60 -40 -20 0 20 40 60 80
Crank Angle /°CA Crank Angle /°CA

(c) (d)
Fig. 13. Comparison of simulated in-cylinder HRR and pressure with experimental data measured at 0% NG, 25% NG, 50% NG, and 75% NG.

39
H. Huang et al. Fuel 236 (2019) 30–42

Experiment 0.020 Experiment

9
Simulation Simulation
0.018
8

0.016

Soot [g/kW.h]
NOX[g/kW.h]

7
0.014
6
0.012
5
0.010
4 0.008

3 0.006
0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75
NG ratio NG ratio
(a) (b)
10 16
Experiment Experiment
Simulation 14 Simulation
8
12
CO[g/kW.h]

6
HC[g/kW.h]
10

8
4
6

2 4

2
0
0
0.00 0.25 0.50 0.75 0.00 0.25 0.50 0.75
NG ratio NG ratio

(c) (d)
Fig. 15. Comparisons of experimental CO, NOX and HC emissions with simulated results.

CA50 CA50 CA90 CA90 CA90 CA90

0% NG

25% NG

50% NG

75% NG

Fig. 16. Temperature, OH, NOX, CO and methane (CH4) spatial distribution contours with different substitution rates of NG.

40
H. Huang et al.
fuel, and a mixture of propane, ethane, and methane was used as the
representative of NG.
2. The optimized mechanisms for n-butylbenzene and n-heptane still
exhibit NTC phenomenon. Using the laminar flame speed, ignition
delay, and HCCI engine combustion to validate the proposed me-
chanism, a good prediction performance was obtained.
3. By coupling the reduced GXU mechanism into CFD, the CO, HC,
NOX and soot emissions and combustion characteristics of different
substitution rates of NG in a dual-fuel engine were simulated for
validation. The results show that the trends of HC, soot, CO and NOX
emissions and combustion characteristics are consistent with the
simulation results.
4. Compared with the NG substitution rate of 0%, the NOX and soot
emissions decrease, and the mean temperature in the cylinder de-
creases, while the CO and HC emissions increase under the condi-
tion of the NG substitution rate of 75%.
Acknowledgement
The research is sponsored by projects of Natural Science Foundation
of Guangxi (project Outstanding Young Scholarship Award, Grant
No.2014GXNSFGA118005), Natural Science Foundation of China
(51076033), Natural Science Foundation of Guangxi Province
(2017GXNSFAA0198376), and Guangxi Science and Technology
Development Plan (AC16380047).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.fuel.2018.08.161.
References
[1] Sangeeta Moka S, Pande M, et al. Alternative fuels: an overview of current trends
and scope for future. Renew Sustain Energy Rev 2014;32(5):697–712.
[2] Fu J, Liu J, Feng R, et al. Energy and exergy analysis on gasoline engine based on
mapping characteristics experiment. Appl Energy 2013;102(2):622–30.
[3] Cho HM, He BQ. Spark ignition natural gas engines—a review. Energy Convers
Manage 2007;48(2):608–18.
[4] Hosmath RS, Banapurmath NR, Khandal SV, et al. Effect of compression ratio, CNG
flow rate and injection timing on the performance of dual fuel engine operated on
honge oil methyl ester (HOME) and compressed natural gas (CNG). Renew Energy
2016;93:579–90.
[5] Poompipatpong C, Cheenkachorn K. A modified diesel engine for natural gas op-
eration: performance and emission tests. Energy 2011;36(12):6862–6.
[6] Sahoo BB, Sahoo N, Saha UK. Effect of engine parameters and type of gaseous fuel
on the performance of dual-fuel gas diesel engines—a critical review. Renew Sustain
Energy Rev 2009;13(6):1151–84.
[7] Yousefi A, Birouk M. Investigation of natural gas energy fraction and injection
timing on the performance and emissions of a dual-fuel engine with pre-combustion
chamber under low engine load. Appl Energy 2017;189:492–505.
[8] Yousefi A, Birouk M, Lawler B, et al. Performance and emissions of a dual-fuel pilot
diesel ignition engine operating on various premixed fuels. Energy Convers Manage
2015;106:322–36.
[9] Papagiannakis RG, Hountalas DT. Combustion and exhaust emission characteristics
of a dual fuel compression ignition engine operated with pilot Diesel fuel and
