close

Вход

Забыли?

вход по аккаунту

?

j.fuproc.2018.07.002

код для вставкиСкачать
Fuel Processing Technology 179 (2018) 238–249
Contents lists available at ScienceDirect
Fuel Processing Technology
journal homepage: www.elsevier.com/locate/fuproc
Research article
Prediction of wall impingement in a direct injection spark ignition engine by
analyzing spray images for high-pressure injection up to 50 MPa
Junkyu Parka, Taehoon Kimb, Donghwan Kimb, Sungwook Parkb,
a
b
T
⁎
Research Institute of Industrial Science, Hanyang University, Seoul 04763, Republic of Korea
School of Mechanical Engineering, Hanyang University, Seoul 04763, Republic of Korea
A R T I C LE I N FO
A B S T R A C T
Keywords:
Direct-injection spark-ignition
Fuel film
Injection strategy
Spray development
Wall impingement
This study was performed to analyze the wall impingement and fuel film formation in a DISI engine with injection strategies using image-based analysis and CFD. The direct injection engine uses a high-pressure injection
strategy to improve the homogeneity of the air-fuel mixture, so the spray behavior was analyzed by spray
visualization for various injection pressures, and the wall impingement was predicted for various engine operating conditions based on the acquired images. The mass distribution of the injected fuel was calculated using
the injection profiles and the spray image, and the amount of fuel that impinges on the piston and wall (i.e., the
geometric boundaries of the cylinder) was calculated using data from the spray behavior for various engine
operation conditions such as load and engine speed. The image-based analysis was limited to understanding the
influence of the injection strategy on the droplet behavior after wall impingement of the fuel spray. Therefore,
CFD using KIVA 3 V code was additionally conducted to analyze the effects of the injection strategies on wall film
formation and droplet rebounding reflecting in-cylinder conditions. In the early- and late-injection conditions,
the initial piston position is high, and most of the injected fuel impinges on the piston. As the injection pressure
increases, the injection timing at which wall impingement occurs is advanced because of the rapid spray development. The results of the 3D analysis for the temperature and the intake flow in the engine cylinder showed
that both the wall impingement and the fuel film were reduced as the injection pressure increased because the
fuel evaporation increased due to improved atomization.
1. Introduction
Gasoline engines are widely used as power sources for automobiles
today because of their quiet operation and high-power output. Injection
systems for gasoline engines have evolved into today's DISI (direct injection spark ignition) engines, in which fuel is supplied into the cylinder directly from the carburetor system through the port injection
system. A three-way catalyst has been developed that helps to reduce
the exhaust emissions of premixed combustion as a response to environmental pollution problems caused by exhaust emissions of vehicles. To maximize the exhaust emission reduction efficiency of the
three-way catalyst, theoretical air-fuel ratio combustion techniques
were generalized, and the existing carburetor was replaced by a fuel
injector for precise fuel mass control. In the PFI (port fuel injection)
injection system, the fuel spray is injected into the intake port and it is
targeted at the high temperature intake valve to increase the homogeneity of the mixture in a short period of time. The direct injection
system has been introduced to improve the efficiency of the gasoline
⁎
engines because of global warming caused by greenhouse gas emissions
from automobiles and fuel efficiency issues. When the fuel is directly
injected into the cylinder, the latent heat of evaporation is absorbed
from the intake air in the cylinder, so that the thermal efficiency can be
increased by applying a higher compression ratio. However, the uniformity of the mixture can deteriorate in the direct injection system as
compared with the port injection system in which the vaporized fuel is
supplied to the cylinder by being mixed with the intake air [1,2]. In the
direct injection engine, fuel evaporation and mixture formation should
sufficiently occur until the end of the compression stroke, wherein
spark ignition occurs. The fuel injection strategy is important in the
DISI engine because the atomization performance determines the exhaust emission characteristics of the engine. To maximize the quality of
the air-fuel mixture in a short period of time, the fuel injection pressure
of direct injection engine is higher than that of the port injection
system, and the injection pressure of the gasoline direct injection engine
has been continuously increased to satisfy emission regulations. PN
(particulate number) is a huge emission problem because of
Corresponding author at: School of Mechanical Engineering, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul 04763, Republic of Korea.
E-mail address: parks@hanyang.ac.kr (S. Park).
https://doi.org/10.1016/j.fuproc.2018.07.002
Received 19 May 2018; Received in revised form 5 July 2018; Accepted 5 July 2018
0378-3820/ © 2018 Elsevier B.V. All rights reserved.
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
Nomenclature
E
Ecrit
m
ms
CFD
DISI
KH
RT
SMD
splash Mach number
critical splash Mach number
incident mass
splashed mass
computational fluid dynamics
direct-injection spark-ignition
Kelvin-Helmholtz
Rayleigh-Taylor
Sauter mean diameter
Abbreviations
BTDC
Before Top Dead Center
development. Also, the fuel atomization and mixing performance are
improved because of larger momentum exchange between the fuel
spray and ambient gas. Schulz et al. [6] visualized the wall film formation in a constant-volume chamber using a laser induced fluorescence technique to quantitatively compare wall film formation for
various conditions because the wall film is a cause of soot and hydrocarbon emissions in the DISI engine. They reported that about half of
the injected fuel was deposited as a wall film at low wall temperature
depending on initial warm-up conditions, but the evaporation characteristics improved as temperature increased. However, a small
amount of the fuel film still existed when injection began at the early
stage of the intake stroke. As the distance between the nozzle and the
wall increased, the amount of fuel in the fuel film decreased, and the
film thickness decreased because of improved atomization with the use
of a high-pressure injection strategy. To reduce the formation of the
wall film, which determines the exhaust characteristics, an appropriate
injection strategy should be selected depending on the engine operating
conditions.
