Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Wall Impingement Process of a Multi-Hole GDI
Spray: Experimental and Numerical Investigation
2012-01-1266
Published
04/16/2012
Alessandro Montanaro
Istituto Motori CNR
Simone Malaguti
Univ of Modena & Reggio Emilia
Salvatore Alfuso
Istituto Motori CNR
Copyright © 2012 SAE International
doi:10.4271/2012-01-1266
ABSTRACT
The Direct Injection (DI) of gasoline in Spark Ignition (SI)
engines is very attractive for fuel economy and performance
improvements in spark ignition engines. Gasoline direct
injection (GDI) offers the possibility of multi-mode
operation, homogeneous and stratified charge, with benefits
respect to conventional SI engines as higher compression
ratio, zero pumping losses, control of the ignition process at
very lean air-fuel mixture and good cold starting.
The impingement of liquid fuel on the combustion chamber
wall is generally one of the major drawbacks of GDI engines
because its increasing of HC emissions and effects on the
combustion process; in the wall guided engines an increasing
attention is focusing on the fuel film deposits evolution and
their role in the soot formation. Hence, the necessity of a
detailed understanding of the spray-wall impingement
process and its effects on the fuel distribution. The
experimental results provide a fundamental data base for
CFD predictions.
In this paper investigations have been performed using a 7hole injector, 0.179 mm in hole diameter, spraying in a
constant volume vessel with optical accesses. To examine the
effects of various factors on development of the spray
impinging on the wall, experiments have been conducted at
different injection pressures, diverse wall inclination angles
and at atmospheric pressure. The acquired images have been
processed for extracting the characteristic parameters of the
impinging fuel at the different operative conditions.
The multi-hole spray has been simulated by Star-CD code
taking into account the commercial gasoline properties and
the real mass flow rate derived from experimental
measurements. In order to correctly reproduce spray
impingement and fuel film evolution, a numerical
methodology has been defined. Lagrangian sub-models and
numerical parameters have been validated against
experimental results.
INTRODUCTION
A great effort is paid by engine manufacturers in order to
gain a deep knowledge of the GDI impact on the engine
design and operation. It is well recognized the many
advantages of the direct injection of gasoline joint with the
supercharged technology in terms of fuel efficiency,
drivability and lower size/power ratios [1, 2]. The charge
cooling through in-cylinder injection and the ability to
improve scavenging at full load operations leads to an
increase in engine brake specific power. Downsized
homogeneous-charged GDI engines with boost lead to
relevant fuel savings compared to current production PFI
gasoline engine. Stratified-charge lean GDI combustion
provides even further fuel consumption reductions. In terms
of CO2 tailpipe emissions, boosted downsized stratified GDI
engines approaches turbo-diesel one for given displacement.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Since the European Commission proposal aims to reduce
average fleet CO2 emissions down to 95 g/km by 2020,
which would imply a halving of fuel consumption with
reference to 1990 and the recently increased limitations to
both NOx and Soot emissions, the engine researcher are
pushed to rediscover gasoline engines. As a consequence,
massive investments for the development of new
technologies and strategies have been made, mainly targeted
at increasing fuel conversion efficiency together with
reducing engine-out emissions. Particular care is devoted to
the study of new injection and combustion systems: the
injection pressure is constantly raising as well as new
complex control strategies are under development. As far as
the fuel spray morphology and its interaction with the
combustion system has to deeply investigated, the fuel
mixture formation within the combustion chamber is
continuously growing in importance.
Among the many complex phenomena occurring during the
fuel mixture preparation such as in-cylinder flow motion,
piston design, injector location, injection pressure, timing and
strategy and spray characteristics, the fuel/wall interaction
could play an important role. When the engine runs at very
low engine speeds and low engine loads, as the idle
condition, the reduced fuel injection pressure and the need of
particular injection timing could lead to fuel deposits on the
combustion chamber [3]. For stratified cold operation, the
spray impingement on piston may yield to film formation and
large HC and soot exhaust emissions [4]. Furthermore wall
guided engine architecture is concerned the fuel mixture
formation (e.g. stratified charge) is mainly influenced by the
spray/wall interaction.
It is well known that an impinging spray has a fine
atomization and good spatial distribution of fuel droplets.