natural gas. Energy Convers Manage 2004;45(18):2971–87.
[10] Cordiner S, Rocco V, Scarcelli R, et al. Experiments and multi-dimensional simu-
lation of dual-fuel diesel/natural gas engines. SAE technical papers 2007-24-0124.
[11] Nithyanandan K, Lin Y, Donahue R, et al. Characterization of soot from diesel-CNG
dual-fuel combustion in a CI engine. Fuel 2016;184:145–52.
[12] Mousavi SM, Saray RK, Poorghasemi K, et al. A numerical investigation on com-
bustion and emission characteristics of a dual fuel engine at part load condition.
Fuel 2016;166:309–19.
[13] Trenberth K E, Jones P, Ambenje P, et al. Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the Fourth Assessment Report of
the Intergovernmental Panel on Climate Change (IPCC). Computational Geometry
2007; 2007(2):1–21.
[14] Zhang Q, Li N, Li M. Combustion and emission characteristics of an electronically-
controlled common-rail dual-fuel engine. J Energy Inst 2015;89(4):766–81.
[15] Zhou L, Liu YF, Wu CB, et al. Effect of the diesel injection timing and the pilot
quantity on the combustion characteristics and the fine-particle emissions in a
micro-diesel pilot-ignited natural-gas engine. Proc Inst Mech Eng, Part D: J
Automob Eng 2013;227(8):1142–52.
[16] Imran S, Emberson DR, Diez A, et al. Natural gas fueled compression ignition engine
41
Fuel 236 (2019) 30–42
performance and emissions maps with diesel and RME pilot fuels. Appl Energy
2014;124(1):354–65.
[17] Mittal M, Donahue R, Winnie P, et al. Exhaust emissions characteristics of a multi-
cylinder 18.1-L diesel engine converted to fueled with natural gas and diesel pilot. J
Energy Inst 2014;88(3):275–83.
[18] Karim GA. An Examination of Some Measures for Improving the Performance of Gas
Fuelled Diesel Engines at Light Load. SAE Paper 912366 1991.
[19] Karim GA, Liu Z, Jones W. Exhaust Emissions from Dual Fuel Engines at Light Load.
SAE Paper 932822 1993.
[20] Singh S, Kong SC, Reitz RD, et al. Modeling and Experiments of Dual-Fuel Engine
Combustion and Emissions. SAE 2004-1-0092.
[21] Aggarwal SK, Awomolo O, Akber K. Ignition characteristics of heptane–hydrogen
and heptane–methane fuel blends at elevated pressures. Int J Hydrogen Energy
2011;36(23):15392–402.
[22] Maghbouli A, Shafee S, Saray RK, et al. A multi-dimensional CFD-Chemical kinetics
approach in detection and reduction of knocking combustion in diesel-natural gas
dual-fuel engines using local heat release analysis. SAE Int J Engines
2013;6(2):777–87.
[23] Zhao W, Yang W, Fan L, et al. Development of a skeletal mechanism for heavy-duty
engines fuelled by diesel and natural gas. Appl Therm Eng 2017;123:1060–71.
[24] Yousefi A, Guo H, Birouk M. An experimental and numerical study on diesel in-
jection split of a natural gas/diesel dual-fuel engine at a low engine load. Fuel
2018;212.
[25] Hockett A, Hampson G, Marchese AJ. Development and validation of a reduced
chemical kinetic mechanism for computational fluid dynamics simulations of nat-
ural gas/diesel dual-fuel engines. Energy Fuels 2016:30(3).
[26] Pitz WJ, Mueller CJ. Recent progress in the development of diesel surrogate fuels.
Prog Energy Combust Sci 2011;37(3):330–50.
[27] Anand K, Ra Y, Reitz RD, et al. Surrogate model development for fuels for advanced
combustion engines. Energy Fuels 2011;25(4):1474–84.
[28] Qian Y, Qiu Y, Zhang Y, et al. Effects of different aromatics blended with diesel on
combustion and emission characteristics with a common rail diesel engine. Appl
Therm Eng 2017;125.
[29] Nakamura H, Darcy D, Mehl M, et al. An experimental and modeling study of shock
tube and rapid compression machine ignition of n-butylbenzene/air mixtures.
Combust Flame 2014;161(1):49–64.
[30] Healy D, Kalitan DM, Aul CJ, et al. Oxidation of C1–C5 alkane quinternary natural
gas mixtures at high pressures. Energy Fuels 2010;24(3).
[31] Wang H, Yao M, Reitz RD. Development of a reduced primary reference fuel me-
