EPA/AA/TDG/94-03
Technical Report
High-Speed/High-Resolution Imaging Of Fuel Sprays
From Various Injector Nozzles For Direct Injection Engines
by
Fakhri J. Hamady
Jeffrey P. Hahn
Karl H. Hellman
Charles L. Gray, Jr.
September 1994
NOTICE
Technical Reports do not necessarily represent final EPA decisions or
positions. They are intended to present technical analysis of issues
using data which are currently available. The purpose in the release of
such reports is to facilitate the exchange of technical information and to
inform the public of technical developments which may form the basis
for a final EPA decision, position or regulatory action.
U. S. Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
Regulatory Programs and Technology Division
" Technology Development Group
2565 Plymouth Road
Ann Arbor, MI 48105
-------
r
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ANN ARBOR. MICHIGAN 48105
OCT 17 1994
OFFICE OF
AIR AND RADIATION
MEMORANDUM
SUBJECT: Exemption From Peer and Administration Review
FROM: Karl H. Hellman, Chief
Technology Development Group
TO:
Charles L. Gray, Jr., Director
Regulatory Programs and Technology Division
The attached report entitled "High-Speed/High-Resolution Imaging of
Fuel Sprays from Various Injector Nozzles for Direct Injection Engines,"
(EPA/AA/TDG/94-03) describes the results of our continuing in-house
investigation to quantify the transient fuel spray properties from new concept
injectors and technology for engine applications. A high-speed laser sheet
imaging system and a laser diffraction technique were used to integrate the
visual observations with the droplet size measurements from various injector
nozzles.
Since this report is concerned only with the presentation of data and its
analysis, and does not involve matters of policy or regulations, your
concurrence is requested to waive administrative review according to the
policy outlined in your directive of April 22, 1982.
Concurrence:
Charles L. Gray, J^.'^DTrector, RPT
Date:
-------
Table of Contents
Page
Number
I. Summary 1
II. Introduction 1
III. Experimental Facilities 3
A. Engine Assembly 3
B. Fuel Injection Systems 4
C. High-Speed/High-Resolution Imaging System. ... 6
D. Malvern Particle Sizer 8
IV. Results and Discussion 8
V. Conclusion 21
VI. Recommendations and Future Efforts 22
VII. Acknowledgments 23
VIII. References 23
-------
I. Summary
A high-speed/high-resolution imaging technique and analysis were
applied to study fuel injector spray timed evolution in ambient air and in
a motored single-cylinder engine. Alcohol fuel was injected from a low-
and mid-pressure fuel injection systems. The two injection systems with
various nozzles were designed for use in the EPA/NVFEL program to
develop clean and efficient engines that use alternative fuels.
A 15W copper vapor laser with a fiber optic delivery system
synchronized with a high-speed drum streak camera was utilized to expose
films at 5,000 frames per second (fps). The spray characteristics were
investigated at injection pressures of 0.68 MPa (100 psi) and 15.0 MPa
(2,200 psi) and injection duration range of 3-5 ms. A sequence of
successive frames was selected from the films to examine the influence of
the injector parameters and the valve lift on the atomization process. The
spray penetration was quantified by analyzing the high-speed films. The
mean droplet size distributions from the nozzles were measured by using
the Malvern particle sizer.
This report is part of an ongoing investigation to obtain information
on the transient spray properties from new concept injectors and
technology for direct injection engine applications. Experimental results
indicated that the low-pressure injectors yielded large droplet and the
mid-pressure injectors generated small droplet on the average of 10
microns (nm). This is due to the spray convergence and droplet
coalescence in the low-pressure injector and to the high fuel jet
momentum, and consequently, the good atomization in the mid-pressure
injector. The visual observations provided valuable information about the
spray features and are used to interpret the droplet size measurements.
II. Introduction
An important area for successful further development of the internal
combustion engine involves convenient control of the fuel-air mixture at
the time that combustion is initiated to promote fuel efficiency, reduce
exhaust emissions, and provide flexibility of operation in a broader range
of speeds and loads. The use of advanced fuel injection systems appears
to be an attractive approach that could improve the engine combustion
process and reduce emissions. The emphasis in this effort is to provide in
a systematic approach the needed experimental data on the spray
dynamics, including atomization, vaporization and mixing, from low- and
mid-pressure direct cylinder fuel injection systems. By better
understanding the simultaneous effects of the injector controlling design
parameters, it is hoped that this work will assist the ongoing fuel
-------
-2-
injection and engine development at the Environmental Protection
Agency's National Vehicle and Fuel Emissions Laboratory (EPA/NVFEL).