However, both the spatial distribution of fuel and air
entrainment into the impingement spray have a great
influence on combustion characteristics. Literature reports the
impingement angle [5] and the position of the impinging
spray [6] are important factors for mixture formation near the
combustion chamber wall. Investigations by means of
experimental and numerical comparison have been performed
by many researchers. A GDI spray-wall interaction inside a
heated pressurized chamber, using various visualization
techniques, results the effects of the on the upward spray
vortex inside the spray and on the horizontal spread after
impingement [7]. The predictive capability of the CFD
impinging models have been investigated by Walkins and
Park [8] and by Bai and Gosman [9], founding that models
based on single-droplet experiments are insufficient to
calculate the spray-wall impingement process. Finally,
literature are deeply studying the impinging fuel spray for
GDI application such as spray generated by hollow cone
swirl injector [10,11,12].
In this study the fuel and wall interaction is investigated for a
spray generated by a second generation multi-hole injector of
a current production GDI engine adopting a wall guided
architecture. Injection pressures, spacing from idle engine
operating condition strategy to full load operating condition,
are investigated by means of optical apparatus. Impinging
spray is then reproduced using commercial code Star-CD,
focusing on the capability of the CFD models. This paper
resumes a preliminary activity focused on the
characterization on the multi-hole GDI injectors, in terms of
wall impact and aims to analyze the effect of the injection
pressure and the inclination angle of the impinging wall. The
aim of this work is to increase the understanding of the
complex phenomena such as the spray and wall interaction in
view of subsequent fuel mixing analysis within the
combustion chamber of a wall-guided GDI engine.
EXPERIMENTAL APPARATUS
The GDI injector considered in the present work is the minisac seven-hole Bosch HDEV 5.1 with solenoid actuation.
This injector is characterized by a hole diameter of 0.179 mm
and with static flow rates of 13.7 g/s at the injection pressure
of 10 MPa. On a plane perpendicular to the injector axis
placed at a distance of 30 mm from the holes themselves, the
seven jet directions give a spray footprint characterized by a
hollow-ellipsoid shape (Figure 1).
Figure 1. Holes distribution and spray footprint on a
plane placed at 30 mm from the injector tip.
Commercial gasoline has been used (ρ=740 kg/m3), within all
the experimental campaign, as delivered by a hydropneumatic injection system without rotating organs. A
reservoir tank of 1.0 dm3, located before the electro-injector,
assures the absorption of the pressure oscillations produced
from the needle opening and the gas recharge. Further details
of the apparatus are reported in ref. [13]. The injection system
and synchronized acquisition set up have been managed by a
programmable electronic control unit (PECU). This is an
open system able to reproduce the injector energizing
currents for the desired strategy in terms of number of
injection pulses, durations, rise and dwell times. The PECU
reproduces in exit a TTL signal, relate to the injection event,
for synchronizing the consecutive images management and
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
acquisitions. The injector exhibits a great flexibility to range
from low to high load engine conditions [14].
The gasoline mass flow rate has been measured by means of
an AVL Fuel Injection Gauge Rate System working on the
Bosch tube principle [15, 16]. The fuel is sent into a small
chamber kept at the constant pressure of 0.8 MPa and
connected to a pipeline of constant internal section, 12 m
long. An AVL GM12D pressure transducer collects the
pressure variation produced by the fuel delivered from the
nozzle. The pressure variation, Δp, is proportional to the
injected quantitiy, q, through the following relationship:
impingement. The plate has been located at 20 mm from the
nozzle tip, thanks to a X-Y-Φ micrometric apparatus, in two
different mode: parallel and inclined of 30° respect the total
spray axis, fully framed by the CCD camera. The average
roughness of the impact-wall is 1.077 µm, measured by the
Stylus Profilometer, Model Surtronic 3 by Rank Taylor
Hobson. The plate was at room temperature. The experiments
on spray development and atomization characteristics of the
impinging sprays have been performed in terms of injection
pressure (Pinj) and wall inclination angle (α) at the impinging
distance (dwall) of 20 mm. All the spray impacts have been
performed at ambient pressure and temperature. The detailed
experimental conditions are listed in Table 1.
Table 1. Experimental conditions for impinging spray
where ρ is the fuel density, Atube the inner section of the pipe
and a the speed of the sound in the fuel. A time resolution of
10 µs is adopted, suited with the 130 KHz natural frequency
cut of the transducer (7.6 µs). The measured q value gives the
injected quantity in the time interval 10 µs, while the total
injected quantity per shot is evaluated as the integral over the
entire pulse duration. The measurements are average over
100 shots and the total amount is compared with the one
collected at the discharge pipe, as weigthed through a high
precision balance and the discrepancies were always lower
than 2%.
