18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
Ultra-high speed PIV measurement for gasoline spray ejected
from a multiple hole DI injector
Shibata, T.*1 Zama, Y.*1 Furuhata, T.*1 Kusano, H.*2
*1:Division of Mechanical science and Technology, Gunma University,
1-5-1 Tenjin-cho, Kiryu, Gunma, Japan, 376-8515
*2: Analytical & Measuring Instruments Division, Shimadzu Corporation,
1, Nisinokyo,Kuwabara-cho, Nakagyo-ku,Kyoto, Japan, 604-8511
Keywords: gasoline spray, velocity field, ambient gas density, injection pressure, PIV
ABSTRACT
Recently, injection pressure of gasoline fuel in a direct injection spark ignition (DISI) engine tends to increase
as compared with a conventional DISI engine in order to promote atomization of the fuel. Moreover, high boosted
engine by a turbo charger tends to be apply in terms of downsizing concept in order to improve the thermal
efficiency of the engine. It means that ambient gas density in the combustion chamber is higher than that of the
engine without the high boost. Here, injection pressure of the fuel and the ambient gas density is strongly related to
atomization process of the fuel, and the atomization of the fuel affects the mixture formation between the fuel and
air. Therefore, it is necessary to understand the atomization process near the nozzle so as to consider strategy of a
high thermal efficiency engine. In order to understand the flow characteristics of the gasoline spray, velocity
measurement in a gasoline spray was needed. In this study, gasoline spray injected to constant volume vessel was
taken by high-speed video camera. Velocity field of the gasoline spray was measured with time-resolved PIV. Effect
of ambient gas density and injection pressure on flow characteristics of the gasoline spray was investigated. Flow
characteristics of the gasoline spray near the nozzle exit of direct injection (DI) injector was investigated. From the
velocity data obtained by PIV, difference of flow characteristics between the spray (two-phase jet) and the single
phase jet was evaluated. It was found that ambient gas density contributed to development of the mixing layer of
the spray. In the region from z=0mm to z=4mm, the axial velocity of the spray at ρ =1.1kg/m slightly increased with
a
3
an increase of the axial distance. The similarity of the normalized velocity distribution of the spray maintained even
though distance from the nozzle was different. Moreover, the normalized velocity distribution of the spray was
similar to that of single phase jet.
Introduction
In a direct injection spark ignition (DISI) engine, gasoline fuel is injected into a
combustion chamber, and stratified charge combustion near a spark plug can be generated by
using a wall guided concept and so on. It is possible to achieve lean burn concept of the engine.
Moreover, air temperature in the combustion chamber can be lower than that in a port fuel
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
injection (PFI) engine due to latent heat of fuel evaporation, and then efficiency of the engine
become high as compared with the PFI engine.
Recently, injection pressure of gasoline fuel in a direct injection spark ignition (DISI)
engine tends to increase as compared with a conventional DISI engine in order to promote
atomization of the fuel. Moreover, high boosted engine by a turbo charger tends to be applied in
terms of downsizing concept in order to improve the thermal efficiency of the engine. It means
that ambient gas density in the combustion chamber is higher than that of the engine without the
high boost. Here, injection pressure of the fuel and the ambient gas density is related to
atomization process of the fuel strongly, and the atomization of the fuel affects the mixture
formation between the fuel and air. Therefore, it is necessary to understand the atomization
process near the nozzle in order to consider strategy of a high thermal efficiency engine.