The advent of high-pressure fuel injection systems has led to a large
number of system arrangements, like the dual-injector, sequential
injection and the two-stage injector systems. With these arrangements
there have been several studies concerned with the injection system
performance. Many important aspects of the spray events were tested
under various ambient pressure conditions utilizing different imaging
techniques. Savey, et al. [1]* used stroboscopic flash photography to
visualize the spray characteristics of a unit injector operating with water
injection into the ambient air. The photographic evidence used in
conjunction with dynamic laser attenuation measurements allowed the
estimation of the spray penetration distance and cone angle. A
shadowgraphic technique was employed by Kato, et al. [2] for the analysis
of non-evaporating fuel spray characteristics. Fuel was injected into a
high pressure nitrogen-filled chamber where spray events were captured
with a high-speed camera. Spray droplet diameter and equivalence ratio
were analyzed using shadow photographs. In-cylinder tests of the
injection process were also conducted, but these did not include any
visualizations. Katsura, et al. [3] photographed the impingement of a
diesel spray on a flat wall in a high-pressure chamber with both
transmitted and scattered light. The effect of ambient density was studied
and it was found that the spray droplet size decreases with increasing
ambient density. The synchronization of a copper vapor laser sheet with a
high-speed camera was employed by Hamady and Morita. [4,5]
Experiments were performed by injecting jet-type fuel through a single-
hole nozzle at high-injection pressure into atmospheric air and in a direct
injection rotary engine. Several event-to-event variations in the injection
process were observed that had not previously been noted.
Engine studies using pressured chambers with air flow were
considered to simulate in-cylinder flow situations. Hiroyasu, et al. [6]
used high-speed Schlieren photographs to observe spray vaporization and
analyze diesel spray behaviors. A repetitive micro-flash used with a drum
camera allowed Yoshikawa, et al. [7] to observe spray behavior. They
also conducted in-engine studies and utilized emission measurements to
optimize the injection conditions. In a firing engine, Werlberger and
Cartellieri [8] used the endoscopic high-speed combustion photography
technique to study wall jet development, the role of swirl, and the effect
of the pilot injection on mixture formation and combustion.
With advances in measuring techniques, other optical systems were
developed and applied, namely, laser diffraction and laser sheet
illumination [9-11] for determining the droplet size distributions, spray
Numbers in brackets designate references at
end of report.
-------
-3 -
penetration, and cone angle. Phase Doppler anemometry [12,13] was used
for examining the dynamic behaviors of the fuel jet such as velocity and
particle size. Laser induced exciplex fluorescence [14,15] and laser
induced Rayleigh scattering [16] were also devised to provide planar
information on the spray characteristics, including vapor and liquid
phases and fuel vapor concentration, of the fuel-air mixture.
In view of the above discussion of various methods to quantify the
spray structures, high-speed/high-resolution imaging is needed to describe
the fuel injection process and to integrate the visual observations with the
quantitative measurements. For instance, spray regions of interest can be
identified and examined for their transient behavior. In this investigation,
emphasis was placed on the analysis of the injector's controlling
parameters which influence the fuel spray structures. Several nozzles of
different designs were manufactured to achieve injection of different
spray patterns. High-speed imaging and laser diffraction systems will be
devised to examine and quantify the spray features of interest. On the
basis of the spray analysis, the information will be used to suggest design
changes which will enhance atomization, spray formation, and the droplet
size distributions. Information of this type is not only important, but
provides a valuable database for the development and evaluation of
concepts that exploit fluid dynamic instabilities or other approaches to
enhance atomization.
III. Experimental Facilities
A. Engine Assembly
For this work, an OH160 Tecumseh single-cylinder air-cooled
engine was modified by bolting an extension on the original piston top.
The extended piston is contained within a cylinder which has three mounts
to support quartz windows for imaging and droplet size measurements.
