Energy Conversion and Management 133 (2017) 511–523

Contents lists available at ScienceDirect

Energy Conversion and Management

journal homepage: www.elsevier.com/locate/enconman

Numerical investigation of the effect of injection strategy on mixture

formation and combustion process in a port injection natural gas rotary
engine
Baowei Fan a, Jianfeng Pan a,⇑, Wenming Yang b, Yangxian Liu a, Stephen Bani a, Wei Chen a
a
School of Energy and Power Engineering, Jiangsu University, Zhenjiang 212013, Jiangsu, China
b
Department of Mechanical Engineering, National University of Singapore, 119260, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: This work aimed to numerically study the influence of injection strategy on mixture formation and com-
Received 2 September 2016 bustion process in a port injection natural gas rotary engine. On the base of a 3D dynamic simulation
Received in revised form 20 October 2016 model which was established in our previous work, some critical information was obtained, which
Accepted 28 October 2016
was difficult to obtain through experiment, in terms of the flow field, the fuel distribution, the temper-
Available online 5 November 2016
ature field and the concentration fields of some intermediates. Simulation results showed that for mix-
ture formation, the movements of fuel in injection stage were mainly controlled by the intensity of the
Keywords:
vortex I for injection timing, and the value of jet flux for injection duration respectively. With retarded
Natural gas
Rotary engine
injection timing, the decreasing intensity of the vortex I resulted in less fuel moving toward the back
Port injection strategy of the combustion chamber. With the extension in injection duration, the decreasing value of jet flux
Mixture formation resulted in more fuel staying at the back of the combustion chamber. For combustion process, the overall
Combustion process combustion rate for injection strategy which had an injection timing of 390 °CA (BTDC) and injection
Three-dimensional dynamic simulation duration of 51.5 °CA (case ID4) was the fastest. This was mainly due to the fact that the accumulation area
of fuel was at the middle and front of the combustion chamber. Meanwhile, fuel concentration near the
leading and trailing spark plugs was conducive for the flame kernel formation. Compared with the injec-
tion strategy which had an injection timing of 450 °CA (BTDC) and an injection duration of 55 °CA (case
IT1), the improved combustion rate of case ID4 had a 23% increase in the peak pressure, but also a certain
increase in NO emissions.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
With energy crisis and environmental pollution occurring,
energy saving and environmentally friendly electric vehicle (EV)
in existence presents a great potential for the future [1,2]. How-
ever, the small range of electric vehicles is still one of the problems
inhibiting the spread of these cars. Range extender (RE) in general
is the realistic possibility to increase the range of plug-in hybrid
electric vehicle (PHEV) in the near future without the need to
improve the technology of the main battery storage. For improving
the efficiency of entire vehicles, RE desperately needs a kind of
power equipment with high-energy density and high power-to-
weight ratio [3,4]. Compared with conventional reciprocating
engine, the advantages of the rotary engine, like compact design,
⇑ Corresponding author at: School of Energy and Power Engineering, Jiangsu
University, 301 Xuefu Road, Zhenjiang 212013, China.
E-mail address: mike@ujs.edu.cn (J. Pan).
multi-fuel capability, low noise and low vibration levels, make it
more suitable as the power equipment of RE [5]. For example,
the combination of a compact rotary engine and a small generator
dynamo is considered one of the most promising power generation
modes for PHEV [6]. Therefore, for the wide application prospects
described above, more attention is being given to the research on
the performance and emission of rotary engine by many countries
and research institutes.
Currently, various fuels such as gasoline, diesel, aviation kero-
sene, natural gas and hydrogen, have been successfully applied to
rotary engines [7]. Relatively speaking, gaseous fuels like natural
gas and hydrogen, are more suitable to be used in rotary engines.
This is mainly due to the fact that, gaseous fuel is easy to form
combustible mixture in the cylinder as compared to diesel, avia-
tion kerosene and other traditional liquid fuel. In addition, acting
as an ecologically sensitive and efficient fuel, natural gas is often
recognized as a promising option put forward by governments
for a sustainable energy system [8]. Therefore, the natural

