Professional Documents
Culture Documents
December 2004
School of Engineering
University of Brighton
Abstract
ABSTRACT
The report presented here describes the work undertaken at the University of Brighton
on a rapid compression machine based on a two-stroke cross flow Diesel engine
(Proteus), with an optical head (spray chamber) to allow visualisation of the incylinder combustion events. These experiments include the study of dense Diesel fuel
spray structure such as spray cone angle, breakup length/time, and the spray
penetration length from single and multi-hole valve covered orifice (VCO) nozzles at
cold and hot intake charge conditions. The characteristic dependence of penetration
length on time after the breakup period, the effect of multi-hole nozzles on spray
penetration, and the effect of injection pressures as well as the in-cylinder pressures
were investigated. From the experimental data, a new empirical constant for the
penetration length correlation was derived.
The study of a single jet spray from a multi-hole nozzle was achieved by designing
and testing a novel spray isolator. This allowed more precise characterisation of the
isolated spray without the influence of other jets from a multi-hole nozzle. These
phenomena were investigated in this study, using the direct imaging technique of sidelit high-speed video photography.
The analysis and measurement of the injected fuel mass for each consecutive injection
for multiple-injection strategies was carried out using the rate gauge technique.
Further experiments are planned to characterise the transient spray under multipleinjection strategies at high in-cylinder and injection pressures. The forthcoming
experiments will include using Phase Doppler Anemometry (PDA) and high-speed
video imaging technique. This will provide measurements of droplet size, velocity,
and number densities (PDA) and spray geometry measures such as spray cone angle,
breakup length, wall impingement and penetration depth (high-speed video imaging).
ii
Declaration
DECLARATION
I hereby certify that this report is my own work except where otherwise indicated. I
have identified my source of information, and in particular have put in quotation
marks any passages that have been quoted word for word and identified their origins.
Signed:
Date:
iii
Acknowledgments
ACKNOWLEDGMENTS
I am most grateful to my supervisory team Dr Cyril Crua, Dr Elena Sazhina, and Prof.
Morgan Heikal from the University of Brighton.
Special thanks are due to Dr Martin Gold from Ricardo Consulting Engineers.
I wish to express my gratitude to Prof. Sergei Sazhin, Dr Steve Begg, Dr Dave
Kennaird, Mr Romain Demory, and Mr Walid Abdelghaffar. Also, I wish to thank Mr
Ralph Wood, Mr Bill Whitney, Mr Ken Maris, Mr Dave Stansbury, and all the other
technical support groups who have contributed to this work in many ways.
In addition, I wish to acknowledge the financial and technical support provided by
Ricardo Consulting Engineering group as well as the funding received from EPSRC,
and the equipment obtained through their loan pool scheme.
iv
Contents
CONTENTS
ABSTRACT ______________________________________________________________________ ii
DECLARATION __________________________________________________________________iii
ACKNOWLEDGMENTS ___________________________________________________________ iv
LIST OF TABLES________________________________________________________________ vii
LIST OF FIGURES ______________________________________________________________ viii
NOMENCLATURE_______________________________________________________________ xii
1.
INTRODUCTION _______________________________________________ 1
2.2
2.3
2.3.1
2.3.2
2.3.3
2.4
3.
Introduction_____________________________________________________ 14
3.2
3.3
Experimental results______________________________________________ 17
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
3.3.6
3.3.7
3.3.8
3.4
4.
5.
Introduction __________________________________________________________ 7
Experimental method ___________________________________________________ 7
Associated errors _____________________________________________________ 13
17
19
21
26
30
33
35
42
Introduction_____________________________________________________ 47
4.2
4.3
Introduction_____________________________________________________ 58
Calibration of mass flow rate with multiple injection strategy___________________
High speed video imaging ______________________________________________
Investigation of multiple injections by means of PDA_________________________
Modelling ___________________________________________________________
59
60
60
61
References _________________________________________________________ 62
Appendices ________________________________________________________ 67
The experimental conditions for determination of compression ratio and Polytropic
coefficient ____________________________________________________________ A-1
The effect of multi-hole nozzle on penetration ______________________________ A-2
Contents
Breakup length & time as a function of injection and in-cylinder pressures _____ A-22
vi
List of tables
LIST OF TABLES
Table 2-1. Summary of errors at motored condition (Crua 2002)............................... 13
vii
List of figures
LIST OF FIGURES
Figure 2-1. Spray isolator. ............................................................................................. 3
Figure 2-2. Position of the spray isolator in the top hat. ............................................... 4
Figure 2-3. CAD drawing of the jets from a 7 hole nozzle, spray development within
the isolator cone and the chamber in the top hat. .................................................. 4
Figure 2-4. Top hat with two different types of injector ............................................... 5
Figure 2-5. Quiescent fuel spray chamber under ambient conditions. .......................... 6
Figure 2-6. Pressure decay against time. ....................................................................... 8
Figure 2-7. Mass flow rate against pressure divided by the square root of temperature.
............................................................................................................................. 10
Figure 2-8. Log of pressure data against log of volume for 8 MPa ICP (cold intake).11
Figure 2-9. PV diagram for measured and computed data at 8 MPa ICP (cold intake).
............................................................................................................................. 11
Figure 2-10. In-cylinder temperature and mass against crank angle at 8 MPa ICP (cold
intake). ................................................................................................................. 12
Figure 3-1. Schematic diagram of the experimental set-up for side-lit spray
photography. ........................................................................................................ 15
Figure 3-2. An image in digital and binary format showing maximum tip penetration
before cluster detachment. The red dots indicate the position of the threshold;
acquired with the Phantom camera...................................................................... 16
Figure 3-3. Graphical representation of threshold level at 10% (chosen level of grey
at 60). ................................................................................................................... 17
Figure 3-4. Spray cone angle at the start and end of injection; acquired with the APX
camera.................................................................................................................. 18
Figure 3-5. Validation of the spray isolator, at different pressures; acquired with the
APX camera......................................................................................................... 18
Figure 3-6. Transverse movement of the needle caused by differential pressure
distribution for a multi hole VCO nozzle. ........................................................... 19
Figure 3-7. A 3 and a 5 hole nozzle, with an injection pressure of 60 MPa, 50 mm3
fuelling; acquired with the APX camera. ............................................................ 20
Figure 3-8. Randomness in hole to hole variation from a 3-hole VCO single guided
0.2 mm Bosch nozzle at 60 MPa injection pressure, 2 MPa ICP, and 50 mm3
fuelling. The images are captured 0.17 ms after first sight of fuel, acquired with
the Phantom camera. ........................................................................................... 20
Figure 3-9. Graphical representation of the injection time delay with injection
pressure. In-cylinder pressure 6 MPa at TDC. .................................................... 21
Figure 3-10. Time delay at different in-cylinder pressures, injection pressure at
160MPa................................................................................................................ 22
Figure 3-11. Comparison between 0.2 mm single-hole, 3-hole, and 5-hole single
guided VCO nozzles, for an injection pressure of 140 MPa, and 30 mm3 fuelling.