In the past, many researchers have investigated atomization characteristics of the gasoline
spray. Han et al.[1] carried out a numerical study of air-fuel mixing in a direct injection spark
ignition (DISI) engine. They reported that the intake flow in the combustion chamber strongly
affected the mixture formation between air and fuel. Lee et al.[2] investigated spray
characteristics such as spray development in the direct-injected (DI) gasoline engine by
analyzing spray tip penetration. They clarified that spray tip penetration and fuel-air mixing
were influenced by high injection pressure of the fuel spray. Mitroglou et al.[3] investigated
stability of the gasoline spray ejected from multi-hole DI injector under various engine operation
conditions. They found that the stability of the spray from DI injector was higher than that of a
swirl pressure atomizer for a gasoline engine. Yamakawa et al.[4] investigated effect of nozzle
hole diameter and L/D in hole gasoline injector on the mixing formation between fuel and air by
using laser absorption scattering (LAS) method. From the result obtained in this study, they
clarified that the mixture formation depended on the nozzle diameter and L/D. It was suggested
that the formation was related to turbulence of fuel flow inside the nozzle due to cavitation
formation.
In the previous works for the gasoline spray, there was a few report regarding flow
characteristics of the gasoline spray near the nozzle exit even though its characteristics strongly
was related to the atomization process of the spray. Kawahara and Tomita [5] investigated flow
characteristics of the gasoline spray at downstream from 40mm by using phase doppler
anemometer (PDA). However, PDA was point measurement even though velocity and droplet
diameter of the fuel were measured simultaneously. Velocity field of the spray was desired in
order to understand the atomization process.
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
Particle image velocimetry (PIV) is the promising way for measurement of velocity field.
The spray ejected from the DI injector is high speed near the nozzle exit, and the spray is
unsteady flow. Therefore, PIV measurement with high spatial and temporal resolutions is
desired. However, it is not sutable to measure the velocity field of the spray near the nozzle exit
by using the typical PIV[6] and the conventional time-resolved PIV technique[7] due to
insufficient spatial and temporal resolutions measurement.
In this study, in order to investigate effect of injection pressure and ambient gas density on
flow characteristics of the gasoline spray near the nozzle exit, ultra-high speed PIV technique
was applied to velocity field measurement of the spray by using the ultra-high speed camera and
the CW laser. From the velocity data obtained by PIV, difference of flow characteristics between
the spray (two-phase jet) and the single phase jet was discussed.
Experimental Setup
Schematic view of experimental set-up is shown in Fig. 1. It consisted of a fuel injection
system, a high pressure vessel and an optical system for visualization of a gasoline spray. The
high pressure vessel was filled with N gas. Test fuel was isooctane instead of gasoline. Using a
2
multiple hole gasoline nozzle with a single shot injection system, test fuel was injected into the
pressure vessel. For velocity measurement inside the spray, continuous laser (CW) was utilized
as a light source, and a laser-light sheet was formed. Then tomographic images of the spray was
obtained. The thickness of laser-light sheet was adjusted and set on less than 1mm to obtain high
quality images for PIV analysis. A digital high speed video camera (Shimadzu, HPV-X2) was
arranged in perpendicular position to the laser-light sheet, and sequential images of the spray
were captured. Experimental conditions are shown in Table 1. Ambient gas temperature in the
high-pressure vessel was 300K, and then ambient gas densities ρ were set on 1.1kg/m , 5.8
a
3
kg/m and 11.6 kg/m under the temperature condition. Injection period was set at 3.0ms.
3
3
Injection pressure was 13MPa and 20MPa. Time-resolved PIV analysis was applied to measure
the velocity field inside the spray. For PIV analysis, correlative tomographic images with narrow
time interval were required. Consequently, an ultra-high frame rate of 2,000,000 f.p.s was needed
Table 1 Experimental condition
Figure 1 Experimental setup
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
for PIV analysis in order to capture PIV image near the nozzle hole. Spatial resolution was
fixed constant at 0.017mm/pixel regardless of ambient gas density and injection pressure. In this
PIV analysis, interrogation spot size was 17pixels×17pixels which corresponding with
0.29mm×0.29mm of real image, and interrogation spot area of 50% was overlapped with the
adjacent interrogation area. It means that grid space in PIV analysis was 0.15mm.