http://dx.doi.org/10.1016/j.enconman.2016.10.070
0196-8904/Ó 2016 Elsevier Ltd. All rights reserved.
512 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523
gas-fueled rotary engine is considered to be a new and clean
energy system. However, the question of how to improve the com-
bustion efficiency of the rotary engine has been a major issue. This
is because flow pattern in the combustion chamber is unidirec-
tional flow at ignition timing. The general direction of the unidirec-
tional flow is from the back to the front of the combustion
chamber, which results in the different flame propagation in the
front and the back of the combustion chamber. Along with the
direction of the flow, it is easy for the flame to propagate toward
the front of the combustion chamber, which makes the fuel in
the front of combustion chamber burn out without delay. How-
ever, against the direction of flow, it is difficult for the flame to
propagate toward the back of the combustion chamber, which
leaves a lot of unburned fuel in the rear of combustion chamber
[9]. Moreover, when natural gas is used in rotary engine, its low
burning speed could also exacerbate the above problem.
Fortunately, the application of stratified combustion mode is con-
sidered as a key solution for these problems [10]. This is mostly
because the combustion rate can be increased by controlling the
fuel distribution, although the unidirectional flow pattern in the
combustion chamber cannot be changed at ignition timing. Port
injection and direct injection in cylinder are two ways of realizing
stratified combustion. As compared to direct injection in cylinder,
port injection is widely used in rotary engine, because of its low
requirement on electronic fuel injection system [11]. Therefore, it
is imperative to study the realization mechanism of different strat-
ified mixture and its effect on flame propagation in the port injec-
tion natural gas rotary engine. Meanwhile, the injection strategy is
one of the most influential factors for engine performance. The
fuel-air mixing procedure significantly depends on the injection
position, injection timing (IT), injection pressure and injection
duration (ID) [12,13]. What is more, the mixture quality in the
combustion chamber determines the combustion characteristics
of the engine. At present, there are many studies on how to impact
the combustion process by the injection strategy in reciprocating
engines. For example, Yang et al. [14] experimentally studied the
effects of natural gas port injection and diesel pilot injection on
the combustion and emissions of a turbocharged common rail
dual-fuel engine at low load. The results indicated that retarded
natural gas injection timing achieved a stratified-like air-fuel mix-
ture in cylinder under the different pilot injection conditions,
which provided a method to improve the combustion performance
and exhaust emissions at low load. Wang et al. [15] investigated
the combustion and emission characteristics of a diesel engine
with dimethyl ether (DME) as port premixing fuel under different
injection timing. It was found that port premixing DME quantity
played an important role in combustion and emission control. Dur-
ing low-temperature reaction phase, the peak value of heat-release
rate (HRR) increased with a rise of port DME quantity because of
relatively richer mixture concentration. Thangavel et al. [16] made
experimental studies on simultaneous injection of ethanol-
gasoline in the intake port of a four stroke spark ignition (SI)
engine. It was found that the efficiency and torque were better at
a mass ratio of 1:1 (50% of ethanol and 50% of gasoline) due to fas-
ter combustion as a result of better mixture preparation, as com-
pared to conventional pre-blended injection of gasoline and
ethanol. Lakshmanan et al. [17] made experimental studies at the
optimized fixed injection duration of 90 °CA and injector opening
time of 5 °CA (ATDC), to investigate the gas flow rate on the com-
bustion, performance and emission of the engine. The experimen-
tal results show that the NO emission decreased with increase in
gas flow rate, as compared to diesel fuel operating at full load. As
seen above, despite previous works suggesting suitable injection
parameters conducive to improve combustion efficiency, research
about the effect of injection parameters on mixture formation
and combustion processes in a port injection natural gas rotary
engine are still rare. For example, as the air flow controls air-fuel
mixture in rotary engine, particle image velocimetry (PIV) method
was applied by Maki et al. [18] to measure the flow field in a rotary
engine. However, due to the limitations of the experimental set-up,
only the 2D flow field at low-speed was measured, that is, the 3D
flow field under normal engine operating conditions could not be
tested. Kawahara et al. [19] developed an optical spark-plug sensor
with a double-pass measurement length using an infrared absorp-
tion technique for measuring hydrocarbon fuel concentrations.
Cycle-resolved measurements were made to investigate the effects
of the fuel concentration near the spark plug on the combustion
characteristics of the commercial rotary engine. However, the dis-
tribution and movement of fuel in other positions of the cylinder
could not be measured. Karatsu et al. [20] developed an optical
rotary engine, and combustion in the combustion chamber was
observed by bottom view and side view simultaneously using
two high-speed cameras. Their work made significant contribu-
tions to the study of the flame propagation in the rotary engine.
This once again demonstrated that it was easy to propagate the
flame toward the front of the combustion chamber, but difficult
to propagate the flame toward the rear of the combustion chamber
by experimental research. However, further research on how to
impact the combustion process by the fuel distribution is still
needed. Spreitzer et al. [21] used the CONVERGE software to ana-
lyze the gas flow and the combustion processes in the three com-
bustion chambers simultaneously. Although the mixture formation
process was not discussed, their work showed that the rotary
engine could be investigated in an unprecedented analysis depth
through simulation. For example, Jeng et al. [22] used the FLUENT
software to numerically investigate the influence of leakage
through the apex seal, fuel type and recess size on the performance
of a rotary engine. Their numerical results gave more insight into
the effect of these parameters on the power output of a rotary
engine.
This work seeks to bridge the gaps identified in literature, using
numerical simulation to study the influence of injection strategy
on mixture formation and combustion process in a port injection
natural gas rotary engine. On the basis of the 3D dynamic simula-
tion model which was established in our previous work [23,24],
some critical information is obtained which is difficult to acquire
through experiment. These are the flow field, the fuel distribution,
the temperature field and some intermediate concentration fields.
Our study also outlines a theoretical guide to determine the best
injection strategy in the port injection natural gas rotary engine
under different working conditions.
2. Geometric model generation and meshing
2.1. Computational domain
In the present study, the rotary engine has two spark plugs,
leading spark plug (LSP) and trailing spark plug (TSP). A schematic
diagram of the rotary engine is shown in Fig. 1(a). For convenience
of discussion, the names of the components of the rotary engine
are shown in Fig. 1(b). Technical specifications of the engine are
listed in Table 1.
The simulations were performed at an engine speed of
3500 rpm and at wide open throttle. Natural gas was injected into
the inlet port and all computations were made with excess air ratio
of 1.5. The appropriate time window of the nozzle for fuel injection
is from 470 °CA (BTDC) to 280 °CA (BTDC) for nozzle A. Therefore,
at a fixed injection pressure, five different injection timings were
used to study the effect of injection timing on mixture formation
and combustion process in the engine. After that, at a fixed injec-
tion timing (390 °CA (BTDC)), eight different injection durations
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523 513