Times are in milliseconds after first frame showing visible liquid spray; acquired
with the APX camera........................................................................................... 23
Figure 3-12. Needle lift trace for a Bosch injector, 0.2 mm VCO single hole single
guided nozzle at 160 MPa injection pressure, 30mm3 fuelling. .......................... 23
Figure 3-13. Penetration with time for single/multi hole 0.2 mm single guided VCO,
and 7-hole Delphi injector, at 140 MPa injection pressure in an atmospheric
pressure and temperature chamber. ..................................................................... 24
viii
List of figures
Figure 3-14. Effect of gas density and viscosity on spray tip penetration; 140 MPa
injection pressure; 30 mm3 fuelling; 0.2 mm single guided 5 hole VCO nozzle in
the atmospheric pressure and temperature chamber............................................ 25
Figure 3-15. Penetration with time for single/multi hole 0.2 mm single guided VCO,
and 7-hole Delphi injector, at 160 MPa injection pressure, and 6 MPa in-cylinder
pressure at TDC. .................................................................................................. 25
Figure 3-17. Fish-bone shaped structure of the spray from a 0.2 mm single hole VCO
single guided nozzle, at 2 MPa ICP, cold intake. Images acquired with the
Phantom camera. ................................................................................................. 27
Figure 3-18. Fish-bone shaped structure of the spray, using the Laplacian edge
detection method of the Phantom camera. .......................................................... 28
Figure 3-19 a. Spray images taken at 2 MPa In-cylinder pressure, and 0.41 ms after
first sight of fuel. Images acquired with the Phantom camera. ........................... 29
Figure 3-20. High-speed video sequence in steps of 29 s showing clusters of droplet
starting to detach along the leading edge of the spray at cold intake; acquired
with the Phantom camera. ................................................................................... 30
Figure 3-21. Spray penetration from a 0.2 mm VCO nozzle, at 160 MPa injection
pressure; 6 MPa ICP. ........................................................................................... 31
Figure 3-22. Rail pressure traces recorded by an AVL Indiset high speed data
acquisition system at 160 MPa injection pressure, 6 MPa ICP, and 50 mm3
fuelling. Broken lines indicate apparent needle full lift / closure. The nozzles
tested were all 0.2 mm single guided VCO. ........................................................ 31
Figure 3-23. Injection current traces recorded by an AVL Indiset high speed data
acquisition system at 160 MPa injection pressure, 50 mm3 fuelling................... 32
Figure 3-24. Rail pressure traces recorded at 60 MPa injection pressure, 6 MPa ICP,
and 50 mm3 fuelling. ........................................................................................... 32
Figure 3-25. Frequency of the Common rail pump and the regulator recorded by an
AVL Indiset high speed data acquisition system. ............................................... 33
Figure 3-26. Effect of injection pressure on spray penetration for a 0.2 mm 3-hole
single guided VCO nozzle, at 6 MPa ICP at cold intake..................................... 34
Figure 3-27. Effect of injection pressure on spray penetration for a 0.2 mm 5-hole
single guided VCO nozzle, at 6 MPa ICP at cold intake..................................... 34
Figure 3-28. Effect of injection pressure on spray penetration for a 0.2 mm 3 and 5hole single guided VCO nozzle, at 6 MPa ICP at cold intake............................. 35
Figure 3-29. Effect of ambient gas density on liquid spray penetration for cold intake;
160 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3
fuelling................................................................................................................. 36
Figure 3-30. Effect of ambient gas density on liquid spray penetration for cold intake;
100 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3
fuelling................................................................................................................. 36
Figure 3-31. Effect of ambient gas density on liquid spray penetration for cold intake;
60 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3
fuelling................................................................................................................. 36
Figure 3-32. Comparison between correlation proposed by Hiroyasu and Arai [1989,
1990] and experimental results for non-evaporating spray (cold intake); 160 MPa
injection pressure; 6 MPa ICP; 36.3 kg m-3 ambient gas density; 0.2 mm single
guided 3 hole VCO nozzle. ................................................................................. 37
Figure 3-33. Comparison between correlation proposed by Hiroyasu and Arai [1989,
1990] and experimental results for non-evaporating spray (cold intake); 60 MPa
ix
List of figures
injection pressure; 6 MPa ICP; 36.3 kg m-3 ambient gas density; 0.2 mm single
guided 3 hole VCO nozzle. ................................................................................. 38
Figure 3-34. Comparison between correlation proposed by Yule and Filipovic [1991]
and experimental results for non-evaporating spray (cold intake); 160 MPa
injection pressure; 6 MPa ICP; 36.3 kg m-3 ambient gas density; 0.2 mm single
guided 3 hole VCO nozzle. ................................................................................. 39
Figure 3-35. Comparison between correlation proposed by Yule and Filipovic [1991]
and experimental results for non-evaporating spray (cold intake); 60 MPa
injection pressure; 6 MPa ICP; 36.3 kg m-3 ambient gas density; 0.2 mm single
guided 3 hole VCO nozzle. ................................................................................. 39
Figure 3-36. The effect of in-cylinder gas density on breakup time for nonevaporating spray (cold intake); 0.2 mm single guided 3 hole VCO nozzle....... 40
Figure 3-37. The effect of in-cylinder gas density on breakup length for nonevaporating spray (cold intake); 0.2 mm single guided 3 hole VCO nozzle....... 40
Figure 3-38. Breakup time as a function of injection pressure for varied in-cylinder
gas density; non-evaporating spray (cold intake); 0.2 mm single guided 3 hole
VCO nozzle. ........................................................................................................ 41
Figure 3-39. Breakup length as a function of injection pressure for varied in-cylinder
gas density; non-evaporating spray (cold intake); 0.2 mm single guided 3 hole
VCO nozzle. ........................................................................................................ 42
Figure 3-40. Comparison between non-evaporating and evaporating spray at 160 MPa
injection pressure; 6 MPa ICP; 50 mm3 fuelling; 0.2 mm single guided 3 hole
VCO nozzle; 0.6 ms after first sight of fuel; acquired with the Phantom camera.