The spray visualization is made by injecting the fuel in a high
pressure quiescent vessel, optically accessible, containing
nitrogen at the pressure of 0.1 MPa and the temperature of
300 K. Images of the spray, enlightened by powerful flashes,
have been collected at different instants from the Start of
Injection (SOI) by means of a synchronized CCD camera,
1376×1040 pixels, 12 bit resolution, 0.5 µs shutter time. An
overview of the experimental scheme for the image
acquisition is reported in ref. [13]. The images have been
acquired using a 50 mm focal lens reaching a calibration
factor of 10 pixel/mm. The captured images have been
processed off-line by means of a proper software able to
extract the parameters characterizing the spray dynamics. A
set of 5 images has been collected for each injection
condition for a statistical analysis of the cycle-to-cycle
dispersion. The images processing analysis has been carried
out in different steps: image acquisition and background
subtraction, filtering, fuel spray edges determination, tip
penetration and cone angle measurements. Background
subtraction and median filter procedures have been adopted
during the image acquisition to remove impulse noise and
stray light so to maintain sharp the spray edge. This is
determined selecting an intensity threshold level for
separating the fuel region from the background ambient gas.
More details about image post-processing procedure can be
found in [17]. Finally, a stainless steel flat plate has been
introduced into the vessel to reproduce the spray-wall
The growth of the impinging gasoline spray, which is a
function of the radial penetration (L) and spray height (H),
has been analyzed in order to estimate the development of the
spray on the wall. The radial penetration has been defined by
the distance between the spray center and the end of spray
behavior while the spray height has been defined by the
height between the plate bottom and the upper spray behavior
[18]. Figure 1 shows the definitions of the spray height (H),
radial spray penetration (Lsx and Ldx), inclination angle of the
wall (α) and impinging distance (dwall) both for the vertically
(left side) and the inclined impinging spray (right side).
For the lateral spray penetration (L) the average between the
corresponding values on the left (Lsx) and on the right (Ldx)
respect the injector axis has been considered. While for the
measures of the spray height only the value on the left side
after the impact has been considered and only for the
condition α=0°.
CFD SETUP
The analyses have been carried out by means of the Star-CD
code, licensed by CD-Adapco [19, 20]; a computational grid
reproducing the experimental quiescent chamber has been
used for the validation of the CFD models; the cylinder of 40
mm of radius and 90 mm of height has been made up of
150000 hexahedral cells, with a average cell size equal to 1.5
mm [21]. As the CFD simulations are compared in terms of
spray impingement on the wall, the computational grid is a
cylinder of 50 mm of radius and 22 mm of height. It is made
by 70000 hexahedral cells while the mesh has been refined
near the impinging region, resulting cells of 1 mm of
dimension. A wall boundary condition has been used to
simulate the impinging region.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Figure 1. Definition of the impinging spray characteristics.
The RNG k-ε model has been chosen as a closure model [22,
23] for the simulation of the turbulence. All the analyses have
been carried out within the framework of the lagrangian
approach [24, 25] to model the fuel spray and a set equal to
170000 computational parcels has been injected. As far as the
primary breakup has been considered, the initial droplets
have been introduced adopting the modified linear analysis of
Kelvin-Helmholtz instability (KH model) [26, 27, 28]: in
order to take into account the effect of the cavitation the
resulting distribution of the atomized droplets has been
modified.
As far as the secondary breakup is concerned a tuning of the
Rayleigh-Taylor model is carried out [29, 27]. Turbulent
phenomena due to the droplet flow interaction have been
taken into account by the stochastic approach described in
[30]; to correctly define the interaction between the droplets
into the toroidal zone of the spray, a collision model [31] has
been adopted.
Figure 2 shows the various impingement regimes of a
droplet-wall interaction. The detailed Bai and Gosman model
[32] has been used for the calculation of the impinging
droplets. In the stick regime, a droplet with low kinetic
energy adheres to the wall in nearly spherical form and
continues to evaporate. In the case of spread, the droplet
impacts with moderate velocity on a dry or wetted wall,
spreads out and mixes with the wall film (wetted wall) or
forms a wall film (dry wall). If rebound occurs, the droplet
bounces off the wall (reflection) and does not break up. This
regime is observed in the case of dry and hot walls, where the
contact between drop and wall is prevented by a vapor
cushion. Rebound also occurs in the case of a wet wall if the
impact energy is low and an air film between drop and liquid
film minimizes energy loss. In the boiling-induced break-up
regime, the droplet disintegrates due to a rapid liquid boiling
on a hot wall. The wall temperature must be near the
Nakayama temperature TN, at which a droplet reaches its
maximum evaporation rate. In the case of break-up, the
droplet deforms into a radial film on the hot surface, which
breaks up due to thermo-induced instability. The splash
regime occurs at very high impact energy. A crown is
formed, jets develop on the periphery of the crown, become
unstable and disintegrate into many droplets.