(a) Schematic

(b) Component name

Fig. 1. The schematic and component name of side-ported rotary engine.

Table 1 were used to study the effect of injection duration on mixture for-
Engine technical specifications. mation and combustion process in the engine. These parameters of
Engine parameters Value
the different injection strategies are defined in Table 2. Among
them, the injection strategy for case IT3 is same as that for case
Number of rotors 1
Cooling Air cooled
ID5, and the two different case names are given for the conve-
Ignition source Spark plug nience of discussion in Section 4.
Displacement 160 cm3
Compression ratio 8:1 2.2. Mesh generation
Generating radius 69 mm
Eccentricity 11 mm
Chamber width 40 mm A user defined function (UDF) was programmed and used to con-
Ignition timing for TSP/LSP 50 °CA (BTDC)/50 °CA (BTDC) trol the motion of the rotor in ANSYS FLUENT software [25]. The
Intake timing Advance angle 70 °CA (ATDC), three combustion chambers have the same working process. There-
Delay angle 63 °CA (ABDC) fore, only one of the combustion chamber was used in the simula-
Exhaust timing Advance angle 62 °CA (BBDC),
Delay angle 70.5 °CA (ATDC)
tion to save calculation time. The mesh for the rotary engine is as
shown in Fig. 2. A grid independence test was conducted and the
grid size of 2.2 mm was chosen for the combustion chamber. In
addition, the grid size of the injection nozzle was set to 0.5 mm [26].

Table 2
3. Mathematical models and model validation
Summary of the different injection strategies.
The model selection in this study is shown in Table 3. The rea-
Case Injection timing Injection Injection Volumetric
duration (°CA) pressure (bar) coefficient
sons for using the selected models can be found in literature
[24,27]. In addition, the reduced chemical mechanism which has
IT1 450 °CA (BTDC) 55 1.2 0.8980
IT 2 420 °CA (BTDC) 55 1.2 0.8940
16 components and 41 reaction steps [28], was used to calculate
IT 3 390 °CA (BTDC) 55 1.2 0.8875 the combustion process. Pressure boundary conditions were set
IT 4 360 °CA (BTDC) 55 1.2 0.8824 at intake and exhaust port respectively. As the engine was natu-
IT 5 330 °CA (BTDC) 55 1.2 0.8794 rally aspirated, the values for the pressure boundaries were set
ID1 390 °CA (BTDC) 44 2.0 0.8872
to that of atmospheric pressure. Meanwhile, non-slip conditions
ID2 390 °CA (BTDC) 46 1.8 0.8878
ID3 390 °CA (BTDC) 48 1.6 0.8884 were set at the walls for the velocities and temperature was set
ID4 390 °CA (BTDC) 51.5 1.4 0.8881 to 400 K. Meanwhile, to validate the selected model, experimental
ID5 390 °CA (BTDC) 55 1.2 0.8875 results from PIV test and indicator diagram, were used to compare
ID6 390 °CA (BTDC) 60 1.0 0.8882 with simulation results. Detailed descriptions on experimentation
ID7 390 °CA (BTDC) 68 0.8 0.8882
ID8 390 °CA (BTDC) 81 0.6 0.8883
and the validation procedures can be found in our previous work
[23,24].
514 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523

4. Results and discussion

4.1. Effect of injection strategy on mixture formation in the rotary

engine

For the convenience of discussion, the flow field in cylinder for

the condition without fuel injection is illustrated in Fig. 3. At the
start of the intake stroke, the motion track of the flow is marked
with two arrows in Fig. 3(a). When the fresh mixture enters the
combustion chamber, it impinges on the lower cylinder head. After
the impingement, two vortexes (vortex I and vortex II) are formed
in the front and back of the combustion chamber respectively.
After that, the flow which comes from the back of the combustion
chamber continues to move forward along the upper cylinder head
as indicated by the arrow located at the back of combustion cham-
ber. In the early stages of the compression stroke as seen from
Fig. 3(b), with the decrease in volume in front of the combustion
chamber, the vortex I has broken down. As seen in Fig. 3(c), during
the late stages of the compression stroke, due to a sharp decrease
in volume in the combustion chamber, the vortex II is also basically
Fig. 2. Mesh of the side-ported rotary engine with fuel nozzle (450 °CA (BTDC)).
disappeared, and the main flow pattern in the combustion cham-
ber is unidirectional flow.