............................................................................................................................. 43
Figure 3-41. Effect of ambient gas density on liquid spray penetration for hot intake at
160 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle and 50
mm3 fuelling. ....................................................................................................... 43
Figure 3-42. Effect of ambient gas density on liquid spray penetration for hot intake at
60 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle and 50
mm3 fuelling. ....................................................................................................... 44
Figure 3-43. Effect of ambient gas temperature on liquid spray penetration length at
160 MPa injection pressure; 0.2 mm single guided 3 hole VCO nozzle and 50
mm3 fuelling. ....................................................................................................... 44
Figure 4-1 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO
Delphi injectors; 60 MPa injection pressure; 0.41 ms injection duration. .......... 48
Figure 4-2 Comparison of injector No 1 and No 2; 60 MPa injection pressure; 0.41 ms
injection duration for injector No1, and 0.46 ms injection duration for No2...... 49
Figure 4-3 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO
Delphi injectors; 60 MPa injection pressure; 1.25 ms injection duration. .......... 50
Figure 4-4 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO
Delphi injectors with multiple injection strategy; 0.6 ms dwell time; 160 MPa
injection pressure; 0.4 ms injection duration....................................................... 51
Figure 4-5 Comparison of injection rate with the filtered rate in mg/ms. 100 MPa
injection pressure; 0.38 ms injection duration corresponding to 5 mm3 volume of
fuel; Delphi injector No 2.................................................................................... 52
Figure 4-6 Differences in the rate of injection between single injection event (solid
lines) and multiple-injection (broken lines). The data for the single injection has
been superimposed on the above Figure twice, with an offset to match the second
List of figures
xi
Nomenclature
NOMENCLATURE
Latin symbols
A------------------------------------------------------------------------------------------Area [m2]
AP.Lp/2----------------------------------------------------------------Projected area of spray [m2]
C-------------------------------------------------------------------------------------------Constant
Cc----------------------------------------------------------------------Coefficient of contraction
Cd------------------------------------------------------------------------Coefficient of discharge
dd----------------------------------------------------------------------------Droplet diameter [m]
d1--------------------------------------------------------------------Mean size based on number
d2---------------------------------------------------------------Mean size based on surface area
d3--------------------------------------------------------------------Mean size based on volume
d10--------------------------------------------------------------------Arithmetic mean diameter
d32--------------------------------------------------------------------------Sauter mean diameter
dG-----------------------------------------------------------------Gaussian beam diameter [m]
De--------------------------------------------------------------Equivalent nozzle diameter [m]
Din----------------------------------------------------------Nozzle orifice diameter (inlet) [m]
Dn-------------------------------------------------------------------Nozzle orifice diameter [m]
Do-------------------------------------------------------Sack chamber diameter of nozzle [m]
Ex-------------------------------------------------------------------------Beam expansion factor
FD------------------------------------------------------------------------------Doppler frequency
FT---------------------------------------------------------Focal length of the transmitting lens
Ln-------------------------------------------------------------------------------Nozzle length [m]
Lp--------------------------------------------------------------------------Penetration length [m]
Lb-----------------------------------------------------------------------------Breakup length [m]
Ls---------------------------------------------------------------------------Spacing of the stripes
m---------------------------------------------------------------------------Mass flow rate [kgs-1]
nrel--------------------------------------------------Ratio of the refractive index of Diesel/air
Nf------------------------------------------------------------------------------Number of fringes
P1--------------------------------------------------------------------------Injection pressure [Pa]
Pg--------------------------------------------------------------------------Ambient pressure [Pa]
Pv----------------------------------------------------------------------------vapour pressure [Pa]
rr-------------------------------------------------------------------Radius of nozzle (round) [m]
xii
Nomenclature
S----------------------------------------------------------------------------------------------Signal
t----------------------------------------------------------------------------------------------Time [s]
tb-----------------------------------------------------------------------------------Breakup time [s]
ti----------------------------------------------------------------------------------Time injection [s]
t*----------------------------------------------------------------------------Shear breakup time [s]
T-----------------------------------------------------------------------------------Temperature [K]
TDP-----------------------------------------------------------------------------------Doppler period
Tphase-------------------Period between the zero crossing of the signal from detector 1 to 2
ub--------------------------------------------------------------------------Breakup velocity [ms-1]
ud---------------------------------------------------------------------------Droplet velocity [ms-1]
ur---------------------------------------------------------------------------Relative velocity [ms-1]
ui-----------------------------------------------------------------------Initial spray velocity [ms-1]
U----------------------------------------------Velocity component normal to the fringe [ms-1]
V----------------------------------------------Mass average velocity through the nozzle [ms-1]
Greek symbols
-------------------------------------------------------------------------Intersection angle [deg]
z--------------------------------------------------------------Length of the probe volume [m]
x--------------------------------------------------------------Height of the probe volume [m]
y---------------------------------------------------------------Width of the probe volume [m]
f------------------------------------------------------------------------------Fringe spacing [m]
---------------------------------------------------------------------------Scattering angle [deg]
---------------------------------------------------------------------------Elevation angle [deg]
--------------------------------------------------------------------------Spray cone angle [deg]
----------------------------------------------------------------------------------Density [kg m-3]
--------------------------------------------------------------------------Surface tension [Nm-1]
------------------------------------------------------------------------------------Viscosity [Pa]
---------------------------------------------------------------------Kinematic viscosity [m2s-1]
-----------------------------------------------------------Wavelength of the laser beam [nm]
P----------------------------------------------------------------------------Pressure differences
t-------------------------------------------------------------------------------Difference in time
xiii
Nomenclature
Subscripts
g---------------------------------------------------------------------------------------Ambient gas
l-----------------------------------------------------------------------------------------------Liquid
Acronyms
ATDC----------------------------------------------------------------------After top dead centre
BDC--------------------------------------------------------------------------Bottom dead centre
BTDC--------------------------------------------------------------------Before top dead centre
CA-------------------------------------------------------------------------------------Crank angle
CCD----------------------------------------------------------------------Charge coupled device
CFD-------------------------------------------------------------Computational fluid dynamics
CO------------------------------------------------------------------------------Carbon monoxide
CN-----------------------------------------------------------------------------Cavitation number
DI----------------------------------------------------------------------------------Direct injection
ICP---------------------------------------------------------------------------In-cylinder pressure
LDA---------------------------------------------------------------Laser Doppler Anemometry
LIF------------------------------------------------------------------Laser induced fluorescence
LSD----------------------------------------------------------------------Laser Sheet Dropsizing
MIE--------------------------------------------------------------------------------Mie Scattering
NOX--------------------------------------------------------------------------------Nitrogen oxide
Oh-----------------------------------------------------------------------------Ohnesorge number
PIV------------------------------------------------------------------Particle image Velocimetry
PMT----------------------------------------------------------------------Photo Multiplier Tube
PM-------------------------------------------------------------------------------Photo Multiplier
PDA----------------------------------------------------------------Phase Doppler anemometry
PTV--------------------------------------------------------------Particle tracking Velocimetry
VCO-----------------------------------------------------------------------Valve covered orifice
Re-------------------------------------------------------------------------------Reynolds number
SMD-----------------------------------------------------------------------Sauter mean diameter
TDC------------------------------------------------------------------------------Top dead centre
TTL-------------------------------------------------------------------Transistor transistor Logic
We---------------------------------------------------------------------------------Weber number
xiv
1.
INTRODUCTION
Diesel engines are widely used because of their high efficiency and cost effectiveness.
Recently, passenger cars have adopted such engines, due to their inherent advantages
over gasoline counterparts. This has increased the demand for higher output, lower
noise and greater emission performance from Diesel engines.
In meeting these
In view of the above criteria, the experiments undertaken so far are mainly focused on
single injection events in order to assess the effects of ambient gas density on
The study of a single jet spray from a multi-hole nozzle was achieved by designing
and testing a novel spray isolator. This allowed more precise characterisation of the
isolated spray without the influence of other jets from a multi-hole nozzle.
The information gathered so far on single and multi-hole single injection events are
crucial for correct analysis of the combustion cycle. This information was obtained
using a side-lit high-speed CCD video imaging technique, to record complete
injection cycles in an atmospheric pressure and temperature controlled environment.
This was in addition to the high pressure and temperature closed loop system
investigation (Proteus).
Further experiments are planned to be carried out during the forthcoming months, on
the calibration of fuel mass flow, and the rate of injection. High-speed CCD video
imaging, and finally Phase Doppler Anemometry (PDA), with a multiple injection
strategies, will be applied in order to understand the characteristics of Diesel spray
formation and the subsequent effect on emissions. This should help to build a new
phenomenological model of dense Diesel spray formation for a multiple injection
strategy.
2.
EXPERIMENTAL APPARATUS AND EVALUATION OF
CURRENT BUILD
2.1
Initially the top hat (Cylinder head) was designed for experimental investigation of
single hole Bosch injectors as described in the literature review. The current study has
redesigned the top hat to incorporate a spray isolator (Figure 2-1). This will allow the
study of a single spray jet from a multi-hole nozzle, without disturbing the flow
through the other holes. In addition, precise characterisation of the isolated spray can
be achieved without the influence of other jets from a multi-hole nozzle.