Figure 2. Illustration of the impacts regimes
Figure 3. Impinging regimes and transition conditions
Bai and Gosman [32] have developed a more detailed
impingement model considering also the splash regime, as
depicted in figure 3 where the transition conditions are
reported too. In the case of a dry wall, the stick and spread
regimes are combined and called the adhesion regime and the
stick regime is neglected in the case of a wetted wall because
of its typically very low impact energy. The model varies the
transition regimes:
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
RESULTS AND DISCUSSION
(1)
is the Weber number, which takes into account the ratio of
the droplet's kinetic energy (vn is the velocity component
normal to the surface, ρl the liquid density, d the droplet
diameter, σ surface tension);
Different injection strategies have been tested changing the
injection duration, pressure and the corresponding total
injected mass. All the investigated configurations are
collected in Table 2. For each configuration, reported in
Table 2, the fuel injection rate has been extracted from the
experiments as the average of one hundred of shots.
Table 2. Injection strategies.
(2)
is the Laplace number, which evaluates the relative
importance of the surface tension and viscous forces (µl is the
dynamic viscosity of the liquid).
For a dry wall, the transition criterion between adhesion and
splash is given as
(3)
where the coefficient A depends on the surface roughness; in
the case of an already wetted wall the transition Weber
numbers for rebound and spread are
and for spread and splash are
(4)
Bai has modeled the post-impingement characteristics
introducing experimental correlation and random numbers in
the definition of the velocity components for the rebounding
droplets and for the radius of the secondary droplets (i.e. in
case of splash regime). As the droplets adhere on the wall a
liquid film formation could occur. A CFD model developed
by Bai and Gosman [33], multi-component evaporation and
condensation on the film surface is based on the theory given
in [34] and [35]. The model accounts for convective transport
of conserved quantities within the film and from/to the gas
phase.
Figure 4 shows the injection profiles obtained for the
investigated injection strategies. The area under each profile
is the total amount of the injected fuel per stroke. The profiles
depend on the injection pressure and duration. Infact higher is
the injection pressure and higher is the stationary phase level
while greater is injection duration and longer the
corrisponding injection profile. The investigaed strategies
represent typical injection conditions for the star-up, idle and
maximum power engine operating conditions for a
turbocharged GDI engine.
It is interesting to note as all the profiles are quite similar in
the overlapping parts in terms of rise time and slope. In the
early stage of the injection the dynamic of the exiting fuel is
independent from the injection pressure and the mass flow
rate reaches the peak value after 140 µs for 12 MPa and 5.5
MPa of injection pressure while for the 3.0 MPa after 150 µs.
The three investigated cases present an incresed mass flow of
83%, 90% and 130% respectevly for 12, 5.5 and 3.0 MPa,
comparing to the mass flow rate during the stationar period of
the injection. A subsequent pressure drop occuring in the
nozzle generates a strong decrease of the flow rate equal to
the −38%, −26% and −35% respect the stationary mass flow
for 12, 5.5 and 3.0 MPa. The two highest injection pressure
cases begin the stationary period of the injection after 700 µs,
while for the 3.0 MPa case it is evident dynamic and unstable
effects are dominant for all the injection period. This
injection pressure, adopted during the idle condition in
current production GDI engine, appears to be border line in
terms of fuel delivering and total injected fuel quantity.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Figure 4. Fuel injection rate profiles for the different
injection conditions.
All the numerical simulations adopt the experimental flow
rates. In order to validate the CFD models a preliminary
analysis has been performed. The spray has been simulated
for the three injection pressure in a quiescent chamber at
ambient temperature and pressure. The atomization and
secondary breakup models have been tuned comparing the
calculated spray with the experimental images, as reported in
figure 5. The pictures report the fuel spray behavior in the
early stages of the injection, the left side images show the
experimental results and the right show the calculated results.