Table 3 4.1.1. Movement of fuel in cylinder under different injection timing

Summary of models selection. At a fixed injection pressure, the injection timing is retarded
Object Mode selection and boundary conditions continually from case IT1 to IT5. The movement of fuel for case
Turbulence Two-equation RNG k–e sub-model IT1, case IT3 and case IT5 are illustrated in Figs. 4–7. It can be seen
Combustion Eddy-dissipation concept (EDC) model from Fig. 4, when the crank angle is 25 °CA after the start of injec-
NO formation ‘‘Thermal NO” and ‘‘Prompt NO” model tion, the movement trend of the fuel is significantly affected by the
vortexes (vortex I and vortex II) in the combustion chamber. For

(a)350°CA (BTDC) (b) 150°CA (BTDC)

(c) 50°CA (BTDC)

Fig. 3. Streamlines in cylinder of the condition without fuel injection.
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523
case IT1, the mass of fuel which flows to the back of combustion
chamber is basically the same as the fuel which flows to the front
of the combustion chamber. For cases IT3 and IT5, the mass of fuel
which flows to the front of the combustion chamber is more than
fuel which flows to the back of the combustion chamber. This is
mainly because with retarded injection timing, although the
impact position of jet flow on the lower cylinder head moves
toward the back of combustion chamber, the decreasing intensity
of the vortex I results in less fuel which moves toward the back
of combustion chamber. As shown in Fig. 5, when the crank angle
is 75 °CA after the start of injection, for case IT1, the fuel in the back
of combustion chamber continues to move forward along the
upper cylinder head. For cases IT3 and IT5, the mass of the fuel
which moves from back to front of the combustion chamber along
the upper cylinder head, is smaller than that for case IT1. When the
crank angle is 200 °CA (BTDC), it can be seen from Fig. 6 that, for
case IT1, the fuel is distributed along the upper cylinder head
throughout the combustion chamber. However, for cases IT3 and
IT5, the fuel is mainly distributed in the middle and front of the
combustion chamber. As shown in Fig. 7, during the later stages
of the compression stroke (100 °CA (BTDC)), for case IT1, the fuel
is distributed evenly throughout the combustion chamber. For case
IT3 and IT5, the fuel distribution has obvious stratification phe-
nomenon. The accumulation area of fuel is at the middle and front
of the combustion chamber for case IT3, and at the front of the
combustion chamber for case IT5.
As seen from above analysis, with retarded injection timing, the
decreasing intensity of vortex I results in less fuel which moves
toward the back of combustion chamber in injection stage. During
the later stages of the compression stroke, with retarded injection
timing, the fuel stratification phenomenon becomes increasingly
obvious, and the accumulation area of fuel moves toward the front
of combustion chamber continuously.
(a) Case IT1 at - 425o CA
(c) Case IT5 at - 305o CA
Fig. 4. The iso-surface of fuel mass fraction and the streamlines in cylinder at 25 °CA after the start of injection for different injection timings.
515
4.1.2. Movement of fuel in cylinder under different injection duration
At a fixed injection timing (390 °CA (BTDC)), the injection dura-
tion extends continually from case ID1 to ID8. The movement of
fuel for case ID1, case ID4 and case ID8 is illustrated in Figs. 8–12.
When the crank angle is 365 °CA (BTDC), it can be seen from
Fig. 8, with the extension of injection duration, the mass which
enters the combustion chamber is reduced gradually at the same
time. Therefore, it can be seen from Fig. 9, when the crank angle
is 315 °CA (BTDC), for case ID1, the injection process is complete,
and the fuel in the back of combustion chamber moves forward
along the upper cylinder head. However, for case ID8, due to the
low jet flux, the injection process is ongoing and fuel mainly stays
in the back of the combustion chamber. That is, with the extension
of injection duration, more fuel stays in the back of combustion
chamber. When the crank angle is 200 °CA (BTDC), as seen in
Fig. 10, for case ID1, case ID4 and case ID8, the fuel in the back
of the combustion chamber continues to move forward along the
upper cylinder head. As the rotor rotates, when the crank angle
is 150 °CA (BTDC), it can be seen from Fig. 11, for case ID1, most
of the fuel in back of combustion chamber has moved to the front
of the combustion chamber. However, for case ID8, a lot of fuel still
remains in the back of the combustion chamber. During the later
stages of the compression stroke (100 °CA (BTDC)), it can be seen
that from Fig. 12, the accumulation area of fuel is at the front of
the combustion chamber for case ID1, at the middle and front of
combustion chamber for case ID4. For case ID8, the accumulation
area of fuel is at the middle and front of the combustion chamber.
Meanwhile, some fuel remains at the rear of the combustion
chamber.
As seen from above analysis, with the extension in injection
duration, the decreasing value of jet flux results in more fuel
staying in the back of the combustion chamber in injection stage.
During the later stages of the compression stroke, with the
(b) Case IT3 at -365o CA
516 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523