Various studies have been carried out on the spray behavior and
wall film formation when applying high pressure injection strategies
[7–9]. However, there is a limit to predicting the wall impingement
phenomena depending on the piston movement and spray development
in the cylinder under various engine operating conditions. It is possible
to visualize in-cylinder spray behavior and mixture formation processes
using an optical engine. However, it is hard to take clear spray images
inhomogeneous mixtures in gasoline direct injection engines [3,4].
Various attempts have been made to reduce the fine particles such as
ultra-high-pressure injection strategies and attaching gasoline particulate filters. When the fuel injection pressure is increased, the momentum exchange between the fuel and the intake air occurs quickly
because of the rapid spray development. A longer time for mixing can
occur until ignition after the end of injection because of the shorter
injection duration. However, in direct injection engines, the fuel wall
film, which adversely affects the exhaust performance, can form on the
top of the piston or at the cylinder liner because of the rapid spray
development when using a high-pressure injection strategy. Therefore,
the injection strategy in direct injection engines is very important, and
it should be able to maximize mixing performance and prevent formation of the fuel wall film. A variety of studies have been carried out
on the injection strategy and wall film formation in gasoline engines.
Lee et al. [5] analyzed the macroscopic spray behavior and atomization performance of the multi-hole DISI gasoline injector using a
high-pressure injection strategy through spray visualization, droplet
size and velocity measurements using a PDPA (phase Doppler particle
analyzer) system. Spray, air entrainment, branch-like structures and
droplet detachment are observed in the atomization process of gasoline.
In addition, jet-to-jet interactions increase as injection pressure rises in
multi-hole sprays. Measurements of droplet size and velocity using
PDPA equipment have shown that the spray head period becomes
shorter as the injection pressure increases because of rapid spray
Common rail
Pneumatic pump
Fuel Chamber
HASKEL PUMP
HSF-300
Test GDI injector
HVC-SL
HVC-SL
Fuel tank
Compact RIO
Metal halide lamp
Metal halide lamp
Data Storage
High speed
camera
Fig. 1. Experimental apparatus for spray visualization.
239
Signal line
Fuel line
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
Nozzle tip
Nozzle tip
Spray cone angle
Spray image
Spray tip penetration ( )
Fig. 2. Detected spray boundary and the definitions of the spray development indices.
Total area : injection quantity, m
Mass flow rate (mg/ms)
Steady-state flow rate
Sectioned area
: injected fuel mass of each spray image,
= injection quantity / number of image frame
¢
Injection delay
£
¤
¥
¦
Spray image interval
Injection duration
Time after start of energization (ms)
¢
Step
£
¤
¥
¦
Injected fuel quantity
Injection profile
Spray image
Fig. 3. Spray development process with time during fuel injection.
depending on the Weber number and wall temperature [10]. CFD is a
useful tool for analyzing the complex physics in the engine. Therefore,
numerical analysis using CFD was carried out to investigate the spraywall interaction under real engine operating conditions.
because of the design constraints and the engine geometry. Also, a
system for observing wall impingement takes considerable effort and
cost to build. In this study, spray behavior as a function of various injection pressures was analyzed through visualization experiments in the
atmosphere using a multi-hole DISI gasoline injector, and the wall
impingement was analyzed from the acquired spray images using the
cylinder boundaries and the piston position for various engine operation conditions. Based on the high-speed spray images, spray behavior
was analyzed in terms of spray tip penetration, spray cone angle and
spatial distribution, and the mass distribution of fuel of the spray was
defined using the injection profiles and spray boundaries. In engine
operating conditions, a lot of physical phenomena affect spray-wall
interactions (e.g., turbulent intake flow and vaporization). Turbulent
intake flow reduces the momentum of the spray or guides droplets to
the wall. High ambient temperature conditions due to residual gas accelerated vaporization and reduced impacting fuel mass. When the fuel
impacts on the wall, some of the impinged droplets formed a fuel film
on the wall, and the others rebound without changing size or splash
2. Methodologies
2.1. Experimental apparatus and conditions
Spray visualization was conducted to analyze the effects of the injection pressure on the spray development and wall-impingement. The
test injector was a side-mount type multi-hole DISI gasoline injector,
which has 6 holes designed for a 33 MPa injection system. The injector
driving signal was generated using a LabVIEW-based imbedded control
system (NI, Compact RIO). N‑heptane was used as the test fuel, and the
fuel was pressurized up to 50 MPa using a pneumatic pump (Haskel,
ASFD-202). To obtain the mass distribution of the injected fuel, scattered light from the fuel droplets was acquired using two metal-halide
240
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
Ai : Area of the region i, mi : fuel mass of the region i
Fig. 4. Method of image-based analysis to predict the wall-impinged fuel mass.
120
Table 1
Engine specifications and analysis conditions to predict spray wall impingement.
Analysis conditions
Bore
Stroke
Injection pressure
Injection quantity
Engine speed
Injection timing
Spray tip penetration (mm)
Engine specifications
110
75 mm
84 mm
5, 10, 15, 25, 33, 50 MPa
14 mg (middle load)
25 mg (high load)
1200, 3600 rpm
50–350 BTDC degree
Table 2
Simulation conditions for spray-wall interaction analysis in the engine.
100
90
80
EXP.
SIM.
Injection presure
15 MPa
25 MPa
33 MPa
50 MPa
70
60
50
40
30
20
10
Engine speed
Intake pressure
Wall temperature
Fuel temperature
Injection mass
Injection pressure
Injection timing
1200, 3600 rpm
0.072 MPa
Head
Liner
Piston
363 K
14 mg
15, 25, 33, 50 MPa
130–350 BTDC degree (interval:
0
0.0
390 K
360 K
360 K
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
Time after start of energizing (ms)
Fig. 5. Spray tip penetration comparison for various injection pressures.
between injections. The injector driving system was synchronized with
the high-speed camera to take the spray images at the same time the
injector was energized. The detailed experimental setup is shown in
Fig. 1.