The numerical and experimental comparison appears to be
quite good, the simulations well capture the spray
morphology for all of the three injection pressure. As the
pressure increases a strong jet to jet interaction is visible in
the experimental images and explained in [21]. Due to the
pressure drop occurring into the center of the spray, for the
case with the two highest injection pressures, the center line
jet shows a greater penetration than the neighbors. The CFD
images report the computational parcels colored in function
of the droplet diameter, since blue is the lowest diameter and
red is the highest, it is visible as the injection pressure
increases the injected droplets are characterized by lower
droplet size.
Figure 6 shows the good experimental and numerical
comparison, in terms of spray tip penetration. The CFD
simulations of the three cases well capture the experimental
penetration in the early stages of the injection, while, after 0.5
ms after SOI, the numeric over-predicts the experiment. The
gap is more relevant for 5.5 and 12 MPa cases and almost
focused on the stationary period of the injection. After 1.2 ms
after SOI the 3.0 MPa and 12MPa better match with
experimental results.
Figure 5. Experimental (left) and numerical comparison
(right) of the spray morphology for the three injection
pressure.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
case. From the above statements, CFD simulation predicts
secondary droplets with an higher tangential momentum than
experiment, leading to an increased L penetration and a
shorter H penetration.
Figure 6. Experimental and numerical comparison of
the tip spray penetration for all the injection pressures.
In figure 7 and in figure 8 the influence of the injection
pressure on the lateral spray penetration is reported for both
inclination angle of the wall (α=0° and α=30°) with the wall
distance of 20 mm. For each curve the experimental starting
point indicates the impact time of the spray against the wall.
The penetration rates just after impingement begin to increase
according to increasing the injection pressure. In Figure 8 the
effects of the wall inclination angle on the lateral spray
penetration can also be seen, it increases with increase of the
wall inclination angle (going from α =0° to 30°). This trend
indicates that the effect of wall impingement is reduced
owing to a reduction in radial droplet momentum as the
inclination angle of the wall is increased.
A slight overestimation of the radial penetration can be seen
in the numerical and experimental comparison, in figure 7
and 8. It is more evident in the simulations of the impinging
spray on the flat plate with inclination angle α=0°. In the first
part of the penetration rate the CFD simulations, in figure 8,
show a good agreement in terms of L penetration, for all of
the injection pressure, while after 0.7 ms after SOI, a slight
gap with the experimental data appears. The slight over
estimation shown in the numerical penetration in figure 6 is a
possible explanation. The simulation computes higher normal
component velocity of the impinging droplets than
experiment, resulting in higher tangential component velocity
of the secondary droplets. Furthermore, as the tangential
velocity component of the droplets, after the impingement in
the splash regime, is function of the tangential velocity
component of the impinging droplets and function of a
friction coefficient cf [32], a tuning activity has been
performed. Literature suggests for water spray a value of 0.7
while in this activity a cf=0.5 has been estimated. Lower
values of cf have not been taken into account.
As far as the numerical calculations are concerned, reaching
the mass flow rate the steady condition, the 5.5 MPa and 12
MPa injection pressure show an under predicted height
penetration H of the spray, as visible in figure 9 for the α=0°
Figure 7. Radial spray penetration (L) at various
injection pressures for α=0° inclination angle of the wall.
Figure 8. Radial spray penetration (L) at various
injection pressures for α=30° inclination angle of the
wall.
Figure 9. Height spray penetration (H) at various
injection pressures for α=0° inclination angle of the wall.
Figures 10 (3.0 MPa), 11 (5.5 MPa) and 12 (12MPa) report
the mean Weber Numbers of the spray averaged on a section,
for the two inclination angle: red points regard the impinging
droplets while the black points regard the droplets after the
impact on the wall. Since the spray is hitting the dry wall at
the beginning of injection process, the transition regime is
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
regulated by (3). Due to the high pressure peak, previously
described in the figure 4, the impinging occurs in the splash
regime. Later in the injection period, as the mass flow rate
reaches the steady condition, the Weber Number of the
droplet decreases, as visible in the figure 10, 11, 12 and
spread regime could occur for some droplets, even if the
splash is still the main impinging regime as Wedroplet is
greater than (4).
Figure 12. Weber Number of the droplets before the
impact (red) and after the impact (black) for the 12 MPa
case.
Figure 10. Weber Number of the droplets before the
impact (red) and after the impact (black) for the 3.0 MPa
case.
Figure 11. Weber Number of the droplets before the
impact (red) and after the impact (black) for the 5.5 MPa
case.