(a) Case IT1 at -375o CA (b) Case IT3 at -315o CA

(c) Case IT5 at -255o CA

Fig. 5. The iso-surface of fuel mass fraction and the streamlines in cylinder at 75 °CA after the start of injection for different injection timings.

(a) Case IT1 at -200o CA (b) Case IT3 at -200o CA

(c) Case IT5 at -200o CA

Fig. 6. The iso-surface of fuel mass fraction and the streamlines in cylinder at 200 °CA (BTDC)) for different injection timings.
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523 517

(a) Case IT1 at -100o CA (b) Case IT3 at -100o CA

(c) Case IT5 at -100o CA

Fig. 7. The iso-surface of fuel mass fraction and the streamlines in cylinder at 100 °CA (BTDC)) for different injection timings.

(a) Case ID1 at -365o CA (b) Case ID4 at-365o CA

(c) Case ID8 at -365o CA

Fig. 8. The iso-surface of fuel mass fraction and the streamlines in cylinder at 365 °CA (BTDC)) for different injection durations.
518 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523

(a) Case ID1 at -315o CA (b) Case ID4 at-315o CA

(c) Case ID8 at -315o CA

Fig. 9. The iso-surface of fuel mass fraction and the streamlines in cylinder at 315 °CA (BTDC)) for different injection durations.

(a) Case ID1at -200o CA (b) Case ID4 at -200o CA

(c) Case ID8 at -200o CA

Fig. 10. The iso-surface of fuel mass fraction and the streamlines in cylinder at 200 °CA (BTDC)) for different injection durations.
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523 519

(a) Case ID1 at -150o CA (b) CaseID4 at-150o CA

(c) Case ID8 at -150o CA

Fig. 11. The iso-surface of fuel mass fraction and the streamlines in cylinder at 150 °CA (BTDC)) for different injection durations.

(a) Case ID1 at -100o CA (b) Case ID4 at -100o CA

(c) Case ID8 at -100o CA

Fig. 12. The iso-surface of fuel mass fraction and the streamlines in cylinder at 100 °CA (BTDC)) for different injection durations.
520 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523

o o
Injection Duration (CA/ ) Crank Angle (CA/ )
40 45 50 55 60 65 70 75 80 85 -40 -20 0 20 40 60 80
37 37 1.0 0.0
Peak Pressure in Cylinder (bar)

Peak Pressure in Cylinder (bar)

36 36

Mass Fraction of CH4 Burned

Mass Fraction of CH4 Burned
35 35 0.8 0.2

34 34 -40 -20
0.0
ID of 44oCA
33 33 0.6 0.4
ID of 51.5oCA
ID of 68oCA
32 32 0.1
ID of 81oCA
0.4 0.6
31 31
0.2
30 30
0.2 0.8
29 29 IT of 450oCA
-460 -440 -420 -400 -380 -360 -340 -320 IT of 390oCA
o IT of 330oCA
Injection Timing (CA/ ) 0.0 1.0
-40 -20 0 20 40 60 80
Fig. 13. Comparison of the peak cylinder pressure of different injection strategies. o
Crank Angle (CA/ )

Fig. 14. Mass fraction of CH4 burnt of different injection strategies.

decrease in injection duration, fuel stratification phenomenon
becomes increasingly obvious, and the accumulation area of fuel
moves toward the front of combustion chamber continuously. There are two peaks for the peak cylinder pressure curve, the first
is at an injection duration of 51.5 °CA (case ID4), and the second is
4.2. Effect of injection strategy on combustion process in the engine at an injection duration of 81 °CA (case ID8). As known, the cylin-
der pressure is determined by the fuel combustion rate. Fig. 14
Fig. 13 shows the peak cylinder pressure of different injection shows the mass fraction of CH4 burnt for different injection strat-
strategy. It can be seen that, with retarded injection timing, the egy. The crank angle duration from 0% to 90% heat release of the
peak cylinder pressure increases initially and then decreases, and total fuel (CA10-90%) is used to reflect the quality of the combus-
the highest peak cylinder pressure appears at injection timing of tion within an engine. It can be seen that with retarded injection
390 °CA (BTDC) (case IT3). At a fixed injection timing (390 °CA timing, the change rule of the flame propagation duration
(BTDC)), with the extension in injection duration, the peak cylinder (CA0-90%) is similar to that of the peak cylinder pressure, and
pressure first increases, then decreases and at last increases again. the fuel combustion rate for case IT3 is the maximum. However,

Table 4
Comparison of the fuel mass fraction distribution at ignition timing and CH2O mass fraction distribution at TDC for different injection timings.