The retrofitted piston with a cylindrical bowl cut into it is equipped with
a quartz insert to allow optical access through the piston. An adjustable
mirror holder including a mirror is located inside the extended piston to
provide optical access into the combustion chamber through the piston.
The Tecumseh cylinder head was replaced by an in-house custom-
fabricated cylinder head to adapt different types of injectors. Figure 1
shows that a fuel injector can be centrally positioned so that its centerline
coincides with the cylinder axis.
The spark plug is mounted close to the injector tip. The valves were
driven by extending the push rods from the base engine with the same
stock timing. Standard piston rings and intake and exhaust manifolds were
used in this assembly. Table 1 shows the specifications of the modified
-------
-4-
Injector
Cylinder
Piston
Quartz
insert
Combustion
chamber
Figure 1 Engine modified cylinder head assembly
engine. The engine assembly was constructed to allow easy changing of
the engine head, piston and quartz. The assembly was mounted onto a
heavy steel plate along with an 11.2 kW (15 hp) variable-speed electric
motor coupled to the engine crankshaft. An IC5460 engine electronic
control unit was used to control the ignition timing and fuel injection
operation.
Table 1 Modified engine specifications
Bore x Stroke (mm)
Displacement (cm3)
Diameter of piston bowl (mm)
Depth of piston bowl (mm)
Compression ratio
Number of valves
88.9 x
73.0
453.3
60.0
10.0
12.0
2
B. Fuel Injection Systems
Two fuel injection systems were used in this investigation, The first
operating at 0.68 MPa and the second at 15.0 MPa fuel pressure. The first
system is equipped with a nozzle having an outward opening spring-
loaded poppet valve which is operated by an electronically controlled
integral solenoid. The outward opening uses the cylinder pressure to hold
the valve against its seat to improve closure process. Due to the low-
injection pressure the injection events will occur during the intake and
early part of compression stroke. The nozzle assemblies are shown in
Figure 2.
In the mid-pressure system, two injector types were used. The first
injector is equipped with a nozzle having an outward opening spring-
loaded poppet valve which is operated by the fluid hydraulic pressure.
-------
-5-
(No scale)
Poppet
valve
tt
p—a
B
Valve
seat
ITOMJ
Peripheral
holes
Cross-sectional view
Figure 2 Low-pressure injector nozzle assemblies
The nozzle assembly is shown in Figure 3. The nozzles and springs can be
varied to study different injector concepts.
The second injector has an inward opening needle valve actuated by
the pressure acting against spring force. The fuel injection system has full
electronic control of the solenoid valve. Initially, the second fuel injector
nozzle tip design incorporated a cruciform slot geometry. The fuel spray
from this nozzle utilized the swirling effect imparted to the fuel as it
(No scale)
Poppet
valve
f/ \u_ Valve
LJlseal
f* Valve lift
Figure 3 Mid-pressure injector poppet type nozzle assembly
-------
-6-
passes through angular slots on the nozzle needle. As shown in Figure 4,
a set of different nozzle designs have been used, in an attempt to improve
the droplet size distributions.
(No scale)
Swirl
vanes
B
1
Vertical
vanes
W...:
Pintle
Figure 4 Mid-pressure injector nozzle assemblies
C. High-Speed/High-Resolution Imaging System
The high-speed imaging system is shown schematically in Figure 5.
It consists of a copper vapor laser, fiber optic delivery system, light sheet
forming optics, laser drum streak camera, and a laser V shot controller.
The copper vapor laser (CVL) Model CU15-A, is a gas discharge
device of 15W average output power manufactured by Oxford Lasers. It
emits short pulses of 5-30 kHz at pulse energy 2.75 (max.) in the green
and yellow regions of the visible spectrum at wavelengths of 511 and 578
nm. The pulse duration is approximately 20 ns and pulse jitter is ± 2 ns.
The laser main components are a plasma tube, high-voltage DC power
supply and pulsed discharged system (thyratron driver and thyratron). The
thyratron driver is a high-voltage, high-repetition pulse generator which
provides trigger pulses to the thyratron. The driver can be operated using
an internal oscillator or by applying the external pulse to the trigger
select.
-------
-7-
The laser beam (25 mm in diameter) can be directed toward the
region of interest by the fiber optic delivery system. The fiber diameter
used is 600^im. The fiber has a standard mount adapter (SMA) termination
at both ends. Overall injection transmission efficiencies are in excess of
70 percent for fiber diameters greater than or equal to 600 jam.