Isolator legs
Isolator cone
Isolator outlet
With the isolator positioned in the head, the injector resides between the two legs that
secure the isolator in place (Figure 2-2).
The particular nozzle hole of interest is then aligned with the central outlet of the
spray isolator, thus allowing the spray jet of interest to exit. The remaining fuel from
the other spray jets is captured in the base of the isolator cone. The captured fuel then
exits down the excess fuel disposal outlet, behind the isolator deflector plate, and is
extracted through the exhaust system as shown in Figure 2-2 and 2-3. The spray
isolator is only capable of restricting up to six holes in a seven hole spray nozzle. This
is due to the decreased separation angle between each spray jet, if the number of
nozzle holes exceeds seven.
Top hat
Depending on the number of holes in the multi hole nozzle (between 2 to 7 in this
study), some of the spray jets are formed in a small chamber which has been
machined out in the modified top hat. The remaining spray jets (maximum of 2 jets if
a 7 hole nozzle is experimented with) are formed within the isolator cone (Figure 2-3).
Isolator leg
Top hat
Injector nozzle
Isolator excess fuel disposal outlet
Isolator outlet
Isolator cone
Spray jet
Figure 2-3. CAD drawing of the jets from a 7 hole nozzle, spray development within the isolator cone
and the chamber in the top hat.
The volume of the chamber that was machined out in the top hat is identical to the
volume of the spray isolator that fits into the spray chamber. This was to maintain the
compression ratio of the original build.
The redesign of the top hat also permits two different types of injectors (Bosch &
Delphi) to be studied and compared as shown in Figure (2-4).
Top hat
Bosch injector
Delphi injector
2.2
An atmospheric pressure and temperature chamber was designed and constructed for
the investigation of the fuel spray characterisation and validation of the spray isolator.
The spray chamber is shown in Figure 2-5. The chamber is composed of 4 glass
windows held together by aluminium angled beams. Two 10 mm thick aluminium
plates were machined to form the basis of the chamber frame. The top plate held the
weight of top hat, while the bottom plate was mounted on the fuel bowl. Grooves in
the top and bottom plates ensured that the glass chamber was firmly secured in place.
A cork gasket was utilised between the grooves in the plates and the glass windows to
ensure that fuel vapour did not leak from the chamber. The two plates were separated
by four tubular steel posts. The tubes also acted as part of the clamping device for the
chamber to be held together by the use of four threaded studs that ran the height of the
rig. The fuel bowl had several functions; the first was to allow liquid fuel to drain
from its centre, and the second was to house the extraction pipe for the low-speed
purge system which assisted the removal of spent fuel vapour from the chamber,
without altering the injected fuel spray formation. Additionally, incorporated within
the fuel bowl was a very fine wire mesh. This allowed the breakup of large droplets
on impact at the base, and consequently, a reduction in window fouling.
5
Top hat
Stud
Small jets
Top plate
Spray isolator
Tubular steel post
Air supply to small jets
Angled beam
Bottom plate
Fuel bowl
Liquid fuel outlet
Air inlet
Extraction outlet
Aluminium leg
Wooden support
Air from the small air jets was also directed towards the windows to assist a reduction
in window fouling.
The low-speed purge system and the air to the small jets was controlled by a throttle
valve air regulator. The extraction pipe was attached to a filtration system housed
within a tank, absorbing the fuel vapour.
The complete chamber was mounted upon a set of aluminium legs, and then attached
to a set of wooden supports which absorbed some of the vibrations caused by the high
and low pressure pumps. The entire rig was mounted on a portable fuel delivery
system, identical to the one described in the literature review section 3.2. The
dampening of the vibration allowed accurate imaging acquisition to be achieved.
2.3
2.3.1
Introduction
2.3.2
Experimental method
The volumetric capacity of the Proteus combustion chamber had been determined by
calculating the swept volume and the clearance volume using specified engine data.
For the compression stroke, it was necessary to take the start of compression from the
moment the inlet and exhaust ports were covered by the top compression ring at 243o
crank angle.
This experiment was conducted in several stages. The first stage was, by using the
combustion chamber as a fixed volume, the chamber was filled with nitrogen at 1
MPa at a fixed crank angle (the crankshaft was locked) corresponding to a known
chamber volume. The decay in pressure was measured in order to find out the mass
lost through blow-by past the piston (Figure 2-6).
The transducer used for this part of the experiment was a piezoelectric 4045-A10
Kistler transducer measuring absolute pressures up to 1 MPa, operating between
temperatures of 273.15 and 413.15 K, with a thermal sensitivity shift of 1%.
Pressure drop
1.2
1
0.8
0.6
0.4
0.2
0
-5
10
15
20
25
30
Time (s)
Assuming that the decay in pressure against time is exponential (Figure 2-6), then,
p = a ebt
(2.1)
Where p is the pressure, a and b are constants, and t is the time. Taking the natural
log of Equation 2.1,
ln ( p ) = ln e H ebt
ln( p ) = H + bt
(2.2)
(2.3)
(2.4)
since
R
p = mT
V
(2.5)
then
dp R dmT R mdT Tdm
=
=
+
dt V dt
dt
V dt
(2.6)
However, when the gas temperature (T) is constant, dT/dt = 0. By rearranging and
substituting Equation 2.4, the mass flow rate becomes:
m=
dm V dp V
bt
=
=
a b e
dt RT dt RT
(2.7)
In addition, due to the pressure ratios between the crankcase pressure and the incylinder pressure being much less than 0.5, the system is said to be choked. Therefore,
by using Lavals nozzle theorem, the maximum mass flow rate past the piston rings
occurs when the Mach number is unity assuming adiabatic conditions. Hence the
maximum flow rate can be written as:
m = t At ut
(2.8)
Where u t , At , t are the critical values of velocity of the sound, cross-sectional area,
and the density at the throat. Therefore:
ut = RTt
(2.9)
Substituting Equation 2.9 into 2.8, the mass flow rate becomes:
m = t At RTt
(2.10)
Also, the maximum flow rate can be expressed in terms of the reservoir conditions
o , po and To , where
1
2 ( 1)
t = o
( + 1)
(2.11)
and
2
Tt = To
( + 1)
(2.12)
giving
1
2 ( 1)
m = o
At
+ 1
2 R To
(2.13)
po
, the expression for
R To
2 ( +1) / ( 1)
R + 1
p
m = At o
To
(2.14)
and
2 ( +1) / ( 1)
k = At
R + 1
(2.15)
therefore,
m=k
po
To
(2.16)
From the mass flow rate data collected using Equation 2.7 under isothermal condition,
the results are then applied to Equation 2.16.
By plotting the new data, a best fit trendline gives a value for the slope (k) as shown
in Figure 2-7.
0.0008
0.0007
dm/dt (kg/s)
0.0006
0.0005
0.0004
0.0003
0.0002
0.0001
0.0000
0
10000
20000
30000
40000
50000
60000
P/Sqr(T)
Figure 2-7. Mass flow rate against pressure divided by the square root of temperature.
For the second stage of the experiment, Proteus was motored at 500 rpm. For varied
intake conditions (such as air temperature), the in-cylinder pressure was measured.
