City, University of London Institutional Repository
Citation: Mitroglou, N., Nouri, J. M., Gavaises, M. and Arcoumanis, C. (2006). Spray
characteristics of a multi-hole injector for direct-injection gasoline engines. International
Journal of Engine Research, 7(3), pp. 255-270. doi: 10.1243/146808705X62922
This is the unspecified version of the paper.
This version of the publication may differ from the final published
version.
Permanent repository link: https://openaccess.city.ac.uk/id/eprint/1515/
Link to published version: http://dx.doi.org/10.1243/146808705X62922
Copyright: City Research Online aims to make research outputs of City,
University of London available to a wider audience. Copyright and Moral
Rights remain with the author(s) and/or copyright holders. URLs from
City Research Online may be freely distributed and linked to.
Reuse: Copies of full items can be used for personal research or study,
educational, or not-for-profit purposes without prior permission or
charge. Provided that the authors, title and full bibliographic details are
credited, a hyperlink and/or URL is given for the original metadata page
and the content is not changed in any way.
City Research Online: http://openaccess.city.ac.uk/ publications@city.ac.uk
City Research Online
Spray characteristics of a multi-hole injector for
direct - injection gasoline engines
N. Mitroglou, J. M. Nouri, M. Gavaises and C. Arcoumanis
Centre for Energy and the Environment
School of Engineering and Mathematical Sciences, City University, London, UK
Abstract: The sprays from a high-pressure multi-hole
nozzle injected into a constant volume chamber have been
visualised and quantified in terms of droplet velocity and
diameter with a two-component phase Doppler
amenometry (PDA) system at injection pressures up to
200bar and chamber pressures varying from atmospheric
to 12bar. The flow characteristics within the injection
system were quantified by means of an FIE 1-D model,
providing the injection rate and the injection velocity in
the presence of hole cavitation, by an in-house 3-D CFD
model providing the detailed flow distribution for various
combinations of nozzle hole configurations, and by a fuel
atomisation model giving estimates of the droplet size
very near to the nozzle exit. The overall spray angle
relative to the axis of the injector was found to be almost
independent of injection and chamber pressure, a
significant advantage relative to swirl pressure
atomisers. Temporal droplet velocities were found to
increase sharply at the start of injection and then to
remain unchanged during the main part of injection
before decreasing rapidly towards the end of injection.
The spatial droplet velocity profiles were jet-like at all
axial locations, with the local velocity maximum found at
the centre of the jet. Within the measured range, the effect
of injection pressure on droplet size was rather small
while the increase in chamber pressure from atmospheric
to 12bar resulted in much smaller droplet velocities, by
up to fourfold, and larger droplet sizes by up to 40%.
Key words: gasoline direct injection engines, high-
pressure multi-hole injectors, phase Doppler
anemometry, nozzle flow CFD simulation, atomisation
modelling
1. Introduction
The objective of introducing direct-injection gasoline
engines into the market is to reduce fuel consumption
through charge stratification under overall lean
conditions, to increase volumetric efficiency and to
reduce exhaust emissions. There are numerous feasible
design configurations for spark-ignition gasoline direct
injection engines, which are classified depending on the
relative position of the injector to the spark plug and
piston crown shape, the injection timing and the air
motion and mixture preparation strategy. They are
classified as wall-, air-, or spray-guided combustion
systems, employing central or side fuel injection. In all
concepts, good combustion is achieved by formation of a
stable and ignitable mixture around the spark plug at the
time of ignition. The major component of the fuel
injection system that is responsible for preparing such a
fuel/air mixture cloud is the high-pressure injector. Thus,
knowledge of the spray characteristics, including spray
structure, tip penetration and distribution of droplet
velocities and diameters as a function of nozzle design,
injection and chamber pressures, is essential.
Previously published investigations [1-8] have
mainly focused on swirl pressure atomisers, known as
first-generation injectors. In general, this type of injector
can produce very finely atomised droplets with diameters
(SMD) in the range 15-25μm over a moderate range of
injection pressures (50-100bar). Their disadvantage is
that the spray generated from these injectors is very
sensitive to the operating and thermodynamic conditions.