Injector -
Camera
controller
Laser 'n1 shot
controller
External events
(from injector)
Drum camera
0-ring
Rulon ring
Laser light sheet
Cylinder sleeve
s >-J
Fib
CVL Laser
Cylinder
head
Piston
Quartz
insert
Piston
extension
Lower
piston
Sheet forming optics
Base engine
Figure 5 High-speed/high-resolution imaging system and optical engine assembly
The camera is a Cordin Model 321 laser drum camera, which
exposes a streak/frame record of 25 mm wide and 1,000 mm long at a
maximum rate of 0.30 mm/us on a standard 35 mm, 400 ASA color
negative film and has an effective aperture of f/3.2. It is designed for
framing use with the CVL pulsed laser at 35,000 frames per second
(max.). The film is loaded and recovered with a single-shot, daylight
cassette. The camera has a reflex viewer to simplify focus. The camera
housing is evacuated to at least 20 mm Hg absolute pressure during
operation. This reduces image aberrations from density variations in the
air and thermal effects of friction. The camera operational parameters are
controlled by its controller Model 447A.
The laser 'n1 shot controller is used to switch the thyratron driver of
the CVL from internal to external modes. In the external mode, the pulse
generator of the controller triggers the laser the number of pulses at the
frequency desired during the filming process. At the end of the film, the
controller switches the laser back to the internal mode. The controller can
also be used to synchronize fuel injection events and ignition timing with
the camera. Operation of the system can be used to study time-resolved
-------
-8-
fuel sprays during single-injection events and to study individual engine
combustion events.
D. Malvern Particle Sizer
The Malvern particle sizer is an optical system which uses line-of-
sight measurement through the spray. It is applied to measure the
ensemble characteristics of the spray like the droplet size distributions.
The system measures simultaneously the intensity of the forward
diffracted light at several angles as the laser beam passes through the
spray.
The Malvern software can provide calculated information such as
the Sauter mean diameter (SMD) and other derived parameters including
volume concentration and specific surface area. More details on the
accuracy and limited applications of the Malvern system are found in
references [18-20].
IV. Results and Discussion
This section presents an analysis of alcohol fuel sprays from a low-
and mid-pressure injection systems injected in ambient air at atmospheric
pressure and in-cylinder at engine speed of 1,000 rpm. High-speed
imaging of the spray events were taken at 5,000 fps in axial planes along
the nozzle axis. From the flow visualizations, important spray features
were identified and integrated with the droplet size measurements and
injector controlling parameters to quantify spray properties.
In the first phase of experiments, sprays from ten different
geometric configurations of nozzle "A" (see Figure 2) of the low-pressure
injection system were examined. The nozzle tip configurations were based
on flush-closed and flush-open between the poppet valve and nozzle body.
Figure 6 shows successive frames of the entire injection event from
nozzle "A" with a flush-open configuration. This figure demonstrates
clearly the detrimental effect of spray convergence and droplet
coalescence on the atomization process and, consequently, further in the
event formation of large droplets dominated the spray pattern. The visual
observations from the other nine nozzles showed similar phenomena
prevailing the spray structures; therefore, photographs will not be
presented.
To quantify the spray features described in Figure 6, droplet size
measurements were conducted along the spray axis. Figure 7 presents the
variation in droplet size at 2.5 and 5.0 cm from the nozzle exit. From the
measurements large droplets were found to exist through the end of
-------
-9-
Figure 6 Successive frames of one injection event injected from the low-pressure injector (nozzle "A")
into the atmosphere at 0.68 MPa injection pressure and 5000 fps
-------
- 10-
^_^
E
C/5
•1
tJ
"5,
2
Q
£.
91
'35
p l
o
Q
JJU
300
250
200
150
100
50
A
Vj
"inn
JUU
250
200
150
100
50
(
(a) Flush closed nozzle
Distance from nozzle
I A A 2.5 cm
- \ « — -« 5.0cm
1 V-°-^
*^ i / \ *
; --A^A \/ \/ \
Al \ ^>.j/ V V *^_«
— Viu • ^A — *"• \ ' '
f '^ 'A. \ / \
\ \ A
\ * ^
\.