Taking a log of pressure and volume data (Figure 2-8), the polytropic coefficient and
the constant, d are calculated using the relationships below.
10
9.0
y = -1.3 X + 2.5413
8.0
7.0
5.0
4.0
Log (P)
6.0
3.0
2.0
1.0
-4.0
-3.5
-3.0
0.0
-2.0
-2.5
Log (V)
Figure 2-8. Log of pressure data against log of volume for 8 MPa ICP (cold intake).
pV n = d
(2.19)
p = dV n
(2.20)
(2.21)
Y = AO + A1 X
(2.22)
d = 10 AO
(2.23)
n = A1
(2.24)
Using the calculated values of n and d , a smoothest pressure curve data was
computed and plotted against the measured data in order to smooth out the pressure
fluctuations as shown in figure 2-9.
10
Computed
Pressure (MPa)
Measured
0
0
0.0004
0.0008
0.0012
0.0016
0.002
Volume (m^3)
Figure 2-9. PV diagram for measured and computed data at 8 MPa ICP (cold intake).
11
Finally, by using the computed pressure data in the second stage of the experiment,
and applying the value of k obtained in the first stage, as well as applying Equation
2.5 and 2.16, a series of mass, temperature and mass flow rate data are compiled using
iterative computation for the motored test. From the new set of data, the temperature
and in-cylinder mass from the start of compression for each crank angle up to TDC
are gathered (figure 2-10). In the case of hot air intakes, the engine was fired for
several minutes between each test to ensure that the working temperature of the build
was reached. For the motored tests, the pressure at the inlet manifold and in-cylinder
were measured using piezoelectric Kistler pressure transducers 4045-A10, and 6061
respectively. The operating temperature range of the water cooled Kistler 6061 is
between 223.15 and 623.15 K, and a working pressure range between 0 and 25 MPa,
with a sensitivity shift of 0.5% between 323.15 and 623.15 K. The 6061 pressure
transducer operates at gauge pressure, therefore absolute pressure levels were
established in further treatment of the pressure signals (the exact test conditions can
be found in Appendix A).
Temperature
480
0.01703
Mass
460
0.01702
440
0.01701
0.017
400
380
0.01699
360
0.01698
Mass (kg)
Temperature (K)
420
340
0.01697
320
0.01696
300
280
180
0.01695
210
240
270
300
330
360
Crank angle
Figure 2-10. In-cylinder temperature and mass against crank angle at 8 MPa ICP (cold intake).
From the correlations above for all test cases, an average polytropic coefficient of 1.3,
and a compression ratio of 9:1 was established.
12
2.3.3
Associated errors
Despite the information that could be obtained from the polytropic coefficient and
compression ratio, its determination could be subjected to errors such as, inaccurate
reference level for the pressure along the cycle. A summary of errors for motored
condition is given in Table 2-2.
Error magnitude
Assessment method
0.01 %
Measured
< 0.5 %
Measured
0.5 %
2.4
Conclusion of chapter 2
The ratio of specific heat capacities at constant pressure and volume is a very useful
variable in determining and studying the compression of a reciprocating internal
combustion engine. The effective value of polytropic coefficient can be estimated
from pressure and volume measurements as a function of crank angle (or time) during
a motored test. The associated measurement errors have a direct influence on the
estimation of compression ratio and the polytropic coefficient. From the data obtained
an average polytropic coefficient of 1.3 and a compression ratio of 9:1 were
established.
13
3.
3.1
Introduction
As explained in the literature review, information on spray tip penetration and spray
cone angle is crucial in the design process of an internal combustion engine. This
information not only aids to improve efficiency, but also assists the understanding of
parameters used to judge fuel spray performance.
For example, the penetration length must neither be too long nor short in a
combustion chamber. If too long, impingement could occur. This will result in wetting
the cylinder walls and/or piston crown. This would consequently lead to formation of
soot and wastage of fuel. If too short, reduction in mixing efficiency and good
combustion has been compromised.
The current work describes the experiments undertaken to quantify liquid fuel
distribution of Diesel fuel injected through a modern common rail fuel injection
system. These investigations were conducted using an atmospheric pressure and
temperature chamber, and the Proteus rig.
Characteristics of Diesel spray jet such as liquid core length, penetration rate, spray
cone angle, and droplets breakup can be assessed by direct imaging with the aid of a
camera (Hattori et al. [2002], Heimgartner and Leipertz [2000], Huh et al. [1991],
Karimi [1989], Kennaird et al. [2002]) . There are a number of direct imaging
strategies that can be applied. High-speed visualisation is the adopted technique in this
study.
3.2
A Photron Ultima APX Fastcam high-speed camera was used in this series of
experiments. The camera featured a 10-bit monochromatic CMOS sensor, and a
global electronic shutter speed up to 4 s. With a recording rate adjustable from 60 to
2000 frames per second at maximum resolution (1024 1024), and 4000 to 120000
frames per second at progressively reduced resolution.
Compromise between
acquisition rate and resolution was obtained with a frame rate of 20000, 30000, and
14
40000 frames per second, with a corresponding maximum resolution of 128 256,
256 128, and 512 64 respectively. With over exposure protection, optimum image
quality was ensured for every pixel regardless of illumination levels within the
recorded images (Photon 2004). In order to maximise intensity of the recorded
images, the gamma correction factor was set to 1.0.
In addition, a Phantom V 7.1 high-speed camera was also used in this series of
experiments. The camera featured a 8-bit monochromatic CMOS sensor, and a global
electronic shutter speed up to 2 s. With a recording rate adjustable up to 4800
frames per second at maximum resolution (800 600), and 4800 to 150000 frames
per second at progressively reduced resolutions. Compromise between acquisition rate
and resolution was obtained with a frame rate of 34300 frames per second, with a
corresponding maximum resolution of 128 320. In order to maximise intensity of
the recorded images, the gamma correction factor was set to 1.0. To ensure the best
possible image recording, a black reference level (background grey level reference)
was performed prior to capturing and recording of the images. This black reference
was set automatically by the camera at a level of 40 (out of 256 level of grey for an 8bit camera).
Two 125 W halogen flood lights fitted with a diffuser were used as shown in Figure
3-1, each focused on the fuel spray axis on opposite sides to each other (side lighting).
Flood light
FIE controller
Diffuser
High-speed video camera
Optical Chamber
Injector
PC
Figure 3-1. Schematic diagram of the experimental set-up for side-lit spray photography.
This setup gave the best performance for a homogeneous illumination of the spray
background (as oppose to backlighting).
15
The camera trigger was synchronised with the transistor transistor logic (TTL) output
of the FIE controller. The captured images were then downloaded to a computer
directly connected to the high-speed cameras. The fuel line pressure and camera
trigger were also recorded by an AVL Indiset high speed data acquisition system,
when the ambient pressure and temperature rig was in use. Under motored conditions,
the in-cylinder pressure, intake pressure, fuel line pressure, exhaust pressure, and
injection current was also recorded with the AVL Indiset high speed data acquisition
system.
The processing of the video images for measurement of the spray penetration was
preformed by purpose-developed software (Crua 2002). Suitable pixel thresholding
was carried out in order to pick out the tip of unbroken portion of the spray outline,
furthest from the nozzle on the spray axis ( 1 pixel) from the background. Figure 3-2
shows a typical image before and after thresholding.