Spray visualization was performed on various injection pressure
conditions from 5 MPa to 50 MPa to analyze the effect of injection
pressure on spray development and wall-impingement in the side
20)
lamps and a high-speed camera operating at a speed of 10,000 frames
per seconds. The experiment was performed at room temperature and
atmospheric ambient pressure, and the spray images of each injection
condition were taken 20 times and overlapped to reduce deviation
241
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
28
24
Pinj = 15 MPa
EXP.
SIM.
20
30mm
40mm
50mm
24
16
12
8
4
0
0.0
0.4
0.8
1.2
1.6
2.0
Time after Start of Energizing (ms)
Pinj = 33 MPa
30mm
40mm
50mm
12
8
0
0.0
2.4
EXP.
SIM.
0.4
0.8
1.2
1.6
2.0
Time after Start of Energizing (ms)
2.4
(b) Injection pressure: 25 MPa
28
Legend for color Legend for symbol
20
30mm
40mm
50mm
24
Pinj = 50 MPa
Legend for color Legend for symbol
EXP.
SIM.
20
16
SMD ( m)
SMD ( m)
EXP.
SIM.
4
28
12
8
4
0
0.0
Legend for color Legend for symbol
16
(a) Injection pressure: 15 MPa
24
Pinj = 25 MPa
20
SMD ( m)
SMD ( m)
28
Legend for color Legend for symbol
30mm
40mm
50mm
16
12
8
4
0.4
0.8
1.2
1.6
2.0
Time after Start of Energizing (ms)
0
0.0
2.4
(c) Injection pressure: 33 MPa
0.4
0.8
1.2
1.6
2.0
Time after Start of Energizing (ms)
2.4
(d) Injection pressure: 50 MPa
Fig. 6. SMD comparison for various injection pressures.
distributed in the portion corresponding to the added spray area because of the supply of fuel.
The process of defining the fuel quantity distribution in the spray
using the spray images that developed with time is shown in Fig. 3. If
the total injection amount is m and the number of image frames photographed during injection is n, the injection amount in the first image
(tasoi = 0.1 ms) is m/n, and the spray image at this time is ① of Fig. 3.
The spray image in the next frame is as shown in ② of Fig. 3, where the
spray in the first image develops into the extended region, and the fuel
injected in the second injection section of the injection profile is assumed to be in the existing first spray position (tasoi = 0.2 ms). This step
was repeated until the end of the injection to calculate the fuel distribution in the spray over the injection duration. After this, the amount
of impingement fuel was calculated from the area where the spray
overlaps the cylinder wall surface and the piston based on the cylinder
geometry information such as bore, stroke, injector mounting angle and
the cylinder boundary conditions including the position of the piston
(Fig. 4). This computation was performed using MATLAB code to predict the wall-impingement for various injection pressures, load, injection timings, and engine speed conditions. The detailed engine specifications and analysis conditions for wall-impingement are shown in
Table 1.
This method has limitations in that it cannot reflect some physical
conditions in actual engine systems such as in-cylinder temperature and
intake flow. In addition, there is a need to discuss wall-wetting and
rebounding phenomena that occur after wall-impingement in engines to
analyze the effects of wall-impingement on droplet atomization.
mount type DISI engine. For a DISI gasoline engine with a 398 ml displacement volume, the fuel injection quantity was fixed to 14 mg (for
middle load conditions) and 25 mg (for high load condition) regardless
of the injection pressure conditions to analyze the wall-impingement for
various engine operating conditions such as engine speed and load.
2.2. Image-based prediction of wall impingement
MATLAB image processing code was used to analyze the effect of
injection pressure on the injection duration and spray development.
The distance from the nozzle tip to the spray boundary at the farthest
position on the plane was defined as the spray tip penetration. In the
first 1/3 of the spray tip penetration, the spray is not significantly affected by air entrainment, and this region has very straight spray.
Therefore, this region can be assumed to be an arc centered at the
nozzle tip. The spray cone angle was calculated from the area and
length of the arc. The definitions of the spray development indices are
shown in Fig. 2.
The amount of fuel injected through the injector nozzle per unit
time of 0.1 ms (which is the interval of image acquisition of the fuel
spray) can be obtained using the measured injection quantity results
and assuming that the injection profile for each injection condition was
a square wave. That is, the amount of fuel injected per unit time is
calculated by dividing the injection quantity for each condition by the
number of image frames in which the injection occurs (Fig. 3). It can be
assumed that the fuel injected per unit time is evenly distributed in the
region of the developed spray observed in each image frame acquired
using the high-speed camera. Looking at the images of the spray that
developed over time, it can be seen that the area of the spray increased
gradually with the fuel being supplied during the injection duration
with a constant spray angle. In this case, the mass distribution of the
spray in the frame can be calculated assuming that the fuel is uniformly
2.3. Numerical analysis for real engine operating conditions
KIVA-3V release 2 code was used to analyze the spray-wall interactions in an engine. In-cylinder turbulent flow was predicted by the
242
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
6.0
Yarin and Weiss [15].
Injection duration (ms)
5.5
Injection quantity
14mg
25mg
5.0
1.8·10−4 (E 2 − Ecrit 2) for Ecrit 2 < E 2 < 7500 ⎫
ms
=⎧
⎬
⎨
0.75
for E 2 > 7500
m
⎭
⎩
4.5
Although this model does not reflect the effects of wall temperature
on spray-wall interaction, this study focuses on the effects of intake flow
and vaporization characteristics on spray-wall interactions in engine
operating conditions, and reliable results can be achieved using this
model.
For DISI gasoline engines, fuel can be injected with a fuel temperature of −10 °C in cold start conditions to over 150 °C in high-load
conditions (measured at the tip of injector) [16]. When the gasoline is
heated to 60 °C, flash boiling spray can form. In this study, the flash
breakup model developed by Kim and Park [17] was used to predict
flash breakup. Vaporization of fuel was calculated based on the energy
valence equation [18], and the superheat effect suggested by Adachi
et al. [19] was also included.