An unavoidable ‘collapse’, i.e. a reduction of spray angle
and penetration at elevated chamber pressures
(corresponding to the late-injection strategy of spray-
guided systems) has been reported. A different type of
injector, employing a multi-hole nozzle, has been recently
introduced by fuel injection manufacturers, aiming to
overcome this dependence of the spray characteristics on
thermodynamic and operating conditions by introducing
several holes in a configuration similar to diesel injector
nozzles. Up to now there have been limited investigations
on this type of injectors [9-12], who confirmed the
improved stability of the spray at elevated chamber
pressures relative to that of swirl injectors. Also,
enhanced air entrainment has been observed due to the
N.Mitroglou, J.M.Nouri, M.Gavaises and C.Arcoumanis
separated spray jets, and the larger surface area, which
can be independently directed at desired locations,
achieving improved matching between the injector and
the combustion chamber designs. There is a variety of
multi-hole injector nozzle configurations that have been
designed and manufactured, associated with the
flexibility in hole positioning throughout the injector
nozzle cap (e.g. 6 holes symmetrically distributed, 5 holes
plus one in the centre, 12 holes, and all possible
combinations as shown schematically in Fig.1).
In the present investigation a six-hole injector has
been used to provide a quite symmetrical spray pattern.
The aim is to quantify the effect of injection pressure up
to 200bar and chamber pressure up to 12bar on the spray
structure, using a pulsed light source and a CCD camera,
and on the droplet velocities and sizes as measured with a
phase-Doppler anemometer (PDA). The interpretation of
the results is assisted by CFD simulations predicting the
flow distribution within the injection system, in the
nozzle tip itself and the near nozzle fuel atomisation
process. The following sections describe the experimental
arrangement, the measurement systems and the computer
model, followed by presentation of the results and a
summary of the main conclusions.
2. Experimental arrangement and
instrumentation
A common rail system shown schematically in Fig.2,
with the six-hole injector installed inside a constant-
volume chamber, has been used in this investigation. A
three-piston-type pump coupled to an electric motor is
responsible for delivering high-pressure fuel (up to
200bar) to the common rail, which has been specifically
built with one injector outlet. This common rail was
connected to the injector via a pipe with specific diameter
and length which was, in turn, fixed to the high-pressure
chamber that is equipped with four quartz windows and
connected to a pressurised bottle of nitrogen for
maintaining the required back pressure inside the
chamber (up to 25bar). A fuel pressure regulator attached
to the common rail, a solenoid valve in the chamber’s
exhaust pipe and the injector were all controlled
electronically.
Two prototype 6-hole injectors with a nominal
overall spray cone angle of 90°, a hole diameter of
~140μm, forming an L/D (length/hole diameter) ratio of
2.14, and an operating pressure of up to 200bar were
tested. The first one has a central hole with one of the side
holes missing, while the second one has a symmetric hole
arrangement. Tests have been carried out at two,
relatively high, injection pressures of 120 and 200bar and
at four chamber pressures of 1, 4, 8 and 12bar. The
duration of the injection triggering signal (i.e. the
injection quantity) was kept constant at 1.5ms. Iso-octane
has been selected as the working fluid, since it is safer to
use and more convenient for optical studies than gasoline;
it has a density, kinematic viscosity and surface tension of
692kg/m
3
, 0.78cSt and 0.0188N/m, respectively.
Images of the spray were obtained with a time
resolution of 50μs by a non-intensified, cooled CCD
camera with a spatial resolution of 1280x1024 pixels, a
sensitivity of 12bit and a minimum exposure time of
100ns. A strobe light of 20μs duration was used as the
light source, which was synchronised to the camera.