-
, , , , | , , , , | , , , , |
(b) Flush open nozzle
~~ *'***+-*
\
\ * ^ ^^ - vJ ^
r • ^'^ .^i| J ^"*
- \/ JN
1 1
1 1
1 1
1 1
—
-,,,,,,,,,!,,,, |
) 2 4 6 8 1
Time (ms)
Figure 7 Variation of droplet size with time from the start of injection
for the low-pressure injector
-------
-11-
injection. It was also apparent from the experiments that the atomization
was influenced by the inconsistent performance of the injector solenoid
valve design, plunger travel, and dynamic response. Due to these adverse
effects of the spray convergence and injector performance on the
atomization process, limited results are presented in this report.
However, further injector developments and modifications to promote the
atomization are underway.
In the second phase of experiments, sprays from the mid-pressure
injectors were analyzed. For poppet valve injector depicted in Figure 3, a
sequence of successive frames were selected to demonstrate the spray
development in ambient air at 0.7 mm valve lift. At the start of injection,
Figure 8 shows a narrow cone-shaped spray formed at the nozzle exit. In
the early stage the fuel jet moved outward from the nozzle, where the
leading edge traveled with a high-penetrating velocity before dissipating.
After reaching the breakup length the spray penetration velocity
decreased significantly. This spray behavior is a consequence of the
aerodynamic interaction between the fuel jet and the ambient air, which
usually leads to enhancement of the fuel atomization and fuel-air mixing
process. As the injection event proceeds the fuel jet continues to
penetrate the ambient air but at a lower rate. At the outer surface of the
jet the mixing region grows in breadth and the liquid core disintegrates
into droplets. Between 2.2 and 4.6 ms, Figure 5 demonstrates these
features of the fuel jet divergence and the decrease in penetration
velocity. In the final stage of the injection event (after 4.6 ms) large
droplets can be observed due to valve closing and droplet coalescence.
Experiments were also conducted for valve lift of 0.3 mm. The general
spray features resemble those described previously; however, photographs
indicated a better atomization process as displayed in Figure 9.
From the high-speed films, the spray penetration distances were
measured and plotted for two poppet valve lifts in Figure 10. It is
important to note that after 2 ms the slopes of these curves start to
decrease, quantifying the visual observation of liquid core breakup. The
figure also indicates that by limiting the valve lift from 0.7 to 0.3 mm the
corresponding penetration distances are increased. This can be attributed
to restricting the nozzle exit which elevates the pressure differential at
the valve seat and consequently the fluid speed.
The spray mean droplet sizes represented by SMD were measured
along the spray axis at different distances from the nozzle exit for 5 ms
pulse duration. The droplet size distribution profiles in Figures ll(a) and
ll(b) show a decrease in the droplet size with the measurement delay
times from the start of injection, and an increase near the end of injection
event. The large droplet formation at the start of injection is caused by
-------
-12-
0.2 0.4
0.6
0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
(a) Successive frames from the start of injection
2.6ms
3.4 3.6
3.8
4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4
(b) Successive frames through the end of injection
5.8ms
Figure 8 Selected frames of one injection event injected into the atmosphere from the mid-pressure
poppet valve injector at 15MPa injection pressure (0.7 mm valve lift and 5000 fps)
-------
-13-
0.8 1.0 1.2 1.4 1.6 1.8 2.0
(a) Successive frames from the start of injection
2.4 2.6ms
6.6 6.8
7.0
7.2 7.4 7.6 7.8 8.0 8.2 8.4
(b) Successive frames through the end of injection
8.8
9.0ms
Figure 9 Selected frames of one injection event injected into the atmosphere from the mid-pressure
poppet valve injector at 15MPa injection pressure (0.3 mm valve lift and 5000 fps)
-------
- 14-
IT
0
03
«
-------
-15-
o
N
80
70
60
50
40
J3 30
EX
s
Q 20
10
0
80
70
60
50
c/a
40
•g, 30
2
Q 20
10
t
(a) Valve lift = 0.70 mm
Distance from nozzle
A—A 5.0cm
e—• 7.6 cm
(b) Valve lift = 0.30 mm
8
10
0246
Time (ms)
Figure 11 Variation of droplet size with time from the start of injection
for the mid-pressure poppet valve injector
-------
-16-
s"
—
"a,
o
80
70
60
50
40
30
20
10
0
Distance from nozzle = 5.0 cm
Valve lift :*- A 0.3mm
o—e 0.7mm
t
2468
Time (ms)
Figure 12 Valve lift effect on the droplet size distribution
for the mid-pressure poppet valve injector
10
In this part of the experiments, high-speed imaging was applied to
characterize the sprays from three different nozzle assemblies "A", "B"
and "C" as shown in Figure 4. Nozzle "A" has swirl vanes and a single
hole of length to diameter ratio of 2 (1/d = 2, d = 0.2 mm), Nozzle "B"
has vertical vanes and a single hole of 1/d = 4 (d= 0.6 mm) and nozzle "C"
has swirl vanes and a pintle valve. For nozzle "A", Figure 13 displays one
injection event in ambient air at 15 MPa fuel injection pressure. In this
figure spray development is analogous to that described in Figure 9.