Spray analysis for varying injection pressure with the same injection
quantity was carried out in a constant volume chamber to validate the
breakup model. The grid size of the constant volume chamber was set to
2.5 mm which is the average cell size of engine mesh. Predicted spray
tip penetration and SMD were compared with experimental results.
Droplet diameter was measured using PDPA (Phase Doppler particle
analyzer) system. For the verification of the analytical results, 10,000
droplets data per condition were obtained under room temperature and
atmospheric pressure conditions, and it was averaged every 0.25 ms.
SMD results of the experiment and analysis were compared at 30, 40
and 50 mm from the nozzle along the nozzle direction. A DISI gasoline
engine with a 398 ml displacement volume was used for simulations.
Maximum cell size is 4.0 mm and minimum cell size is 0.5 mm approximately. Average cell size is 2.5 mm approximately, which is the
common cell size used in engine simulation. Total number of cells including intake and exhaust ports is 62,000 at BDC position. The injector
is mounted at the side of the engine. To reflect the in-cylinder condition
filled with exhaust gas, simulation started from the exhaust process. Incylinder gas prior to exhaust valves open (EVO) was assumed to be
completely burned gas in stoichiometric condition with temperature of
1400 K and pressure of 0.29 MPa. This condition was achieved by the
previously conducted combustion simulation with same fuel mass and
stoichiometric ratio. Since CFD is time-consuming-work, simulation
conditions were reduced from the experimental study. Engine speed
was set to 1200 and 3600 rpm. Fuel injection mass was set to 14 mg,
and the intake pressure was set to 0.072 MPa, which corresponds to
stoichiometric conditions for the given injection quantity. The liner and
piston were set to 360 K, and the head was set to 390 K. Fuel was injected at 350° BTDC to 50° BTDC at pressures of 15, 25, 33 and 50 MPa.
Table 2 summarizes the simulation conditions for spray-wall interaction
analysis in the engine.
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0
5
10
15
20
25
30
35
40
45
50
55
Injection pressure (MPa)
(a) Injection durations for the various injection pressures
42
41
Spray cone angle (o)
40
39
38
37
36
35
34
33
32
0
5
10
15
20
25
30
35
40
45
50
55
Injection pressure (MPa)
(b) Spray cone angles for the various injection pressures
120
110
Spray tip penetration (mm)
100
90
80
70
60
50
40
30
20
10
0
Solid line: 14mg
Dashed line : 25mg
Injection pressure
5 MPa
15 MPa
25 MPa
33 MPa
50 MPa
3. Results and discussion
-10
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
3.1. Spray breakup model validation for CFD
Time after start of energization (ms)
(c) Spray tip penetrations for the various injection pressures
The spray breakup model was validated with experimental results
prior to analysis of the spray-wall interactions in a DISI gasoline engine.
A spray tip comparison was carried out for the fuel injection period.
Fig. 5 shows the spray tip penetration comparison between the simulation and experiment for various injection pressures. As the injection
pressure increased, injection duration decreased for injecting the same
amount of fuel. Spray tip penetration increased with increasing injection pressure. However, the spray reached almost the same distance
when fuel injection ended. Simulations showed similar results to the
experiments for all injection pressure conditions.
Fig. 6 shows a SMD comparison between simulation and experimental results. As the distance from the nozzle increased, the SMD
Fig. 7. Comparisons of the spray development for the various injection conditions.
RNG k-ε model modified by Han and Reitz [11]. The KH-RT breakup
model was used for predicting spray breakup [12]. Droplet dispersion
by turbulent flow was predicted using a model by O'Rourke [13]. Spraywall interaction modeled by O'Rourke and Amsden are implemented in
the KIVA code [14]. When the splash Mach number exceeded a critical
value, droplet was assumed to splash. For droplets predicted to splash,
only a part of the incident mass is splashed based on the experiment by
243
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
Injection Pressure
5 MPa
15 MPa
25 MPa
33 MPa
50 MPa
Injection Quantity
14
mg
WDVRH PV
PV
PV
PV
PV
WDVRH PV
PV
PV
PV
PV
WDVRH PV
PV
PV
PV
PV
WDVRH PV
PV
PV
PV
PV
25
mg
14
mg
25
mg
Fig. 8. Spray development process and the final spray images at the end of the injection for various injection pressures.
condition (an injection quantity of 25 mg), the spray tip penetration is
slightly reduced at the lower injection pressures because the time
during which the spray tip receives drag from the surrounding gas increased.
Fig. 8 shows the spray development process and the final spray
image at the end of the injection when the same amount of fuel (corresponding to middle load and high load) is injected under various
injection pressure conditions. At the end of the injection, the spray tip
penetration is not affected by the injection pressure, but the spray cone
angle increases because of the higher turbulent kinetic energy of the
fuel supplied as the injection pressure increases. Looking at the macroscopic spray shape analyzed using the spray boundary detection, the
vertical flow direction of the spray changes as the injection pressure
increases. Under low injection pressure conditions, the drag force that
the spray droplet receives from the ambient gas is small because of the
small nozzle outlet velocity. So, the area occupied by the spray is small,
and spray flow is formed in the direction of each nozzle hole. At higher
injection pressure conditions, the drag force that the fuel spray receives
from the surrounding gas increases because of the increase in the droplet velocity, and the air-entrainment increases because of the large
shear force formed during spray development. As a result of these
phenomena, the spray has a larger spray angle, and the area occupied
by the spray is expanded as the injection pressure increases. In particular, vertical spray plumes are most affected by air-entrainment with
increasing injection pressure, which contributes greatly to the expansion of the spray area. That is, the injection pressure affects the spray
behavior in terms of spray tip penetration, spray cone angle, and spray
area. Since the injection pressure affects the fuel distribution in the
cylinder, the fuel injection pressure makes a difference in wall-
increased. This is clearly seen for low injection pressure cases. At high
injection pressures, the SMD predicted by the simulation showed similar values regardless of the distance from the nozzle. However,
overall trends of increasing injection pressure were predicted well with
the spray breakup model.