Figure 2 shows sequential instantaneous velocity field of the spray for P =20MPa and
inj
ρ =1.1kg/m . In this study, the nozzle which has 6 holes was utilized, and the velocity field of the
a
3
spray ejected from one of 6 holes was measured. Near the nozzle exit, velocity field of the spray
ejected from the DI injector could be obtained by time-resolved PIV using the ultra-high speed
camera, and the spray velocities from z=0 to z=4mm were mainly velocity component of vertical
downward direction. Then, in this study, the spray axis direction was defined as the vertical
downward direction. From those results, mean velocity fields of the spray were analyzed for
various injection pressures and ambient gas densities.
z
0mm
1mm
2mm
tinj=4,000ns
tinj=10,000ns
tinj=16,000ns
100m/s
3mm
4mm
Figure 2 Instantaneous velocity field of direct injection spray in time series.
Mean velocity of axial component on the spray axis for various injection pressure and
ambient gas densities.
In the analysis of the mean velocity field, sequential instantaneous velocity fields of the
spray during injection period of quasi-steady state were averaged, and then the velocity fields of
6 shot sprays were averaged. Figure 3 shows the mean velocity filed of the spray from the nozzle
exit to 4mm downstream at P =20MPa and ρ = 1.1kg/m . According to knowledge regarding the
inj
a
3
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
velocity distribution of a free jet, the axial velocity distribution in radial direction was well
expressed with a Gaussian function. In the past, Zjinen et al. [8] applied a Gaussian function for
the fitting analysis of the velocity distribution of a single-phase free jet. In this study, by using
200
ρa =1.1kg/m3
ρa =1.1kg/m3
Pinj =20MPa
Pinj =20MPa
Mean velocity v [m/s]
Mean velocity v [m/s]
200
100
Z=1.0mm
Z=2.0mm
Z=3.0mm
Gaussian function
Z=2.0mm
100
Z=4.0mm
0
−2
0
Radial distance r [mm]
2
0
−2
0
Radial distance r [mm]
2
Figure 3 Mean axial velocity distribution of
Figure 4 Radial distribution of mean axial
the spray near the nozzle exit.
velocity fitted with a Gaussian function.
(P =20MPa, ρ = 1.1kg/m )
inj
3
a
data of the mean velocity field, radial distributions of axial velocity component were also fitted
with Gaussian function expressed with Eq.(1) as shown in Fig.4.
⎧ (r − rc )2 ⎫
v(r ) = v0 exp⎨
⎬
⎩ σ ⎭
(1)
v(r) is axial velocity of the spray, v0 is maximum axial velocity, rc is real origin in radial direction,
and σ means the standard deviation of velocity distribution in the Gaussian function. From Fig. 4,
velocity distributions of the spray in radial direction could be well fitted with the Gaussian
function as well as that of single-phase free jet.
From the fitting analysis with the Gaussian function, maximum axial velocity distribution
along the spray axis (i.e. velocity distribution on the center axis) could be obtained. The effect of
ambient gas density and injection pressure on the maximum velocity distribution in the spray
axis was investigated. Figure 5 shows the maximum velocity distribution from the nozzle exit to
z=40mm for various ambient gas densities and injection pressures. The axial velocity at
P =13MPa and ρ =1.1kg/m slightly increased with an increase of the axial distance from z=0mm
inj
a
3
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
to z=4mm. In the downstream from z=4mm, the velocity decreased with an increase of the
distance from the nozzle. As for P =20MPa and ρ =1.1kg/m , the axial velocity from z=0mm to
inj
3
a
z=4mm was almost the same as well as the increase trend of P =13MPa and ρ =1.1kg/m . In the
inj
3
a
study of the gas jet, Shakouchi et al.[9] reported that the axial velocity of the jet near the nozzle
exit also increased with an increase of the axial distance, and they concluded that it was caused
by contraction flow inside the nozzle. By considering this report, it seems that the increase trend
of the axial velocity in the spray was also due to the contraction flow inside the nozzle. On the
other hands, at P =13MPa, 20MPa and ρ =11.6kg/m which was high ambient gas density
inj
3
a
condition, the region where the axial velocity hardly changed disappeared, and the axial velocity
dramatically decreased away from the nozzle. It means that its region disappeared with an
increase of shear force between ambient gas and the spray droplets. In addition, at ρ =5.8kg/m ,
3
a
the region where the axial velocity hardly changed decreased by increasing the injection pressure
from P =13MPa to P =20MPa. It means an increase of shear force between ambient gas and the
inj
inj
spray droplets by the promotion of atomization of the spray. Therefore, development of mixing
layer in the spray was promoted with ambient gas density increase as compared with injection
200
Pinj =13MPa
ρa =1.1kg/m3 ,Ta =300K
ρa =5.8kg/m3 ,Ta =300K
ρa =11.6kg/m3 ,Ta =300K
150
ρa =1.1kg/m3
100
ρa =5.8kg/m3
50
ρa =11.6kg/m3
0
0
5
10
15
20
25
30
35
40
Distance from nozzle z [mm]
45
Mean velocity of axial component v0 [m/s]
Mean velocity of axial component v0 [m/s]
pressure increase.