The threshold level was subjectively chosen by selecting one image from the batch of
images generated by a test run, and varying the threshold level to obtain optimum
results. This threshold value became suitable for all images in that batch since the
quality and the illumination of the spray images remained constant.
Figure 3-2. An image in digital and binary format showing maximum tip penetration before cluster
detachment. The red dots indicate the position of the threshold; acquired with the Phantom camera.
Due to the inclination of the nozzle holes, the central axis of the spray trajectory was
towards the camera, and not perpendicular to the focal lens (Figure 3-1). A
trigonometric correction factor was implemented in order to establish the actual
penetration length.
16
A number of repeatability tests for all test conditions were conducted in order to
assess variability in penetration. The variation in penetration was found to be 4.6%
for a typical set of tests at the same condition. This variation only applies from t = 0 to
the time it takes to reach the tip of unbroken portion of the spray outline, furthest from
the nozzle on the spray axis. The resolution selected with the Phantom camera (128
320), gave a pixel size of 0.15 0.15 mm per pixel. This gave an uncertainty value
of 0.075 mm. With a variation of 10% for the threshold level chosen, the change
in measured penetration length was 2% (Figure 3-3).
45
Threshold at 66-ICP at 6 MPa-Injection at 160 MPa
40
Threshold at 60-ICP at 6 MPa-Injection at 160 MPa
35
30
25
20
15
10
5
0
0
0.2
Time (ms)
0.4
0.6
Figure 3-3. Graphical representation of threshold level at 10% (chosen level of grey at 60).
3.3
Experimental results
3.3.1
In order to substantiate the use of a spray isolator, high-speed video tests were
performed with the setup described in Chapter 2, and Section 3.2.
A 5 hole Bosch single guided VCO nozzle, with hole diameter 0.2 mm was used. The
injection pressures were set at 60, 100, 140 MPa. The volume of fuel injected varied
between 30 and 50 mm3 for all injection pressures.
Prior to alignment of the nozzle hole selected with the central outlet of the spray
isolator, high-speed video images of the complete nozzle were taken. This allowed
selection of a spray jet with the largest spray cone angle at the beginning and the end
of spray duration (Figure 3-4).
17
1
2
2
3
3
4
Figure 3-4. Spray cone angle at the start and end of injection; acquired with the APX camera.
From images shown in Figure 3-4, it is clear that the widest spray cone angle is
produced via hole number 4 (the variation in the spray cone angle and the penetration
length will be covered in the forthcoming section). Therefore, hole number 4 was
selected as the worst case scenario in order to reduce ambiguity in the performance of
the isolator. Figure 3-5 shows spray 4 out of the spray isolator at 30mm3 fuelling.
60 MPa injection pressure
Figure 3-5. Validation of the spray isolator, at different pressures; acquired with the APX camera.
From the high speed video images taken, the spray isolator exhibits good isolation of
the spray that is being studied. This is more evident at high injection pressures, when
sudden interruption of the spray at the end of injection from needle closure causes a
large spray cone to be developed as shown in Figure 3-5.
This device has enormous potential when PDA technique is applied. With Diesel
spray being highly dense, the use of multi-hole nozzle in a combustion chamber, in
conjunction with PDA technique can be counterproductive, if utilisation of an
18
isolating device is not applied. This is due to the increased possibility of multiple
Diesel droplet occupancy in the probe volume of the incident laser beam transmitted
from the receiving optics as explained in the literature review.
3.3.2
For the single hole nozzles tested, sudden spray interruption was observed during the
early phase of injection (Section 3.3.3). This was not the case for multi hole nozzles
tested. However, transverse movement of the needle still occurred at the very early
stage of injection. This transverse movement appeared in the form of
variation in the
spray cone angle and the penetration length from each individual hole from the multi
hole nozzle regardless of injection or in-cylinder pressures as shown in Figure 3-4.
This hole to hole variation was also observed by Bae and Kang [2000].
Vertical movement
Transverse movement
19
This continues until the pressure around the needle becomes more stable with the
vertical movement of the needle. Once the needle uncovers all the holes, the
penetration rate and the spray cone angle of each nozzle hole becomes almost
identical (Figure 3-7). Nevertheless, transversal as well as vertical oscillations of the
needle still continue while the injector is energized.
3-hole nozzle
5-hole nozzle
Figure 3-7. A 3 and a 5 hole nozzle, with an injection pressure of 60 MPa, 50 mm3 fuelling; acquired
with the APX camera.
For the 3 and the 5 hole nozzle tested, the hole to hole variation was always
symmetrical and occurred from the same nozzle hole regardless of the injection
pressure. However, this variation did not have the same magnitude for every cycle.
The images shown in Figure 3-8 are taken at three different cycles with an injection
pressure of 60 MPa, and in-cylinder pressure of 2 MPa for a 3-hole nozzle.
Cycle 1
Cycle 2
Cycle 3
Figure 3-8. Randomness in hole to hole variation from a 3-hole VCO single guided 0.2 mm Bosch
nozzle at 60 MPa injection pressure, 2 MPa ICP, and 50 mm3 fuelling. The images are captured 0.17
ms after first sight of fuel, acquired with the Phantom camera.
20
As can be seen from Figure 3-8, during the first cycle the variation in penetration
length and the spray cone angle from hole number 3 is quite profound. In the third
cycle however, there is little difference in the spray cone angle between the holes,
while the penetration length is slightly longer with the third nozzle hole. This
irregularity also suggests the variation in magnitude of transversal needle oscillation.
3.3.3
In comparing the data collected from a single hole, three-hole, and five-hole Bosch
single guided VCO 0.2 mm nozzles, a variation of time delay between the injection
signal and the first sight of fuel has been observed (Figure 3-9). The same phenomena
was also observed with the 7-hole Delphi injector, with an estimated hole size of
0.132 mm (Cd = 0.76).
This delay was observed using the high-speed video imaging in the ambient pressure
and temperature chamber, as well as the Proteus rig at elevated ICP.
Closer assessment of the data has shown, that the injection delay is independent of the
in-cylinder pressure (Figure 3-10), and that, it is directly linked to the injection
pressure. It is therefore acceptable to neglect the influence of in-cylinder pressure
when considering injection delay.
0.8
0.7
7-Hole Delphi
0.6
0.5
0.4
0.3
0.2
40
60
80
100
120
140
Injection Pressure (MPa)
160
180
Figure 3-9. Graphical representation of the injection time delay with injection pressure. In-cylinder
pressure 6 MPa at TDC.
21
As can be seen from Figure 3-9 and 3-10, the Delphi injector exhibits little or no
change in injection delay regardless of injection or in-cylinder pressure. This would
suggest, not only different types of injectors have different response time, but also a
varied level of dependency on the injection pressure that causes the time delay.
0.8
0.7
0.5
0.4
0.3
0.2
1
Figure 3-10. Time delay at different in-cylinder pressures, injection pressure at 160MPa.
Considering Figure 3-11, at about 0.15 ms after visibility of the liquid spray from the
0.2 mm single guided VCO single-hole nozzle, the flow seems to be interrupted. This
spray interruption appears to be caused by injector needle closure, long before the
actual end of the injection period. At about 0.075 ms after the initial interruption
(0.225 ms after SOI), the flow through the nozzle continues without any further delay.
This behaviour was also observed with the double guided VCO single-hole nozzle,
although to a lesser extent. With the multi-hole nozzles tested, no such pattern was
observed.