3.2. Effects of injection pressure on the spray development
To analyze the effect of the injection pressure on the spray behavior,
the injection duration, spray cone angle and spray tip penetration for
various injection pressures are compared under identical fuel quantity
conditions (Fig. 7). As the injection pressure increased, the injection
duration decreased drastically, and the injection durations of the
50 MPa conditions were reduced 30% compared to that of 5.0 MPa.
This tendency was consistent regardless of the injection quantity. In the
DISI gasoline injector, the change in the injection duration is not significantly affected by the injection quantity because the injection profile is close to a square wave. On the other hand, the effect of the injection pressure on the spray tip penetration at the end of the injection
was insignificant. The larger momentum supply causes the spray to
develop rapidly at higher injection pressures, but it has a shorter injection duration. At low injection pressure conditions, the momentum
supply from the nozzle outlet is small, but the injection duration becomes long to inject the same amount of fuel. Average spray cone angle
during the injection period is compared in Fig. 7(b). Spray cone angle
increased with injection pressure, but the increase of spray cone angle
decreased under higher pressure injection condition. Fig. 7(c) show that
the final spray tip penetrations at the end of the injection are similar
regardless of the injection pressures. In the high load injection
244
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
15
that of injected fuel are opposite, which results in a greater gradient of
impinged fuel quantity with time.
The early injection conditions occur during the intake stroke, in
which the piston descends. In this region, the development direction of
the spray and the motion direction of the piston coincide with each
other. The wall impingement is determined by the development speed
of the spray and the speed of the piston. At low speed, i.e., middle load
conditions, the effect of the injection pressure on wall impingement is
not significant because the space occupied by the spray in the cylinder
is not large, and the piston displacement during the injection duration is
also small. In the early stage of the intake stroke, the results show that
there is no significant difference in wall impingement depending on the
injection pressure except at 5.0 MPa. On the other hand, when the fuel
is injected during the compression stroke, the effects of the injection
pressure on wall impingement are remarkably observed. The wall impingement of the fuel spray consistently decreased as the injection
pressure increased. Lower injection pressures under the same injection
timing condition resulted in longer piston displacement during the injection period. So, a small amount of spray impinges on the piston
under the high-pressure injection condition with a short injection
period.
The engine load and speed are the most dominant factors affecting
the wall impingement of the fuel spray among the engine operating
parameters. The differences in fuel injection quantity with engine load
cause a change in the spray tip penetration and the size of the space
where the spray is located in the cylinder at the end of injection. The
speed of the engine is also an important driving variable related to wall
impingement since it determines the piston displacement during fuel
injection. The wall impingement on the piston and cylinder liner are
compared depending on the injection timing at various operating conditions (fuel quantity, engine speed) to analyze the effect of injection
pressure on wall impingement in a high-pressure direct injection system
(Fig. 10). At low speed and middle load conditions, the injection
pressure does not affect the wall impingement significantly in the case
of early-injection during the intake stroke, but there is a significant
difference in the other operating conditions. In the case of early-injection, the amount of impinged fuel mass increases with injection pressure, and this tendency is observed when the load increased or the
engine speed increased. When the engine load increases (i.e., when the
fuel injection quantity increases, Fig. 10(b)), the injection timing where
spray wall impingement occurs became wider because the spray tip
penetration increased at the end of the injection because of the longer
injection period. The spray area in the cylinder also expanded. When a
large amount of fuel is injected at a high pressure, the spray is transferred in the direction of the cylinder liner because of the strong momentum of the fuel spray, and part of the fuel spray impinges on the
cylinder liner as well as the piston.
Changes in engine speed have a greater impact on wall impingement than changes in load (Fig. 10(c)). At the high-speed condition, the
effect of the injection pressure on the impinged fuel mass shows the
same tendency as the low speed condition. The difference in the amount
of impinged fuel mass among the injection pressures becomes larger,
but the injection timing where the wall impingement occurs is similar
regardless of engine speed. At high speed conditions, the injection
timing where the wall impingement occurs is advanced because the
piston speed is faster. Wall impingement does not occur at a 5 MPa
injection pressure because the piston speed is faster than the spray
development when the fuel is injected during the early stage of the
intake stroke, whereas wall impingement occurs at all injection timing
conditions when the fuel is injected during the compression stroke. The
injection timing at which the wall impingement occurs at high speed is
advanced at all injection pressures, although there is a difference in
degree. Therefore, the injection timing should be advanced to reduce
wall impingement. As the engine speed increases, the injection timing
where the wall impingement occurs is advanced, and the non-impingement-timing is reduced when the load increases. These patterns
Pinj : 5.0 MPa, 1200 RPM, mf : 14 mg
Impinged fuel mass (mg)
12
9
Injection timing
ATDC -270o
ATDC -290o
ATDC -310o
ATDC -330o
ATDC -350o
6
3
0
-3
-0.5
0.0
0.5 1.0 1.5 2.0 2.5 3.0
Time after start of energization (ms)
3.5
4.0
3.5
4.0
(a) Intake stroke
15
Pinj : 5.0 MPa, 1200 RPM, mf : 14 mg
Impinged fuel mass (mg)
12
9
Injection timing
ATDC -50o
ATDC -70o
ATDC -90o
ATDC -110o
ATDC -130o
6
3
0
-3
-0.5
0.0
0.5 1.0 1.5 2.0 2.5 3.0
Time after start of energization (ms)
(b) Compression stroke
Fig. 9. Comparisons of impinged fuel mass with time for various injection
timings.
(Pinj: 5.0 MPa, Engine speed: 1200 RPM, mf: 14 mg).
impingement under actual engine operating conditions.