(a) P =13MPa
inj
200
ρa =1.1kg/m3 ,Ta =300K
ρa =5.8kg/m3 ,Ta =300K
ρa =11.6kg/m3 ,Ta =300K
Pinj =20MPa
150
ρa =1.1kg/m3
100
ρa =5.8kg/m3
50
ρa =11.6kg/m3
0
0
5
10
15
20
25
30
35
40
45
Distance from nozzle z [mm]
(b) P =20MPa
inj
Figure 5 Comparison of mean velocity in axial component v0 for various ambient gas densities.
Normalized velocity distribution in radial direction
According to Fig.4, the radial distribution of the axial velocity in the spray was well fitted
with Gaussian function as well as a single-phase flow such as a gas jet. Normalized axial velocity
distributions in radial direction for various ambient gas densities and injection pressures were
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
evaluated in order to compare with the velocity distribution of the single-phase jet. Figure 6
shows the radial distribution of axial velocity normalized with the maximum axial velocity for
the various distance from the nozzle. v/v0 in vertical axis means that axial velocity of the spray
was normalized with maximum spray velocity obtained with Gaussian fitting analysis. r/rv0/2 in
horizontal axis means that radial distance r was normalized with the radial distance where half
of the maximum velocity was shown. Regardless of ambient gas densities as shown in Fig.6(a),
(b) and (c), the normalized velocity distributions were asymmetry. The velocities in left side with
respect to center axis were higher than that in the right side. It was caused by the entrainment of
the spray ejected from the adjacent nozzle hole. Therefore, the velocity distribution in the right
side was focused on the discussion. From the figure, it was found that similarity for radial
distribution of the axial velocity maintained for various ambient gas density conditions even
though the distances from the nozzle and injection pressure were changed. In the figure, the
normalized velocity distribution obtained with PIV was compared with that of the Goertler
solution[10] for a single phase jet. The Goertler solution was depicted with a solid line. From the
comparison, it was found that the normalized velocity distribution of the spray was almost the
same as that of the single phase jet even though the spray ejected from the DI injector was twophase flow. It suggested that feasibility for application of ultra-high speed PIV technique to
measurement of velocity field near the nozzle exit was confirmed.
1
1
Pinj [MPa]
13 20
Z=5.0mm
Z=10.0mm
Z=20.0mm
Z=30.0mm
Z=40.0mm
ρa
Ta =300K
0
−2
0
2
Normalized radial distance r/rv0/2 [-]
(a) ρ =1.1kg/m
a
3
Goertler
Normalized velocity v/v0 [-]
0.5
=1.1kg/m3
1
Goertler
Normalized velocity v/v0 [-]
Normalized velocity v/v0 [-]
Goertler
0.5
Pinj [MPa]
13 20
Z=5.0mm
Z=10.0mm
Z=20.0mm
Z=30.0mm
Z=40.0mm
ρa =5.8kg/m3
Ta =300K
0
−2
0
2
Normalized radial distance r/rv0/2 [-]
(b) ρ =5.8kg/m
a
3
0.5
Pinj [MPa]
13 20
ρa =11.6kg/m3
Ta =300K
0
−2
Z=5.0mm
Z=10.0mm
Z=20.0mm
Z=30.0mm
Z=40.0mm
0
2
Normalized radial distance r/rv0/2 [-]
(c) ρ =11.6kg/m
a
3
Figure 6 Normalized velocity distribution in radial distance for various ambient gas densities vs.