Previous studies of such injectors by Kennaird et al. [2002], Morgan et al. [2001], and
Crua [2002] have revealed the same phenomenon.
22
3-hole nozzle
5 hole nozzle
0.05 ms
0.075 ms
0.1
ms
0.125 ms
0.15 ms
0.175 ms
0.20 ms
0.225 ms
0.250 ms
0.275 ms
0.30 ms
Figure 3-11. Comparison between 0.2 mm single-hole, 3-hole, and 5-hole single guided VCO nozzles,
for an injection pressure of 140 MPa, and 30 mm3 fuelling. Times are in milliseconds after first frame
showing visible liquid spray; acquired with the APX camera.
Graphical representation of needle lift traces shows, the needle continues to lift during
this hesitation period (Figure 3-12). This would ultimately indicate that the needle
shuts the orifice by moving transversally while lifting vertically.
0
-0.5
0.5
1.5
2.5
3.5
Figure 3-12. Needle lift trace for a Bosch injector, 0.2 mm VCO single hole single guided nozzle at 160
MPa injection pressure, 30mm3 fuelling.
23
Figure 3-13 graphically represents the liquid spray tip penetration of the nozzles
tested in the atmospheric pressure and temperature chamber. It can be clearly
observed that the early hesitation in the opening phase of the single hole single guided
VCO nozzle, attributes to the delay in spray penetration with time. This behaviour
was repeatable for all the tests carried out, including tests in the rapid compression
machine (Figure 3-15).
The data for the multi hole nozzles in Figure 3-13 have been offset by 0.15 ms, in
order to allow direct comparisons, with the single hole nozzle.
140
Liquid tip penetration length (mm)
120
5-hole nozzle
7-hole delphi
100
80
60
40
20
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Time (ms)
Figure 3-13. Penetration with time for single/multi hole 0.2 mm single guided VCO, and 7-hole Delphi
injector, at 140 MPa injection pressure in an atmospheric pressure and temperature chamber.
As can be seen from figure 3-13 (in the atmospheric pressure and temperature
chamber) an average spray tip velocity of 220 m/s is reached linearly within 0.5 ms.
This semi linearity can be explained by the shape of the penetrating spray. Figure 3-14
shows the spray jet from a 5 hole 0.2 mm VCO nozzle with a pointed spray tip, and
what appears to be a high concentration of fuel mass fraction along the central axis of
the penetrating spray (a barrel of jet). This tendency is the result of low drag force
which causes very little build up in pressure ahead of the spray tip (which depends
essentially on the ambient gas viscosity and density). This phenomenon was observed
for all injection pressures and nozzle types experimented with in the atmospheric
pressure and temperature chamber.
24
Figure 3-14. Effect of gas density and viscosity on spray tip penetration; 140 MPa injection pressure;
30 mm3 fuelling; 0.2 mm single guided 5 hole VCO nozzle in the atmospheric pressure and
temperature chamber.
35
30
5-hole nozzle
7-hole Delphi
25
20
15
10
5
0
0
0.2
0.4
Time (ms)
0.6
0.8
Figure 3-15. Penetration with time for single/multi hole 0.2 mm single guided VCO, and 7-hole Delphi
injector, at 160 MPa injection pressure, and 6 MPa in-cylinder pressure at TDC.
Graphically, the hesitation observed from the single hole nozzle at 6 MPa in-cylinder
pressure at TDC (Figure 3-15), appears to be less gradual when compared with the
ambient atmospheric condition. However, the period of hesitation is almost identical
in both atmospheric chamber and 2 to 6 MPa in-cylinder pressure at TDC (there exists
a 0.02 ms difference for the single hole nozzle, this is considered negligible).
In addition, the following conclusions have been drawn from the hesitation and delay
of VCO nozzles.
Temporarily, there is an unbalanced pressure around the needle, once the needle
lift commences, especially with single hole nozzles. This unbalanced pressure
leads to transversal movement of the needle as it lifts.
There is a finite time required for the fuel to occupy the space evacuated by the
needle (Kennaird et al. [2002]).
At the end of injection, after apparent needle closure, a small amount of liquid fuel
exits the nozzle. This fuel exits as ligaments and spherical droplets the same size as
the nozzle orifice. This phenomenon was only observed with the single hole VCO
25
nozzle (for all the tests carried out at different injection and combustion pressures).
Figure 3-16 shows a comparison between the 0.2 mm single hole VCO and double
guided VCO nozzles.
Single guided
Double guided
3.3.4
The spray structure of Diesel fuel was observed with the high-speed Phantom video
camera for three different ICP at TDC (2, 4, and 6 MPa), for four different injection
pressures of 60, 100, 140, and 160 MPa, at two different air intake temperatures of
293 and 373 K (cold and hot intakes). The tests carried out were all for a 0.2 mm
26
single guided VCO multi and single hole nozzles for two fuelling rates of 30 and 50
mm3. Each test was repeated a number of times in order to ascertain repeatability for
liquid spray penetration and spray cone angle. The full tip penetration length was not
observed in this study due to the size restriction of the viewing windows within the
optical chamber.
On the basis of the images obtained, there exists a high density portion (intact liquid
core) where the spatial distribution of the liquid fuel is uniform relatively close to the
nozzle tip. Further downstream of the nozzle tip, the high density portion begins to
breakup, and the surface wave takes place on the outer edge of the spray structure
(Figure 3-17). This appeared as a repeatable fish-bone shaped structure, and was
evident for single and multi-hole nozzles tested. This phenomenon was much more
evident when the experiment was carried out in an ambient pressure (0.1 MPa) and
temperature chamber because of reduced air density.
For high ICP environment, the result of surface waves on the spray structure are more
obscured due to the greater absorption of light through the dense air.
Fish-bone structures due to surface wave have also been reported by Tsue et al.
[1992], Koo and Martin [1990], and Yule and Salters [1995]. These surface waves are
repeatable, and with an increase of the axial distance from the nozzle, the distance
between neighbouring stripes (fish-bones) increased with an increase in the
downstream distance.
60 MPa
100 MPa
140 MPa
160 MPa
Fish-bone structure
(Stripes)
Figure 3-17. Fish-bone shaped structure of the spray from a 0.2 mm single hole VCO single guided
nozzle, at 2 MPa ICP, cold intake. Images acquired with the Phantom camera.
27
An exact array of the photographs shown in Figure 3-17 is processed using the
Laplacian edge detection option described in literature review (Figure 3-18).
Figure 3-18 shows the axial distance between the high density portion of the spray
and the neighbouring stripes. Close examination of data has revealed the injection
pressure or the ICP has little effect on the distance between the stripes or the existence
of liquid core. In addition, the distances between the stripes are random and there is a
variation from frame to frame, as well as cycle to cycle (injection to injection) for any
given combination of ICP with injection pressure. Nevertheless, closer to the nozzle
the stripes are highly packed, and as the axial distance from the nozzle increases, the
distance between stripes increase.
60 MPa
100 MPa
140 MPa
160 MPa
Fish-bone structure
(Stripes)
Figure 3-18. Fish-bone shaped structure of the spray, using the Laplacian edge detection method of the
Phantom camera.
Figure 3-19 a and b shows a series of test images at 0.41 ms after first sight of fuel, at
an ambient temperature of 576 K at TDC, if an average polytropic coefficient of 1.3
and a compression ratio of 9:1 is assumed (from Chapter 2).