However, photographs indicate better atomization during the early stage
of injection due to nozzle geometry, fluid velocity and valve opening. In
addition, this spray pattern combined with the fluid droplet air interaction
continued to promote the atomization process during the entire injection
event. At the valve closing relatively small droplets can be observed.
Under similar conditions, high-speed imaging was performed to
describe the spray patterns from nozzles "BH and HC". Photographs from
both nozzles showed resemblance to the general spray features from
nozzle "A". Quantitative description of the spray penetration distances
and droplet size measurements are presented below.
Figure 14 presents a comparison of penetration distances from the
three nozzles. It is important to note that nozzle "A" has a larger
-------
-17-
0.2 0.4
0.6
0.8 1.0 1.2 1.4 1.6 1.8 2.0
(a) Successive frames from the start of injection
2.2
2.6ms
7.4 7.6
7.8
8.0 8.2 8.4 8.6 8.8 9.0 9.2 9.4
(b) Successive frames through the end of injection
9.8 ms
Figure 13 Selected frames of one injection event injected from the mid-pressure single hole nozzle
(nozzle "A") into the atmosphere at ISMPa injection pressure (5000 fps)
-------
- 18-
penetration in the early stage of the injection process and consequently
higher velocity. This phenomenon led to better fuel atomization as
illustrated in Figure 13.
I
o
-------
- 19-
/u
60
/gv
3 50
1 40
1 30
•4-*
1 2°
P
10
0
70
,-s 6°
s
3 50
1 40
8
•a 30
"§• 20
P
10
0
: (a) Nozzle " AH
~ Distance from nozzle
-
: A— -A 5.0 cm
: • — • 7.6 cm
-
-
-
-.a ^...A
r ^^^^^^^ti^-^^-ft^i^^^^-^--*-*
- 1 i i i i i i i i i i i i i i i i i i i i i i i
: (b) Nozzle "BH
'- \
; \
i- v^
- ^.^
^ K^ + +^+*
!- ''^t^-i-^-^^-^"^' ~*^'
-
—
: I , I , 1 , , , , 1 , • , , 1 1 , 1 , 1 I I , 1
7fl
/u
60
E
•^
3 50
1 40
8
•a 30
w
t 20
p
10
0
: (c) Nozzle "CH
—
-
f \
L \ >»-»-^
- \ ^
; V X
— ^>VA \ ^
• \ ^ A~ -*-"*'' "*~*~*— «
r \A_^ ^^^..^.^-L.^--*-^*
—
** i i i i i i i i i i i i i i i i i i i i i i i i
024681
Time (ms)
Figure 15 Variation of droplet size with time from the start of injection
for the mid-pressure single hole injectors
-------
-20-
88° BTDC
86° BTDC
84° BTDC
82° BTDC
80° BTDC
78° BTDC
72° BTDC
70° BTDC
60° BTDC
Figure 16 Selected frames of one injection event at 5000 fps and injection pressure of 15MPa from
a poppet valve injector during compression in a motored single-cylinder engine at 1000 rpm
-------
-21-
In the present study, observation of the spray characteristics in the
engine is limited to the poppet-type mid-pressure fuel injection system.