Goertler solution.
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
Conclusions
In this paper, flow characteristics for the gasoline spray ejected from the DI injector was
investigated by using ultra-high speed PIV technique. The results obtained in this study are as
follows;
1.
In the region from z=0mm to z=4mm, regardless of injection pressure, the axial velocity of
the spray at ρ =1.1kg/m slightly increased with an increase of the axial distance.
a
2.
3
The region where the axial velocity hardly changed disappeared with an increase of ambient
gas density regardless of injection pressure.
3.
The similarity of the normalized velocity distribution in the spray maintained even though
distance from the nozzle was changed.
4.
The normalized velocity distribution of the spray was similar to that of single phase jet.
Acknowledgments
This work was supported by Council for Science, Technology and Innovation(CSTI), Crossministerial Strategic Innovation Promotion Program (SIP), “Innovative Combustion Technology”
(Funding agency: JST).
References
[1] Z.Han, L.Fan, R.D.Reitz, M.Takahashi “Multidimensional Modeling of Spray Atomization
and Air-Fuel Mixing in a Direct-Injection Spark-Ignition Engine” SAE Technical Paper 970884,
1997.
[2] C.S.Lee, M.S.Chon, D.S.Kim “Spray Structure and Characteristics of High-Pressure Gasoline
Injectors for Direct-Injection Engine Applications” Atomization and Sprays, vol.11, pp.35-48,
2001.
[3] N.Mitroglou, J.M.Nuri, M.Gavaises, C.Arcoumanis “Spray Characteristics of a Multi-Hole
Injector for Direct-Injection Gasoline engines” International Journal of Engine Research, 7(3),
pp.255-270, 2006.
[4] K.Sato, J.Kanzaki, M.Yamakawa, K.Nishida “Mixture Formation Process of Spray Injected by
Hole Type Injectors for D.I. Gasoline Engines (Effects of Diameter and L/D of Injector Hole)”
Atomization and Sprays. vol.16, No.55, pp47-55, 2007
[5] N.Kawahara and E.Tomita “Simultaneous Measurements of Droplet Diameter and Velocity
Using Phase Doppler Anemometer (PDA/PDPA)” The Japan Society of Fluid Mechanics.
NAGARE27, pp303-310, 2008
18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics・LISBON | PORTUGAL ・JULY 4 – 7, 2016
[6] G Rottenkolber, J. Raposo, K. Dullenkopf, W. Hentscel, S. Wittig, U. Spicher, W. Merzkirch,
Spray analysis of a gasoline direct injector by means of two-phase PIV, Exp. Fluids, 32,2002.
[7] Lulin Jiang and Ajay K. Agrawal, Spray feature in the near field of a flow-blurring injector
investigated by high-speed visualization and time-resolved PIV, Exp. Fluids, 56, 103, 2015.
[8] Hinze, J. O., and Van der Hegge Zijnen, B. G., Transferof heat and matter in the turbulent
mixing zone of an axially symmetrical jet, J. Appl. Sci. Res., vol. A1, pp.425–461, 1949.
[9] T.Shakouchi, M.Kito, T.Sakamoto, K.Thujimoto, T.Ando “Flow Analysis of Orifice Free Jet
(Effects of Contraction Ratio)” Transactions of JSME. Vol.74, No.737, pp42-48, 2008.
[10] Rajaratnam, N., (1876) Turbulent Jets, Amsterdam and New York, Elsevier Publishing Co, p.
27-49.