A 2-D phase-Doppler anemometer shown
schematically in Fig.3, has been used for the
measurement of the axial and radial droplet velocities and
diameters. According to the manufacturer, a droplet size
range of 0.5μm to 100μm can be detected from the
system and a typical accuracy of the measured size
distributions is 4%, although it depends to a large extent
on the optical configuration. The transmitting and
receiving optics were installed on a 3-D traverse
mechanism with a resolution of 12.5μm in the X, Y axes
and 6.25μm in the Z axis, relative to the injector position.
A wall-mounted Argon-Ion laser with a maximum power
of around 1.5W was used and the output beam was
aligned with the fibre optic unit.
Fig. 1 Schematic of possible multi-hole injector nozzle
configurations (6-hole nozzles employ a L/D ratio of
2.14, while 12-hole nozzles appear to have twice the
L/D ratio of the 6-hole nozzles).
Spray characteristics of a multi-hole injector for direct-injection gasoline engines
0
¦
1
¦
¦
¦
¦
2
¦
¦
¦
¦
3
¦
¦
¦
¦
4
¦
¦
¦
¦
5
¦
¦
¦
¦
6
¦
¦
¦
¦
7
¦
¦
¦
¦
8
¦
¦
¦
¦
9
¦
¦
¦
¦
1
0
¦
¦
¦
¦
1
1
B a r
B a r
0
¦
¦
¦
¦
1
0
¦
¦
¦
¦
3
0
¦
¦
¦
¦
5
0
¦
¦
¦
¦
7
0
¦
¦¦
¦
9
0
¦
¦
¦
¦
1
1
0
¦
¦
¦
¦
1
3
0
¦
¦
¦
¦
1
5
0
¦
F ue l ta n k
L o w pre s su re
p u m p
F ilte r
S ole no id v alv e
E xh a ust p ipe
C on tro l b ox
P re ssu re ra il
L o w pre s su re fu e l p ip e s
H ig h p re ss ure fu e l p ipe s
F ue l re turn p ip e
C on tro l p ulse s
M oto r H ig h p re ss ure
p u m p
In je cto r
Fig. 2 Schematic of the constant volume chamber test rig
This unit was responsible for the splitting of the laser
beam into two pairs of different wavelengths; each pair
consisted of two equal intensity beams. The first pair was
green light with a wavelength of 514.5nm, responsible for
the axial velocity component, while the second pair was
blue light with 488nm wavelength providing the radial
velocity component.
A Bragg cell unit positioned inside this fibre optical
unit provided a 40MHz frequency shift. The transfer of
the four laser beams to the transmitting optics was
through a fibre-optic cable. The collimating and focusing
lenses formed an intersection volume with major and
minor axes of approximately 2.863 and 0.092mm for the
green, and 2.716 and 0.088mm for the blue component.
Light scattered by the droplets was collected by a 310mm
focal length lens positioned at 30° to the plane of the two
incident green beams to ensure that refraction dominated
the scattered light (Fig.3). The signal from the four
photomultipliers was transmitted to the processor unit
where all the data processing was carried out. The
processor was connected to a desktop computer via an
ethernet adaptor, where all the acquired data were saved
for further analysis. Up to 1000 validated sample data
were collected for each measurement location and a time
window of 0.1ms over many injection cycles, to allow
ensemble averages to be estimated. The measurements
were synchronised with the needle lift by an external
reset pulse, and restricted to the first 2.5ms after the start
of the injection process, depending on the axial location
and the pressure in the chamber.
Difficulties in measurements were encountered
during the main injection period especially in the central
part of the individual sprays jets and near the nozzle exit
region under certain test conditions due to the attenuation
of the incident laser beams and the scattered light. The
problem was more pronounced in the case of injection
against elevated chamber pressures, where the system
was unable to detect adequate signals during the main
part of injection up to an axial distance of 20 mm from
the nozzle exit.
Re c e ivin g O p tic s
Tra n sm isio n O p tic s
In je c to r
To p Vie w
3 0 d e g re e s
Fig. 3 Optical configuration of the phase Doppler
anemometer (PDA) system.