Case Fuel distributions at ignition timing CH2O concentration field at TDC

IT1

IT3

IT5
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523
with the extension in injection duration, the change rule of the
flame propagation duration (CA0-90%) is different from that of
the peak cylinder pressure. Although the flame propagation dura-
tion (CA0-90%) for case ID1 is smaller than that for case ID8, the
peak cylinder pressure for case ID8 (34.0 bar) is bigger than that
for ID1 (35.4 bar). The reason is that as can be seen from the partial
enlarged drawing in Fig. 14, in the early stages of the combustion
stroke (from 50 °CA (BTDC) to 20 °CA (BTDC)), the fuel combustion
rate for case ID8 is the maximum.
Through the above-mentioned discussion, combustion charac-
teristics are greatly affected by the injection strategies. To investi-
gate further, fuel distribution at ignition timing and the CH2O
concentration field at TDC are shown in Tables 4 and 5. In particu-
lar, the edge of CH2O concentration field represents the edge of the
flame. This is chiefly because the concentration of CH2O reaches a
maximum before the combustion reactions takes place. It can be
seen from Table 4, with retarded injection timing, the accumula-
tion area of fuel moves toward the front of the combustion
chamber continuously at ignition timing, as described in
Section 4.1.1. Due to the unidirectional flow in the combustion
chamber, the flame mainly propagates in a forward direction [9].
For case IT1, the fuel is distributed evenly throughout the combus-
tion chamber, which results in a lot of unburnt fuel in the rear
Table 5
Comparison of the fuel mass fraction distribution at ignition timing and CH2O mass fraction distribution at TDC for different injection durations.
Case Fuel distributions at ignition timing
ID1
ID4
ID7
ID8
521
combustion chamber at TDC. For case IT3, the accumulation area
of fuel is at the middle and front of the combustion chamber, which
avoids a lot of unburnt fuel in the rear of the combustion chamber.
For case IT5, although the accumulation area of fuel is mainly at
the front of the combustion chamber at ignition timing, fuel com-
bustion rate for the case IT5 is lower than that for case IT3. This is
essentially because the fuel concentration near the trailing spark
plug (TSP) is too low, which hinders the flame kernel formation.
As shown in Table 4 of the CH2O concentration field at TDC for case
IT5, the flame front area near the trailing spark plug is small and
confines to the upper cylinder head.
It can be seen from Table 5 that, at a fixed injection timing
(390 °CA (BTDC)), with the decrease in injection duration, the accu-
mulation area of fuel moves toward the front of the combustion
chamber continuously at ignition timing, as described in
Section 4.1.2. For case ID1, the accumulation area of fuel is at the
front of combustion chamber, and the fuel concentration near
the trailing spark plug (TSP) is too low like case IT5, which hinders
the flame kernel formation. For case ID4, less fuel stays in the rear
of the combustion chamber compared to case IT3, which acquires a
higher fuel combustion rate. For case ID7, the fuel concentration
near the leading spark plug (LSP) is lower than that of case ID4,
which prevents the flame around the LSP from spreading rapidly
CH2O concentration field at TDC
522 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523