The high-speed films were taken by using the optical engine assembly
shown in Figure 5. Selected photographs of the injected fuel are displayed
in Figure 16 at 15.0 MPa from the start of injection at about 88° CA
BTDC. In this set of photographs, the impinged fuel jet diffuses
symmetrically in a disk-shaped flow structure originating at the center of
the combustion chamber. It is also apparent from 72 to 60° CA BTDC that
the fuel distribution is uniform and diffuses symmetrically as the
compression proceeds. In this case, a combustion chamber with strong
squish flow could be used to enhance fuel-air mixing, improve dilution
tolerance and combustion stability. Consequently, a cylinder head with
high flow rate and reduced intake air resistance caused by high swirl
motion could be used to replace a conventional high swirl direct injection
cylinder head. Further development is currently underway at the NVFEL
lab to identify, evaluate, develop and utilize advanced injector
technologies in direct injection engine applications.
V. Conclusion
The spray characteristics of a low- and mid-pressure fuel injection
systems injected into the atmosphere and in-cylinder have been described
by analyzing high-speed films of the process. In addition to the
qualitative description of the sprays the films were used to measure fuel
penetration and give more physical aspects to the laser diffraction droplet
size measurements. Comparison of the spray features from the different
nozzle configurations provided useful information on the injector
parameters and its effect on the sprays. The following concluding remarks
can be made:
1. The high-speed imaging system developed at the EPA/NVFEL offers an
efficient method to test injectors and study transient spray behaviors.
This is to ensure that new concepts are adequately assessed for
feasibility and successful design.
2. From the low-pressure injector, the spray convergence and droplet
coalescence adversely influenced the atomization and, consequently,
contributed to the formation of large droplets.
3. For the poppet valve nozzle of the mid-pressure injector, the
penetration distances and droplet sizes were significantly influenced by
the valve lift. Smaller droplets were generated by decreasing the valve
lift without markedly altering the fuel flowrate.
-------
-22-
4. The ensemble characteristics of the spray from nozzle HA" of the mid-
pressure injector showed a high level of atomization. This phenomenon
is clearly demonstrated in the high-speed photographs, and the droplet
size measurements which yielded on the average SMD values of under
8 jim.
5. The penetration distance and droplet size measurements from the three
nozzles ("A", "B" and "C") of the mid-pressure injector were
consistent with the visual observations. Spray features were
significantly influenced by the nozzle design.
6. In an engine the spray from a centrally located mid-pressure injector
diffuses symmetrically and shows uniform distribution of the fuel.
Accordingly, combustion chamber with relatively strong squish can be
used to promote fuel-air mixing and flame propagation.
Finally, progress has been made in demonstrating the effects of
various nozzle configurations on the spray patterns and droplet size
distribution. In addition, the high-speed images have contributed greatly
to the physical understanding of the transient spray characteristics.
VI. Recommendations and Future Efforts
To avoid the spray convergence and its adverse effect on the
atomization in the low-pressure injection system, the following design
modifications should be made: (a) replace the vertical slots with swirling
vanes on the valve stem, (b) use tapered nozzle tip, and (c) restrict the
diametral clearance between the valve and the nozzle body.
In the mid-pressure injection system fine atomization was achieved
from some nozzle designs; however, the bouncing in the other injectors
influenced significantly the spray characteristics. In this case, modifying
the accumulation volume or the spring in the injectors is required.
The continuing efforts of the Technology Development Group of the
Office of Regulatory Programs and Technology at the EPA/NVFEL lab
will focus on fuel spray characterization and combustion analysis in a
direct injection engine. New injector concepts and technology will be
tested using both high-speed imaging and laser diffraction techniques.
Results will be incorporated into automotive engine developments to
promote clean combustion and reduced emissions for future generation
engines.
-------
-23-
VII. Acknowledgments
The authors would like to express their appreciation for the efforts
and support of L. Kocher and W. Willingham of the EPA/NVFEL for
machining the cylinder head, J. Martin for his technical assistance, and
the efforts of J. Criss and L. Johnson for their editing support. Particular
recognition is given to P. Dingle of Lucas Automotive for providing the
poppet valve injector to modify and for his helpful comments. The authors
would like also to thank AMBAC International for fabricating the single-
hole nozzles and for their assistance during the design process.