An overall qualitative investigation is reported in the last
figures, from 13 and 14. Each figure depicts impinging spray
from a front view, comparing the experimental images with
the simulations for all of the investigated injection pressures.
The figures show the effects of the injection pressure on the
macroscopic behavior of the fuel spray impinging on the
wall. As shown in the figures, the injection pressure hardly
affects the spray development process; at the highest injection
pressures the development of spray is very high. The spray
tips on the impingement wall for the injection pressure of
12.0 MPa are sharper than 5.5 and 3.0 pressures because of
the smaller droplet size produced. Higher is the velocity at
the point of impingement, due to the high injection pressure,
and higher is the radial velocity near the wall and results in a
sharper shape of the spray tip. In figure 13 as the injection
pressure has increased, the spray height increases. In the case
of inclined wall (figure 14), the downward penetration of the
spray is very large. It can indicate that the increased angle of
wall inclination reduces the effect of impingement on the
macroscopic spray characteristics (the momentum loss during
the impingement is less). As the inclined angle of the
impingement wall has increased, the spray path behavior is
enhanced.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Figure 13. Experimental (top) and numerical (bottom) comparison, images of impinging spray for all injection pressures and
α=0°
The simulated spray has been reproduced by means of
computational parcels depicted in constant diameter aiming
to better show the after impinged droplets behavior. Focusing
on the figure 13 the simulations well describe the wake of
secondary droplets and despite at the end of the injection
CFD could be improved, globally the details are captured.
Similar considerations are evident for the 30° of inclination
angle (figure 14): vortexes due to the lost radial momentum
of the droplets are described by simulations since the end of
the injection where some differences appear. As for the 12
MPa injection pressure case which shows for the simulation a
middle wake located between the impinging spray and the
edge of the spray is missing.
CONCLUSIONS
An experimental and numerical study have been performed to
investigate the effects of injection pressure and wall
inclination angle on the macroscopic behavior of a multi-hole
GDI spray impinging on the wall. The experimental
characterization has been carried out in terms of fuel injection
rate and spatial and temporal evolution of the spray
impinging on the wall. The spray visualization has been made
by injecting the fuel in a high pressure quiescent vessel,
optically accessible, containing nitrogen at the pressure of 0.1
MPa and the temperature of 300 K. A stainless steel flat
plate, at the temperature of 300 K and distance of 20 mm
from injector nozzle, has been introduced into the vessel to
reproduce the spray-wall impingement. The growth of the
impinging spray has been analyzed in terms of radial
penetration (L) and spray height (H).
The experimental data together with optical images have been
available for the validation by means of the Star-CD code. In
order to validate the lagrangian CFD models, preliminary
simulations have focused on the spray evolution within a
quiescent chamber at ambient temperature and pressure.
Since the acceptable experimental and numerical
comparisons, both in terms of spray penetration and
morphology, the CFD analysis have investigated the spray/
wall interactions.
The main results can be summarized as follows:
• The fuel injection rate profiles, for the three investigated
injection pressures, show a similar behavior in terms of rise
time and slope; higher is the injection pressure and higher is
the level of the stationary phase.
• A complete description of the dynamic behavior of the
exiting fuel is described in terms of delay.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
Figure 14. Experimental (top) and numerical (bottom) comparison, images of impinging spray for all injection pressures at
α=30°
• The highest injection pressures promote a longer radial
penetration and spray height of the impinging spray.
• Good agreement between experimental and computed
results in terms of spray cone angle, tip penetration and spray
morphology for all the three injection pressures before and
after the impingement;
• Since spray impact is simulated at ambient temperature and
at high injection pressure, the splash regime is correctly
modeled. A tuning of the CFD sub models has been
performed improving the comparison of the radial spray
penetration.
• Differences are evident in the H penetration of the spray at
the end of the injection process leading to the conclusion the
CFD simulation slightly over predicts the secondary droplets
momentum.
This activity aims to analyze the effects of the impinging
multi-hole spray in different configurations and conditions
and the results reported in the paper are a preliminary step
towards the investigation of such a complex set of events.
Future works will focus on new impinging configurations
(i.e. different injector distances, different wall inclinations)
and the effect of the wall temperature will be take into
account and investigated.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
REFERENCES
1. Solomon, A.S., Anderson, R.W., Najt, P.M., Zhao, F.
“Direct Fuel Injection for Gasoline Engines”, Society of
Automotive Engineers, Inc., Warrendale, PA,
978-0768005363, 1999.