in front of the combustion chamber. For case ID8, due to long injec- NO reaches a maximum value (0.0019) for case ID4, which shows
tion duration, the fuel stratification phenomenon is not obvious. a 58% increase compared with case IT1 (0.0012).
The fuel distributes more evenly for case ID8 than the other cases,
which results in a higher combustion rate in the early stages of the
5. Conclusions
combustion stroke (from 50 °CA (BTDC) to 20 °CA (BTDC)). How-
ever, its overall combustion rate is smaller than case ID4.
In the present study, on the basis of a 3D dynamic simulation
It can be seen from the above analysis, compared with the other
model which was established in our previous work, further simu-
injection strategies that, the overall combustion rate for case ID4 is
lations were conducted to investigate the impact of mixture forma-
the fastest. The peak pressure in combustion process reaches a
tion and combustion processes by injection timing and injection
maximum value (36.3 bar) for case ID4, which shows a 23%
duration in a port injection natural gas rotary engine. The simula-
increase compared with case IT1 (29.5 bar). This is mainly because
tions were performed at an engine speed of 3500 rpm and at wide
the accumulation area of fuel is at the middle and front of the com-
open throttle, and the following conclusions were drawn.
bustion chamber. Meanwhile, the fuel concentration near the lead-
ing and the trailing spark plug is conducive to the flame kernel
1. For injection timing, with retarded injection timing, the
formation.
decreasing intensity of vortex I results in less fuel which moves
Figs. 15 and 16 show the average temperature and mass frac-
toward the back of the combustion chamber in injection stage.
tion of NO for cases IT1, IT3, IT5, ID1, ID4, ID7 and ID8, respectively.
During the later stages of the compression stroke, with retarded
As known, NO formation is mostly affected by the local gas
injection timing, the fuel stratification phenomenon becomes
temperature, oxygen concentration and residence time. The earlier
increasingly obvious, and the accumulation area of fuel moves
discussions show that the gross combustion rate for case ID4 is the
toward the front of the combustion chamber.
highest, contributing to the peak average cylinder temperature as
2. For injection duration, with the extension in injection duration,
shown in Fig. 15. Therefore, the drawback of case ID4 is a certain
the decreasing value of jet flux results in more fuel which stays
increase in NO emissions at the later stages of the combustion
in the back of the combustion chamber in injection stage. Dur-
stroke (80 °CA (ATDC)), as shown in Fig. 16. The mass fraction of
ing the later stages of the compression stroke, with the decrease
in injection duration, the fuel stratification phenomenon
becomes increasingly obvious, and the accumulation area of
Crank Angle (CA/o)
fuel moves toward the front of the combustion chamber.
-40 -20 0 20 40 60 80 3. Compared with the other injection strategies, the overall com-
2200 400
bustion rate for the injection strategy which had an injection
Average temperature in cylinder (T/K)

Average temperature in cylinder (T/K)

2000 600 timing of 390 °CA (BTDC) and injection duration of 51.5 °CA
1800 800 (case ID4), is the fastest. This is due to the fact that the accumu-
lation area of fuel is at the middle and front of combustion
1600 1000
ID of 44oCA
chamber. Meanwhile, the fuel concentration near the leading
1400 ID of 51.5oCA 1200 and the trailing spark plug is conducive for flame kernel
ID of 68oCA formation.
1200 1400
ID of 81oCA 4. Under the computational condition, compared with the injec-
1000 1600 tion strategy which had an injection timing of 450 °CA (BTDC)
800 1800 and an injection duration of 55 °CA (case IT1), the improved
IT of 450oCA combustion rate of case ID4 has a 23% increase in the peak pres-
600 IT of 390oCA 2000
sure, but also a 58% increase in NO emissions.
IT of 330oCA
400 2200
-40 -20 0 20 40 60 80
o
Crank Angle (CA/ ) Acknowledgements
Fig. 15. Average temperature in cylinder of different injection strategies.
We gratefully acknowledge financial support from National
Natural Science Foundation of China (Nos. 51576093 and
51606089), Project Funded by thePriority Academic Program
o
Crank Angle (CA/ ) Development of Jiangsu Higher Education Institutions (PAPD),
-40 -20 0 20 40 60 80 and the Startup Foundation of Jiangsu University for Advanced
0.0025 0.0000 Scholars (No. 16JDG045).
o
IT of 450 CA
IT of 390oCA
0.0020 IT of 330oCA 0.0005 References
Mass Fraction of NO
Mass Fraction of NO