VIII. References
1. "Injection Characteristics of High Pressure Accumulator Type
Injector," Savey, C.W., Beck, J.N., Dobovisek, Z., and Gebert, K.,
SAE Paper 890266, 1989.
2. "Spray Characteristics and Combustion Improvement of D.I. Diesel
Engine with High Pressure Fuel Injection," Kato, T., Tsujimura, K.,
Shintani, M., Minami, T., and Yamaguchi, I., SAE Paper 890265,
1989.
3. "Characteristics of a Diesel Impinging on a Flat Wall," Katsura, N.,
Saito, M., Senda, J., and Fujimoto, H., SAE Paper 890264, 1989.
4. "Stratified Charge Rotary Engine Studies at The MSU Engine
Research Laboratory," SAE Paper 890331, Hamady, F.J., Kosterman,
J., Chouinard, E., Schock, H.J., Chun, K., and Hicks Y., SAE
Transactions, 1989.
5. "Fuel-Air Mixing Visualization in a Motored Rotary Engine
Assembly," Morita, T.B., Hamady, F.J., Stuecken, T.R., Somerton,
C.W. and Schock, H.J., SAE Paper 910704, 1991.
6. "Structures of Fuel Sprays in Diesel Engines," Hiroyasu, H., and
Arai, M., SAE Paper 900475, 1990.
7. "Optimizing Spray Behavior to improve Engine Performance and to
Reduce Exhaust Emissions in a Small D.I. Diesel Engine," Yoshikawa,
S., Furusawa, R., Arai, M., and Hiroyasu, H., SAE Paper 890463,
1989.
8. "Fuel Injection and Combustion Phenomena in a High Speed D.I.
Diesel Engine Observed by Means of Endoscopic High Speed
-------
-24-
Photography," Welberger, P., and Cartellieri, W., SAE Paper 870097,
1987.
9. "Experimental and Numerical Analysis of a Diesel Spray," Allocca,
L., Belardini, P., Bertoli, C., Corcione, F.E., and De Angelis, F.,
SAE Paper 920576, 1992.
10. "Diesel Spray Structure Investigation by Laser Diffraction and Sheet
Illumination," Gulder, O.L., Smallwood, G.J., and Snelling D.R., SAE
Paper 920577, 1992.
11. "Characterization of the Transient Spray from a High Pressure Swirl
Injector," Evers, L.W., SAE Paper 940188, 1994.
12. "Droplet Sizes and Velocities in a Transient Diesel Fuel Spray," Koo,
J.Y., and Martin, J.K, SAE Paper 900397, 1990
13. "Study on Atomization and Fuel Drop Size Distribution in Direct
Injection Diesel Spray," Quoc, H.X., and Brun M., SAE Paper
940191, 1994.
14. "Spray and Self-Ignition Visualization in a D.I. Diesel Engine,"
Baritaud, T.A., Heinze, T.A., and LeCoz, J.F., SAE Paper 940681,
1994.
15. "Liquid and Vapor Fuel Distributions from an Air-Assist Injector - An
Experimental and Computational Study," Diwakar, R., Fansler, T.D.,
French, D.T., Ghandhi, J.B., Dasch, C.J., and Heffelfinger, D.M.,
SAE Paper 920422, 1992.
16. "Quantitative 2-D Fuel Vapor Concentration Imaging in a Firing D.I.
Diesel Engine Using Planar Laser-Induced Rayleigh Scattering,"
Espey, C., Dec, I.E., Litzinger, T.A., and Santavicca, D.A., SAE
Paper 940682, 1994.
17. "Preliminary Investigation of Solenoid Activated In-Cylinder Injection
in Stoichiometric S.I. Engines," Lake, T.H., Christie, M.J., and
Stokes, J., SAE Paper 940483, 1994.
18. "Change of Calibration of Diffraction-Based Particle Sizers in Dense
Sprays", Dodge, L.G., Optical Engineering, Vol. 23, No. 25, pp. 626-
630, October 1984.
-------
-25-
19. "Multiple Scattering Effect in Laser Diffraction Measurements of
Dense Sprays with Bimodal Size Distribution", Gulder, O.L., ILASS
Conference, Madison, Wisconsin, June 8-11, 1987.
20. Atomization and Sprays. Lefebvre, A.H., pp. 397-402, Taylor &
Francis, 1989.
------- |