3. Computer simulation model
In this section, the methodology employed in order to
calculate the whole fuel injection process, that comprises
the fuel injection system, the nozzle flow and the
atomisation process of the injected sprays, is briefly
described.
A variety of models have been applied to the
simulation of the fuel injection process. Initially, a 1-D
model has been used for the simulation of the pressure
N.Mitroglou, J.M.Nouri, M.Gavaises and C.Arcoumanis
waves developing inside the fuel injection system. The
model is based on the solution of the mass and
momentum flow conservation equations, expressed in 1-
D, and which are solved numerically using the method of
characteristics. It estimates the transient variation of the
injection pressure inside the nozzle gallery and the flow
rate through the discharge holes using as inputs the
geometric characteristics of the rail, the connecting pipe
and the nozzle itself as well as the nominal pressure value
inside the common-rail. The needle lift, shown in Fig. 4,
as well as the nozzle geometric details are additional
inputs required by the model. The model used has been
found to predict accurately the total fuel injection
quantity as a function of injection pressure and injection
duration, according to Fig. 5, for different needle lifts; a
typical one is shown together with the triggering signal in
Fig. 4. As can be seen, the volumetric capacity of the
injector is almost a linear function of the triggering pulse
width for injection durations greater than 1ms, but less so
for shorter pulse durations. This is related to the fact that
the needle opens fully at around 0.85 ms from triggering.
It is also evident that the volumetric capacity of the
injector at 200bar injection pressure is larger, as
expected, than at 120bar.
-0 .2
0 .0
0 .2
0 .4
0 .6
0 .8
1 .0
1 .2
1 .4
1 .6
1 .8
2 .0
2 .2
0 .0
0 .5
1 .0
1 .5
2 .0
2 .5
3 .0
V oltage (V )
Ti me after trigger ing (m s)
Trigg er ing P ulse
0
10
20
30
40
50
60
70
80
N e ed le Lift
N ee d le Lift (μ m)
Fig. 4 Triggering signal and typical needle lift diagram
Fig. 5 Volumetric capacity of the injector as a function of
injection pulse duration under atmospheric conditions
and for two injection pressures
Past studies on hole-type nozzles have indicated that
hole type nozzles such as that investigated here, cavitate
above a threshold values for the injection pressure, for a
given back pressure. Once cavitation initiates, then the
discharge coefficient reduces asymptotically as function
of the cavitation number [13], which is defined as
CN=(P
UP
- P
BACK
)/(P
BACK
– P
VAPOR
). An empirical formula
allowing for such prediction is used here and the
corresponding result is shown in Fig. 6b. This, in turn,
can lead to the prediction of the hole effective area, which
is the percentage of the cross sectional hole exit area
occupied by liquid, with the remaining part assumed to
consist of cavitating bubbles. In the case of cavitating
nozzle flow conditions, the effective area decreases with
increasing cavitation number (or injection pressure), as
shown in Fig. 6a. The value of the hole effective area is a
measure of the increase of the injection velocity as a
result of the formation of cavitation relative to that under
non-cavitating conditions. More details about this simple
hole cavitation model as well as the 1-D fuel injection
system model can be found in [14].
30 80 130 180 230
0
10
20
30
40
0.6
0.7
0.8
0.9
1.0
0.6
0.7
0.8
0.9
1.0
D
30
(μm)
Injection Pressure (bar)
18bar
12bar
6bar
1bar
Discharge Coeff.
Effective Area
Fig. 6 Predicted nozzle hole effective area (a), hole
discharge coefficient (b) and droplet volume mean
diameter (c) as a function of injection pressure for
different chamber pressure values.
For the simulation of the detailed flow distribution
inside the sac volume and the injection holes, a multi-
dimensional turbulent CFD flow solver, named GFS, has
been employed. The time-averaged form of the
incompressible Navier-Stokes equations describing the
continuity, momentum and conservation equations for
scalar variables were numerically solved on an
unstructured non-orthogonal and curvilinear numerical
(a)
(b)
(c)