3.3. Wall impingement
Fig. 9 shows the spray wall impingement process for various injection timings under low speed and middle load conditions. When
injecting fuel during the intake stroke, the total amount of fuel impinging on the piston increases as the injection timing is advanced, and
the amount of impinged fuel increases linearly with time during the
injection period. During the intake stroke, the timing at which the spray
starts to impinge on the piston is advanced as the fuel is injected earlier
because of the higher initial piston position. Near the TDC, the piston
velocity is slow, but the spray velocity is constant irrespective of the
injection timing. Therefore, the gradient of the impinged fuel quantity
with time becomes larger as the injection timing is advanced (Fig. 9(a)).
In the early injection of fuel during the intake stroke, the initial piston
position is lowered as the injection timing is retarded, and a region
where spray impingement does not occur appears. Thereafter, wall
impingement of the spray is observed under the late injection timing
condition for injecting fuel during the compression stroke, and the
amount of the impinged fuel increases because of the higher initial
piston position as the injection timing is advanced (Fig. 9(b)). During
the compression stroke, the relative velocity between the fuel spray and
the piston is greater because the direction of motion of the piston and
245
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
Intake valve open
Intake valve open
18
16
Injection quantity 14mg, 1200 RPM
Injection pressure
5MPa
15MPa
25MPa
12
16
33MPa
50MPa
Impinged mass (mg)
Impinged mass (mg)
14
10
8
6
4
2
0
-2
-360
14
Injection quantity 14mg, 3600 RPM
Injection pressure
5MPa
15MPa
25MPa
50MPa
10
8
6
4
2
0
-330
-300
-270
-240 -210 -180 -150 -120
Injection timing (degree ATDC)
-90
-60
-30
-2
-360
0
-330
-300
-270
-240 -210 -180 -150 -120
Injection timing (degree ATDC)
-90
-60
-30
0
-60
-30
0
(b) Middle load, High speed
(a) Middle load, Low speed
Intake valve open
Intake valve open
35
35
Injection quantity 25mg,1200 RPM
Injection pressure
5MPa
15MPa
25MPa
Injection quantity 25mg, 3600 RPM
Injection pressure
5MPa
15MPa
25MPa
30
33MPa
50MPa
Impinged mass (mg)
30
Impinged mass (mg)
33MPa
12
25
20
15
10
5
0
33MPa
50MPa
25
20
15
10
5
0
-5
-360
-330
-300
-270
-240 -210 -180 -150 -120
Injection timing (degree ATDC)
-90
-60
-30
0
-5
-360
-330
-300
-270
-240 -210 -180 -150 -120
Injection timing (degree ATDC)
-90
(d) High load, High speed
(c) High load, Low speed
Fig. 10. Comparisons of impinged fuel mass for various engine operating conditions.
100
90
80
80
Percentage (%)
Injection Pressure:
15 MPa
40
Injection Mass:
14 mg
30
Engine speed:
3600 rpm
70
60
50
Fuel Film
Rebounded Fuel
Vaporized Fuel
90
Engine Speed:
1200 rpm
70
Percentage (%)
100
Fuel Film
Rebounded Fuel
Vaporized Fuel
60
Injection Pressure:
15 MPa
50
40
Injection Mass:
14 mg
30
20
20
10
10
0
360 340 320 300 280 260 240 220 200 180 160 140 120
0
360 340 320 300 280 260 240 220 200 180 160 140 120
Injection Timing (BTDC degree)
Injection Timing (BTDC degree)
Fig. 11. Probability of fuel impingement for various injection timings.
(Engine speed: 1200 rpm, injection pressure: 15 MPa, injection mass: 14 mg).
Fig. 12. Probability of fuel impingement for various injection timings.
(Engine speed: 3600 rpm, injection pressure: 15 MPa, injection mass: 14 mg).
are consistent under high speed and high load conditions. The amount
of wall impingement is reduced early in injection (during the intake
stroke) because of the rapid piston motion, and the impinged fuel mass
increases in the late-injection during the compression stroke. At an
injection pressure of 5 MPa, wall impingement does not occur in the
early stage of the intake stroke, but it occurs from the end of the intake
stroke, and all injected fuel impinges on the piston at 130° BTDC.
Early injection timings are used to form homogenous mixtures in
common DISI gasoline engines with side-mount-type multi-hole injectors. Therefore, early injection timings were focused on in the CFD
study. Fig. 11 shows the probability of film formation and rebounding
for various injection timings. The percentage of fuel film formation and
rebounding was counted until the droplet evaporated. Therefore, if the
rebounded droplets after first impingement formed a fuel film, they
were counted as part of the film mass. The percentage of fuel film and
rebounded fuel indicates the tendency of the fuel to form a fuel film and
rebound over the lifecycle of the droplet. The percentage of vaporized
fuel indicates the part of the droplet that evaporated without impinging
on the wall. The sum of the percentage of fuel film and rebounded fuel
can be regarded as the impingement percentage. When the fuel is injected early, the impingement mass increased. This is because the
piston is located near TDC, and the small in-cylinder volume increases
the probability of impingement. However, over the half of the injected
fuel vaporized without impinging on the in-cylinder wall. The engine
used in this study has no valve overlap. Therefore, a high in-cylinder
pressure and temperature condition was observed near the TDC. This
in-cylinder condition promotes atomization by increasing the vaporization rate. When the fuel is injected after BTDC 290°, the in-cylinder
volume increased, and intake flow developed inside cylinder to prevent
fuel impingement on the in-cylinder wall.
Fig. 12 shows the probability of fuel impingement for various injection timings with an engine speed of 3600 rpm. Fuel impingement
for early injection timing was reduced compared to 1200 rpm. As the
engine speed increased, the piston was located farther down during the
246
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
100
80
Engine speed:
1200 rpm
70
Percentage (%)
Fig. 13 show the effect of injection pressure on fuel impingement on
the wall when fuel is injected at BTDC 350°. As the injection pressure
increased, evaporation rate increased, and the percentage of fuel impingement decreased. As the injection pressure increased, the fuel film
mass decreased due to a decrease in impingement mass. This is the
opposite result from the experimental results. The main difference that
caused the opposite trend was the effect of the evaporation of the fuel.