[1] Alves J, Baptista PC, Gonçalves GA, Duarte GO. Indirect methodologies to
0.0015 ID of 44oCA
0.0010 estimate energy use in vehicles: application to battery electric vehicles. Energy
ID of 51.5oCA Convers Manage 2016;124:116–29.
ID of 68oCA [2] Zhao D. Waste thermal energy harvesting from a convection-driven Rijke-Zhao
0.0010 ID of 81oCA 0.0015 thermo-acoustic-piezo system. Energy Convers Manage 2013;66(2):87–97.
[3] Zhao D, Li L. Effect of choked outlet on transient energy growth analysis of a
thermoacoustic system. Appl Energy 2015;160:503–10.
0.0005 0.0020 [4] Ribau João, Silva Carla, Brito Francisco P, Martins Jorge. Analysis of four-stroke,
Wankel, and microturbine based range extenders for electric vehicles. Energy
Convers Manage 2012;58:120–33.
0.0000 0.0025 [5] Ji Changwei, Su Teng, Wang Shuofeng, et al. Effect of hydrogen addition on
-40 -20 0 20 40 60 80 combustion and emissions performance of a gasoline rotary engine at part load
o and stoichiometric conditions. Energy Convers Manage 2016;121:272–80.
Crank Angle (CA/ ) [6] Hubmann Christian, Beste Frank, Friedl Hubert, Schöffmann Wolfgang. Single
cylinder 25 kW range extender as alternative to a rotary engine maintaining
Fig. 16. Mass fraction of NO in cylinder of different injection strategies. high compactness and NVH performance. SAE paper 2013-32-9132; 2013.
 B. Fan et al. / Energy Conversion and Management 133 (2017) 511–523
[7] Amrouche F, Erickson PA, Park JW, et al. Extending the lean operation limit of a
gasoline Wankel rotary engine using hydrogen enrichment. Int J Hydrogen
Energy 2016;41(32):14261–71.
[8] Tang Qijun, Fu Jianqin, Liu Jingping, et al. Performance improvement of
liquefied natural gas (LNG) engine through intake air supply. Appl Therm Eng
2016;103:1351–61.
[9] Fan Baowei, Pan Jianfeng, Liu Yangxian, Zhu Yuejin. Effects of ignition
parameters on combustion process of a rotary engine fueled with natural
gas. Energy Convers Manage 2015;103:218–34.
[10] Muroki Takumi, Moriyoshi Yasuo, Takagi Masahide. Combustion
characteristics of a direct injection stratified charge rotary engine using
spark ignition and pilot flame ignition systems. SAE paper 2002-32-1791;
2002.
[11] Amrouche F, Erickson PA, Varnhagen S, Park JW. An experimental study of a
hydrogen-enriched ethanol fueled Wankel rotary engine at ultra lean and full
load conditions. Energy Convers Manage 2016;123:174–84.
[12] Agarwal AK, Dhar A, Gupta JG, et al. Effect of fuel injection pressure and
injection timing of Karanja biodiesel blends on fuel spray, engine performance,
emissions and combustion characteristics. Energy Convers Manage
2015;91:302–14.
[13] Fu JQ, Liu JP, Feng RH, Yang YP, et al. Energy and exergy analysis on gasoline
engine based on mapping characteristics experiment. Appl Energy
2013;102:622–30.
[14] Yang Bo, Xi Chengxun, Wei Xing, et al. Parametric investigation of natural gas
port injection and diesel pilot injection on the combustion and emissions of a
turbocharged common rail dual-fuel engine at low load. Appl Energy
2015;143:130–7.
[15] Wang Ying, Zhao Yuwei, Xiao Fan, Li Dongchang. Combustion and emission
characteristics of a diesel engine with DME as port premixing fuel under
different injection timing. Energy Convers Manage 2014;77:52–60.
[16] Thangavel V, Momula SY, Gosala DB, et al. Experimental studies on
simultaneous injection of ethanol–gasoline and n-butanol–gasoline in the
intake port of a four stroke SI engine. Renew Energy 2016;91:347–60.
523
[17] Lakshmanan T, Nagarajan G. Experimental investigation of port injection of
acetylene in DI diesel engine in dual fuel mode. Fuel 2011;90:2571–7.
[18] Maki T, Moriyoshi Y, Okui N, Tabata M. PIV measurement of flow field in a
rotary engine. J Visual Soc Jpn 2008;28(1):205–6 [in Japanese].
[19] Kawahara Nobuyuki, Tomita Eiji, Hayashi Kenta, et al. Cycle-resolved
measurements of the fuel concentration near a spark plug in a rotary engine
using an in situ laser absorption method. Proc Combust Inst 2007;31
(2):3033–40.
[20] Karatsu Y, Minota S, Hashimoto H, et al. Simultaneous observation of
combustion in optical rotary engine by bottom view and side view. SAE
paper 2015-01-1891; 2015.
[21] Spreitzer J, Zahradnik F, Geringer B. Implementation of a rotary engine
(Wankel engine) in a CFD simulation tool with special emphasis on
combustion and flow phenomena. SAE paper 2015-01-0382; 2015.
[22] Jeng D-Z, Hsieh M-J, Lee C-C, Han Y. The numerical investigation on the
performance of rotary engine with leakage, different fuels and recess sizes. SAE
2013-32-9160; 2013.
[23] Fan Baowei, Pan Jianfeng, Tang Aikun, et al. Experimental and numerical
investigation of the fluid flow in a side-ported rotary engine. Energy Convers
Manage 2015;95:385–97.
[24] Fan BW, Pan JF, Pan ZH, et al. Effects of pocket shape and ignition slot locations
on the combustion processes of a rotary engine fueled with natural gas. Appl
Therm Eng 2015;89:11–27.
[25] Ansys Inc. ANSYS FLUENT release 14.0. Canonsburg, PA 15317; 2011.
[26] Fan Baowei, Pan Jianfeng, Yang Wenming, et al. Effects of hydrogen blending
mode on combustion process of a rotary engine fueled with natural gas/
hydrogen blends. Int J Hydrogen Energy 2016;41:4039–53.
[27] Han Z, Reitz RD. Turbulence modeling of internal combustion engines using
RNG-k-e models. Combust Sci Technol 1995;106(4–6):267–95.
[28] Yang B, Pope SB. An investigation of the accuracy of manifold methods and
splitting schemes in the computational implementation of combustion
chemistry. Combust Flame 1998;112(1–2):16–32.