2. Zhao, F., Harrington, D.L., Lai, M-C.D., “Automotive
Gasoline Direct-Injection Engines,” Society of Automotive
Engineers, Inc., Warrendale, PA, 978-0-7680-0882-1, 2002,
doi:10.4271/R-315.
3. Zhao, F., Lai, M., and Harrington, D., “A Review of
Mixture Preparation and Combustion Control Strategies for
Spark-Ignited Direct-Injection Gasoline Engines,” SAE
Technical Paper 970627, 1997, doi: 10.4271/970627.
4. Habchi, C., Foucart, H. and Baritaud, T. “Influence of the
Wall Temperature on the Mixture Preparation in DI Gasoline
Engines”, Oil & Gas Science and Technology - Rev. IFP,
Vol. 54 (1999), No. 2, pp. 211-222.
5. Tsunemoto, H. (1996). “Process of mixture formation of
impinging spray on the wall in a hole type nozzle”, Int. J.
JSME 27, 2, 36-45.
6. Ishikawa, N. (1998), “Observation of spray impingement
behavior in a small DI combustion chamber”, Int. J. JSME
29, 2, 23-28.
7. Park, J., Xie, X., Im, K., Kim, H. et al., “Characteristics of
Direct Injection Gasoline Spray Wall Impingement at
Elevated Temperature Conditions,” SAE Technical Paper
1999-01-3662, 1999, doi:10.4271/1999-01-3662.
8. Walkins, A.P., Park, K., 1996. “Assessment and
application of a new spray wall impaction model” Proc. Inst.
Mech. Eng. Conf. Comput. Reciprocating Gas Turbines,
1-10.
9. Bai, C. and Gosman, A., “Development of Methodology
for Spray Impingement Simulation,” SAE Technical Paper
950283, 1995, doi: 10.4271/950283.
10. Shim, Y. S., Gyung-Min, C., Kim, D., “Numerical and
experimental study on effect of wall geometry on
wallimpingement process of hollow-cone fuel spray under
various ambient conditions”, International Journal of
Multiphase Flow 35, 2009, 885-895.
11. Inamura, T., Tomoda, T. “Characteristics of sprays
through a wall impinging injector”, Atomization and sprays,
vol. 14, pp. 375-395, 2004.
12. Lee, S. Y., Ryu, S. U. “Recent progress of spray-wall
interaction research”, Journal of Mechanical Science and
Technology, Vol. 20, N.9, pp. 1101-1117, 2006.
13. Alfuso, S., Allocca, L., Greco, M., Montanaro, A.,
Valentino, G., “Time- and Space Characterization of Multihole GDI Sprays for IC Engines by Images Processing and
PDA Techniques”, Paper ILASS08-071, 2008.
14. Costa, M., Sorge, U., Allocca, L., “Numerical study of
the mixture formation process in a four strokes GDI engine
for two-wheels applications”, Simulation Modelling Practice
and Theory Journal, doi: 10.1016/j.simpat.2010.07.006, 2010.
15. Bosch, W., “The Fuel Rate Indicator: A New Measuring
Instrument For Display of the Characteristics of Individual
Injection,” SAE Technical Paper 660749, 1966, doi:
10.4271/660749.
16. Wallace, I. “Injection Rate Gauge: Pass Off Information
and User Instructions” - Fuel & Engine Management
Systems, Graz - December 2002.
17. Alfuso, S., Allocca, L., Caputo, G., Corcione, F. et al.,
“Experimental Investigation of a Spray from a Multi-jet
Common Rail Injection System for Small Engines,” SAE
Technical Paper 2005-24-090, 2005, doi:
10.4271/2005-24-090.
18. Lee, C.H., Lee, K.H. “Experimental Study on
Macroscopic Spray Characteristics after Impingement in a
Slit-Type GDI Injector”, Internal journal of Automotive
Technology, Vol. 9, No. 3, pp. 373-380 (2008), doi: 10.1007/
s12239-008-0045-2.
19. Computational Dynamics, “STAR-CD User Guide”,
2010, London (UK).
20. Computational Dynamics, “STAR-CD Methodology”,
2010, London (UK)
21. Malaguti, S., Fontanesi, S., Cantore, G. “Numerical
characterization of a new high-pressure multi-hole GDI
injector”, ILASS Europe 2010, Brno, Czech Republic,
September 2010.
22. Yakhot, V., Orszag, S.A., Thangam, S., Gatski, T.B., and
Speziale, C.G. “Development of turbulence models for shear
flows by a double expansion technique”, Phys. Fluids, 1992.