Therefore, an analysis without evaporation was carried out to clarify
the effects of fuel evaporation on impingement.
Fig. 14 shows the impingement mass for various injection pressures
evaluated with the same criteria as the experimental study. The results
for 3600 rpm were used to investigate the effects of injection pressure
on fuel impingement since the difference in the impinging mass for
various injection pressures was higher when the engine speed was high.
The first impingement on the wall was counted. When the effects of fuel
evaporation are neglected, the impingement mass increased as the fuel
injection pressure increased. Also, the fuel film mass increased as the
impingement mass increased. The higher injection pressure case injected fuel over a short duration. The high injection pressure case ends
fuel injection early, and the piston location is higher than that of the
low injection pressure case. This causes an increase in impingement
mass for the high injection pressure case when evaporation is neglected.
From the above results, the amount of liquid film decreased when the
injection pressure increased due to the increase in the evaporation rate
as the injection pressure increased.
Fig. 15 summarizes the probability of film formation for various
injection pressures and injection timings. The accumulated film mass
evaluated by counting the droplets impinging and forming fuel film was
divided by total injected fuel mass to calculate the probability of film
formation. In the probability map for film formation, higher injection
pressures for injection at BTDC 290° resulted in a greater reduction in
fuel film on the cylinder wall. The high-speed tumble flow generated at
the early stage of the intake process flowed along the in-cylinder wall
and prevented film formation. As the intake flow developed, the speed
of the tumble flow decreased. At this condition, the large droplets
generated by low injection pressure floated inside the cylinder and
formed a liquid film. This causes a slight increase in fuel film when the
fuel is injected at low injection pressure during the intake process.
Fig. 16 shows the probability density function map of the Weber
number for impinging and rebounding droplets. The cases of 350° BTDC
injection and 290° BTDC injection were chosen to represent when the
impinging amounts were the largest and smallest. The cases of 15 MPa
and 50 MPa injection were chosen to represent the high-pressure
Fuel Film
Rebounded Fuel
Vaporized Fuel
90
60
Injection Mass:
14 mg
50
40
Injection Timing:
BTDC 350 degree
30
20
10
0
15
25
33
50
Injection Pressure (MPa)
Fig. 13. Percentage of fuel impingement for various injection pressures.
(Engine speed: 1200 rpm, injection mass: 14 mg, injection timing BTDC 350°).
16
Film Mass
Rebound Mass
14
Mass (mg)
12
10
8
6
4
2
0
15
25
33
Injection Pressure (MPa)
50
Fig. 14. Impingement mass for various injection pressures at non-evaporating
conditions. (Engine speed: 3600 rpm, injection mass: 14 mg, injection timing
BTDC 350°).
fuel injection. Therefore, the impinging mass was reduced for early
injection timings with higher engine speeds. When the fuel is injected
during the intake process, there was no significant difference for different engine speeds, because the in-cylinder volume is relatively large,
and the droplets lost momentum due to impinging on the wall.
Injection pressure (MPa)
50
40
Engine speed:
1200 rpm
0
Injection Mass:
14 mg
6
12
30
18
24
20
30
340 320 300 280 260 240 220 200 180 160 140
Injection timing (BTDC degree)
Fig. 15. Probability of film formation for various injection pressures and injection timings.
(Engine speed: 1200 rpm, injection mass: 14 mg).
247
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
10000
10000
0.000
0.000
8000
7000
Weout
6000
5000
4000
3000
2000
0.000
9000
0.004000
0.008
0.008000
0.01200
0.016
0.01600
0.02000
0.024
0.02400
0.02800
0.03200
0.032
0.03600
0.04000
0.040
0.008
8000
7000
0.016
6000
Weout
9000
5000
0.024
4000
0.032
3000
0.040
2000
1000
1000
2000 4000 6000 8000 10000 12000 14000 16000
2000 4000 6000 8000 10000 12000 14000 16000
Wein
Wein
(a) Pinj = 15 M Pa, C.A.inj = BTDC 350
(b) Pinj = 15 M Pa, C.A.inj = BTDC 290
10000
10000
0.000
0.000
0.004000
0.008000
0.008
0.01200
0.01600
0.016
0.02000
0.02400
0.024
0.02800
0.03200
0.032
0.03600
0.04000
0.040
8000
7000
Weout
6000
5000
4000
3000
2000
9000
0.000
8000
0.008
7000
0.016
6000
Weout
9000
5000
0.024
4000
0.032
3000
0.040
2000
1000
1000
2000 4000 6000 8000 10000 12000 14000 16000
2000 4000 6000 8000 10000 12000 14000 16000
Wein
Wein
(c) Pinj = 50 M Pa, C.A.inj = BTDC 350
(d) Pinj = 50 M Pa, C.A.inj = BTDC 290
Fig. 16. Probability density function map of impinging and rebounding droplet Weber number.
regardless of the injection pressure.
2. Wall impingement occurs for fuel injected from an early injection
and late injection timings. The amount of impinged fuel increases as
the injection pressure increases in early-injection conditions, but it
decreases with injection pressure in the late-injection conditions.
3. As the amount of injected fuel mass increased, the injection timing
region where the wall impingement occurs expanded. When the
engine speed increased, the injection timing at which the wall impingement occurs advanced. The increase in the injection pressure
improved the evaporation due to the superior atomization performance, so that a wider injection timing can be utilized in an engine.