chanism for internal combustion engine combustion simulations. Energy Fuels
2013;27(12):7843–53.
[32] Wang H, Yao M, Yue Z, et al. A reduced toluene reference fuel chemical kinetic
mechanism for combustion and polycyclic-aromatic hydrocarbon predictions.
Combust Flame 2015;162(6):2390–404.
[33] Ra Y, Reitz RD. A reduced chemical kinetic model for IC engine combustion si-
mulations with primary reference fuels. Combust Flame 2008;155(4):713–38.
[34] Ra Y, Reitz RD. A combustion model for IC engine combustion simulations with
multi-component fuels. Combust Flame 2011;158(1):69–90.
[35] Metcalfe WK, Burke SM, Ahmed SS, et al. A hierarchical and comparative kinetic
modeling study of C1–C2, hydrocarbon and oxygenated fuels. Int J Chem Kinet
2013;45(10):638–75.
[36] Darcy D, Mehl M, Simmie JM, et al. An experimental and modeling study of the
shock tube ignition of a mixture of n-heptane and n-propylbenzene as a surrogate
for a large alkyl benzene. Proc Combust Inst 2013;34(1):411–8.
[37] Metcalfe WK, Dooley S, Dryer FL. Comprehensive detailed chemical kinetic mod-
eling study of toluene oxidation. Energy Fuels 2011;25(11):4915–36.
[38] Narayanaswamy K, Blanquart G, Pitsch H. A consistent chemical mechanism for
oxidation of substituted aromatic species. Combust Flame 2010;157(10):1879–98.
[39] Slavinskaya NA, Riedel U, Dworkin SB, et al. Detailed numerical modeling of PAH
formation and growth in non-premixed ethylene and ethane flames. Combust Flame
2012;159(3):979–95.
[40] Sun Y, Reitz RD. Modeling Diesel Engine NOx and Soot Reduction with Optimized
Two-Stage Combustion. SAE Paper 2006-01-0027.
[41] Shen HPS, Steinberg J, Vanderover J, et al. A shock tube study of the ignition of n-
heptane, n-decane, n-dodecane, and n-tetradecane at elevated pressures. Energy
Fuels 2009;23(5):1367–75.
[42] Herzler J, Jerig L, Roth P. Shock tube study of the ignition of lean n-heptane/air
mixtures at intermediate temperatures and high pressures. Proc Combust Inst
2005;30(1):1147–53.
[43] Darcy D, Tobin CJ, Yasunaga K, et al. A high pressure shock tube study of n-pro-
pylbenzene oxidation and its comparison with n-butylbenzene. Combust Flame
2012;159(7):2219–32.
[44] Lipzig JPJV, Nilsson EJK, Goey LPHD, et al. Laminar burning velocities of n-hep-
tane, iso-octane, ethanol and their binary and tertiary mixtures. Fuel
2011;90(8):2773–81.
[45] Mehl M, Pitz WJ, Westbrook CK, et al. Kinetic modeling of gasoline surrogate
components and mixtures under engine conditions. Proc Combust Inst
2011;33(1):193–200.
[46] Rozenchan G, Zhu DL, Law CK, et al. Outward propagation, burning velocities, and
chemical effects of methane flames up to 60 ATM. Proc Combust Inst
2002;29(2):1461–70.
[47] Han Z, Reitz RD. Turbulence modeling of internal combustion engines using RNG k-
e models. Combust Sci Technol 1995;106:267–95.
[48] Beale JC, Reitz RD. Modeling spray atomization with the Kelvin-Helmotz /Rayleigh-
Taylor hybrid model. Atomization Sprays 1999;9:623–50.
[49] Butler TD, Cloutman LD, Dukowicz JK, Ramshaw DJ. Multidimensional numerical
H. Huang et al.
simulation of reactive flow in internal combustion engines. Prog Energy Combust
Sci 1981;7:293–315.
[50] Naber JD, Reitz RD. Modeling Engine Spray/Wall Impingement. SAE technical
paper series 1988.
[51] Senecal PK, Pomraning E, Richards KJ, et al. Multi-Dimensional Modeling of Direct-
Injection Diesel Spray Liquid Length and Flame Lift-off Length using CFD and
Parallel Detailed Chemistry. SAE Technical Paper. 2003-01-1043 2003.
42
Fuel 236 (2019) 30–42
[52] Tao F, Golovitchev VI, Chomiak J. Self-ignition and early combustion process of n-
heptane sprays under diluted air conditions: numerical studies based on detailed
chemistry. SAE 2000-01-2931.
[53] Papagiannakis RG, Rakopoulos CD, Hountalas DT, et al. Emission characteristics of
high speed, dual fuel, compression ignition engine operating in a wide range of
natural gas/diesel fuel proportions. Fuel 2010;89(7):1397–406.