23. Cantore, G., Fontanesi, S., Mattarelli, E., and Bianchi,
G., “A Methodology for In-Cylinder Flow Field Evaluation in
a Low Stroke-to-Bore SI Engine,” SAE Technical Paper
2002-01-1119, 2002, doi:10.4271/2002-01-1119.
24. Bracco, F., “Modeling of Engine Sprays,” SAE
Technical Paper 850394, 1985, doi: 10.4271/850394.
25. Dombrowski, N., Johns, W. R. “The aerodynamic
instability and disintegration of viscous liquid sheets”, Chem.
Eng. Sci., v. 18, p. 203, 1963.
26. Patterson, M. and Reitz, R., “Modeling the Effects of
Fuel Spray Characteristics on Diesel Engine Combustion and
Emission,” SAE Technical Paper 980131, 1998, doi:
10.4271/980131.
27. Beal, J. C., Reitz, R. D., “Modeling spary atomizaion
with the Kelvin-Helmholtz/Rauleigh-Taylor Hybrid modela”,
Atomization and Sprays, vol.9, pp 623-650, 1999
28. Reitz, D. “Modeling Atomization Processes in HighPressure Vaporizing Sprays”, Atomization and Spray
Technology 3, 1987, pp309-337.
Gratis copy for Alessandro Montanaro
Copyright 2012 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Monday, March 19, 2012 05:04:19 AM
29. Taylor, G.I., “The instability of liquid surfaces when
accelerated in a direction perpendicular to their planes”, In
Bachelor, GK, The Scientific Papers of Sir Geoffery Ingram
Taylor, 1963, vol. 3, pp 532-536, Cambridge University
Press.
30. Gosman, A.D., and Ioannides, S.I. 1983. “Aspects of
computer simulation of liquid-fuelled combustors”, AIAA, J.
Energy, 7(6), pp. 482-490.
31. O'Rourke, P.J. “Collective Drop Effects on Vaporizing
Liquid Sprays”. PhD Thesis, University of Princeton, 1981.
32. Bai, C. and Gosman, A., “Development of Methodology
for Spray Impingement Simulation,” SAE Technical Paper
950283, 1995, doi: 10.4271/950283.
33. Bai, C. and Gosman, A., “Mathematical Modelling of
Wall Films Formed by Impinging Sprays,” SAE Technical
Paper 960626, 1996, doi: 10.4271/960626.
34. Sirignano, W.A. 1999. “Fluid Dynamics and Transport of
Droplets and Sprays”, Cambridge University Press, New
York.
CONTACT INFORMATION
Dr. Alessandro Montanaro
Istituto Motori CNR
Via G. Marconi 8, 80125 Napoli, Italy
a.montanaro@im.cnr.it
Dr. Simone Malaguti
Department of Mechanical and Civil Engineering
University of Modena and Reggio Emilia
Via Vignolese 905, 41100 Modena, Italy
Ph.:+39 059 2056114
simone.malaguti@unimore.it
ACKNOWLEDGMENTS
The authors wish to acknowledge CD-Adapco Group, for the
use of the Star-CD code, granted to the University of Modena
and Reggio Emilia.
35. Torres, D.J., O'Rourke, P.J., and Amsden, A.A. 2003.
‘Efficient multicomponent fuel algorithm’, Combust. Theory
Modelling, 7, p. 67.
36. Ruge, J.W., and Stüben, K. 1986. ‘Algebraic Multigrid
(AMG)’ in “Multigrid Methods” (Ed. McCormick, S.),
Frontiers in applied Mathematics, SIAM, 5, Philadelphia.
37. Versteeg, H. K., Malalasekera, W., “An introduction to
computational fluid dynamics. The finite volume method”,
Longman, 1995.
The Engineering Meetings Board has approved this paper for publication. It has
successfully completed SAE's peer review process under the supervision of the session
organizer. This process requires a minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a
retrieval system, or transmitted, in any form or by any means, electronic, mechanical,
photocopying, recording, or otherwise, without the prior written permission of SAE.
ISSN 0148-7191
Positions and opinions advanced in this paper are those of the author(s) and not
necessarily those of SAE. The author is solely responsible for the content of the paper.
SAE Customer Service:
Tel: 877-606-7323 (inside USA and Canada)
Tel: 724-776-4970 (outside USA)
Fax: 724-776-0790
Email: CustomerService@sae.org
SAE Web Address: http://www.sae.org
Printed in USA