4. Promotion of atomization with increasing fuel injection pressure
leads to the spread of an impinging Weber number for early injection cases. When the injection timing is retarded, droplets impinge
on the wall after their momentum is dissipated, and the differences
between the injection pressures are negligible.
injection and the low-pressure injection. When the fuel is injected at
BTDC 350° at 15 MPa, the Weber number of the impinging droplet increased to approximately 3000, and that of the rebounding droplet was
2000. Droplets with high Weber numbers rebounded with low Weber
numbers. That means that the large droplets impinged on the wall and
rebounded with breakup. When the fuel is injected at 350° BTDC at a
pressure of 50 MPa, the distribution of the Weber number for the impinging droplet and rebounding droplet was widely spread. This is
because the distribution of the droplet size spread due to promotion of
atomization. The probability of an impinging droplet with a large
Weber number decreased due to promoted atomization, while the velocity of the droplet increased with increasing injection pressure. When
the fuel was injected at 290° BTDC, most of the droplets impinged on
the in-cylinder wall with a small Weber number regardless of the injection pressure. At these conditions, droplets lost momentum and
disintegrated into smaller droplets. This caused a decrease in the Weber
number of the impinging droplets.
Acknowledgements
4. Conclusions
This work was supported by the Korea Institute of Energy
Technology Evaluation and Planning (KETEP) and the Ministry of
Trade, Industry & Energy (MOTIE) of the Republic of Korea (No.
20172010105770) and the research fund of Hanyang University (HY2018).
This study was carried out to investigate the effect of injection
strategy on wall impingement in a DISI gasoline engine. The results
obtained from the study can be summarized as follows:
1. As the fuel injection pressure increases, the injection timing becomes shorter, but the spray cone angle and spray area increase due
to the supply of a large amount of turbulent kinetic energy. The
spray tip penetration is identical at the end of the injection
References
[1] Q. Liu, A. Cairns, H. Zhao, M.A. Attar, L. Cruff, H. Blaxill, The Effects of Charge
Homogeneity and Repeatability on Particulates Using the PLIF Technique in an
248
Fuel Processing Technology 179 (2018) 238–249
J. Park et al.
[10] C. Bai, A.D. Gosman, Development of Methodology for Spray Impingement
Simulation, SAE Technical Paper, SAE 950283, (1995).
[11] Z. Han, R.D. Reitz, Turbulence modeling of internal combustion engines using RNG
κ-ε models, Combust. Sci. Technol. 106 (1995) 267–295.
[12] R.D. Reitz, J.C. Beale, Modeling spray atomization with the Kelvin-Helmholtz/
Rayleigh-Taylor hybrid model, Atomization Sprays 9 (1999) 623–650.
[13] P.J. O'Rourke, Statistical properties and numerical implementation of a model for
droplet dispersion in a turbulent gas, J. Comput. Phys. 83 (1989) 345–360.
[14] P.J. O'Rourke, A.A. Amsden, A Spray/Wall Interaction Submodel for the KIVA-3
Wall Film Model, SAE Technical Paper, SAE 2000-01-0271, (2000).
[15] A.L. Yarin, D.A. Weiss, Impact of drops on solid surfaces: self-similar capillary
waves, and splashing as a new type of kinematic discontinuity, J. Fluid Mech. 283
(2006) 141–173.
[16] P.G. Aleiferis, et al., Mechanisms of spray formation and combustion from a multihole injector with E85 and gasoline, Combust. Flame 157 (2010) 735–756.
[17] T. Kim, S. Park, Modeling flash boiling breakup phenomena of fuel spray from
multi-hole type direct-injection spark-ignition injector for various fuel components,
Energy Convers. Manag. 160 (2018) 165–175.
[18] Y. Ra, R.D. Reitz, A vaporization model for discrete multi-component fuel sprays,
Int. J. Multiphase Flow 35 (2009) 101–117.
[19] M. Adachi, et al., Characterization of Fuel Vapor Concentration Inside a Flash
Boiling Spray, SAE Technical Paper, SAE 970871, (1997).
Optical DISI Engine, SAE Technical Paper, SAE 2014-01-1207, (2014).
[2] J. Song, T. Kim, J. Jang, S. Park, Effects of the injection strategy on the mixture
formation and combustion characteristics in a DISI (direct injection spark ignition)
optical engine, Energy 93 (2015) 1758–1768.
[3] J. Duan, K. Sun, L. Li, Effects of EGR on PN Emissions Under Operating Parameters
From DISI Engines, SAE Technical Paper, SAE 2017-01-2399, (2017).
[4] F.C.P. Leacha, R. Stonea, D. Richardsonb, J.W.G. Turnerc, A. Lewisc, S. Akehurstc,
S. Remmertd, S. Campbelld, R. Cracknelld, The effect of oxygenate fuels on PN
emissions from a highly boosted GDI engine, Fuel 225 (2018) 277–286.
[5] S. Lee, S. Park, Experimental study on spray break-up and atomization processes
from GDI injector using high injection pressure up to 30 MPa, Int. J. Heat Fluid Flow
45 (2014) 14–22.
[6] F. Schulz, W. Samenfink, J. Schmidt, F. Beyrau, Systematic LIF fuel wall film investigation, Fuel 172 (2016) 284–292.
[7] H. Yu, X. Liang, G. Shu, X. Sun, H. Zhang, Experimental investigation on wall film
ratio of diesel, butanol/diesel, DME/diesel and gasoline/diesel blended fuels during
the spray wall impingement process, Fuel Process. Technol. 156 (2017) 9–18.
[8] C. Rodrigues, J. Barata, A. Silva, Spray impingement modelling: evaluation of the
dissipative energy loss and influence of an enhanced near-wall treatment, Fuel
Process. Technol. 107 (2013) 71–80.
[9] M. Panao, A. Moreira, D. Durao, Statistical analysis of spray impact to assess fuel
mixture preparation in IC engines, Fuel Process. Technol. 107 (2013) 64–70.
249
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 618 Кб
Теги
fuproc, 002, 2018
1/--страниц
Пожаловаться на содержимое документа