As expected, an increase in penetration length resulted from an increase in injection
pressure as well as a decrease in ambient gas pressure. This observation is consistent
with previous studies by Arai et al. [1984], Hiroyasu and Arai [2002], Kennaird et al.
[2002], Naber and Siebers [1996], and Reitz and Diwakar [1987].
28
60 MPa
100 MPa
140 MPa
160 MPa
34 mm from
the nozzle hole
Figure 3-19 a. Spray images taken at 2 MPa In-cylinder pressure, and 0.41 ms after first sight of fuel.
Images acquired with the Phantom camera.
60 MPa
100 MPa
140 MPa
160 MPa
34 mm from
the nozzle hole
Figure 3-19 b. Spray images taken at 6 MPa In-cylinder pressure, and 0.41 ms after first sight of fuel.
Images acquired with the Phantom camera.
As can be observed from the above images, the interaction between the stagnant gas
field and the leading edge of the spray during the initial stage of injection has the
effect of compressing clusters of droplets together as the spray jet penetrates. Further
downstream of the nozzle, the exchange of momentum between the droplets and the
local gas field allows entrainment of dense air along the edges of the spray. The effect
being smaller clusters of droplets stripping back or detaching from the bulk of the
spray. At the leading edge however, when the spray is fully developed, larger clusters
of droplets are starting to detach as the air entrainment increases while the droplets
momentum decreases (Figure 3-20).
29
1.276 ms ASOI
1.305 ms ASOI
1.334 ms ASOI
Figure 3-20. High-speed video sequence in steps of 29 s showing clusters of droplet starting to detach
along the leading edge of the spray at cold intake; acquired with the Phantom camera.
3.3.5
Figure 3-21 graphically represents the liquid spray penetration for a single, 3 and a 5
hole single guided 0.2 mm VCO nozzle obtained by high-speed video imaging
(additional profiles can be found in Appendix A). The curves for the multi hole
nozzles have been offset by 0.17 ms, to allow direct comparison. The number of holes
within a nozzle was found to have a notable effect on the rate of penetration for a
constant quantity of mass injected (injection time was varied according to the nozzle
used). This effect was found to be related to the drop in pressure across the nozzle
holes once the injector needle had reached full lift.
30
45
40
35
30
25
20
1 Hole at 160 MPa injection
15
10
5 Hole at 160 MPa injection
5
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
1.6
Figure 3-21. Spray penetration from a 0.2 mm VCO nozzle, at 160 MPa injection pressure; 6 MPa
ICP.
Figure 3-22 highlights the increased drop in rail pressure recorded by an AVL Indiset
high speed data acquisition system. The drop in rail pressure is clearly more dominant
with the 5 and the 3 hole nozzle as opposed to the single hole. This would indicate
that an increase in the number of holes (for a given orifice size) within a nozzle could
lead to a decrease in rail pressure. Subsequently, this has the effect of reducing the
injection pressure and hence the penetration rate. This phenomenon was observed for
all injection pressures. Figure 3-23 represents the current traces taken for the above
penetration rate with different nozzle configurations. It can be seen from the current
traces that the holding current for opening and closing phase of the nozzles tested are
identical.
180
160
140
120
100
Single Hole
3 Hole
80
5 Hole
Apparent Needle Full Lift
60
20
0
-2
Time (ms)
10
12
14
16
Figure 3-22. Rail pressure traces recorded by an AVL Indiset high speed data acquisition system at 160
MPa injection pressure, 6 MPa ICP, and 50 mm3 fuelling. Broken lines indicate apparent needle full
lift / closure. The nozzles tested were all 0.2 mm single guided VCO.
31
16
14
12
Current, 5 Hole
10
8
6
4
2
0
-2
Time (ms)
Figure 3-23. Injection current traces recorded by an AVL Indiset high speed data acquisition system at
160 MPa injection pressure, 50 mm3 fuelling.
Graphical representation of the rail pressure traces taken with the AVL Indiset data
acquisition system at 60 MPa injection pressure, and 6 MPa ICP is shown in Figure
3-24. In comparing Figure 3-22 and 3-24, it can be seen that the magnitude of drop in
rail pressure is dependent on the injection pressure set for a given nozzle. However,
the initiation of the drop in rail pressure is dependent on the number of holes within a
nozzle.
Single Hole
3 Hole
5 Hole
Apparent Needle Lift
Apparent Needle Closure 5 Hole
Apparent Needle Closure 3 Hole
Apparent Needle Closure Single Hole
100
80
60
40
20
0
-2
10
12
14
16
Time (ms)
Figure 3-24. Rail pressure traces recorded at 60 MPa injection pressure, 6 MPa ICP, and 50 mm3
fuelling.
The fluctuation in rail pressure and hence the pressure drop for all test cases, follows
the frequency of the fuel pressure regulator, which regulates and sets the pressure
32
output from the high pressure pump. The frequency of the pressure regulator is
superimposed on top of the fuel pumps frequency as shown in Figure 3-25.
The pressure regulator is basically an on/off switch with a fixed response time (for a
given pressure), therefore it requires a finite time to build up the rail pressure due to
the lag in the system.
170
160
150
140
130
120
110
100
(71.5 Hz)
90
80
70
60
50
0
10
20
30
40
Time (ms)
Figure 3-25. Frequency of the Common rail pump and the regulator recorded by an AVL Indiset high
speed data acquisition system.
3.3.6
As expected, the fuel rail pressure was found to have a significant effect on the rate of
penetration of the liquid core obtained by the high-speed Phantom camera. Figure 326 shows higher injection pressures produces more developed sprays in a shorter time
(additional profiles can be found in Appendix A), hence reduced breakup period. This
has the tendency to improve the vaporisation processes since the surface area of the
liquid core has increased. Shortly after the breakup period, the liquid penetration
profiles fluctuate around a slowly increasing average. As explained in Section 3.3.4,
these fluctuations were found to be the result of air entrainment, breaking away
clusters of droplets around the periphery of the leading edge of the spray tip.
Studies by Browne et al. [1986], Hiroyasu et al. [1989, 1990], Shimizu et al. [1984]
has shown similar trends for breakup profile.
33
50
45
40
35
30
25
20
15
10
160 MPa injection at 6 MPa ICP
5
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-26. Effect of injection pressure on spray penetration for a 0.2 mm 3-hole single guided VCO
nozzle, at 6 MPa ICP at cold intake.
This phenomenon was observed for all different nozzles tested as shown for the 5 hole
nozzle in Figure 3-27. Although the rate of penetration during the initial stage of
injection is faster for the 3 hole nozzle tested for a given rail pressures when
compared to the 5 hole nozzle (as explained in Section 3.3.5), the breakup period
remained almost the same (Figure 3-28).
50
45
40
35
30
25
20
15
10
5
0
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Time (ms)
Figure 3-27. Effect of injection pressure on spray penetration for a 0.2 mm 5-hole single guided VCO
nozzle, at 6 MPa ICP at cold intake.
34
50
45
40
35
30
25
60 MPa injection at 6 MPa ICP-3 Hole Nozzle
100 MPa injection at 6 MPa ICP-3 Hole nozzle
20
15
10
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-28. Effect of injection pressure on spray penetration for a 0.2 mm 3 and 5-hole single guided
VCO nozzle, at 6 MPa ICP at cold intake.
3.3.7
From the collected data the penetration length was observed to be directly dependent
on the ambient gas density as well as the injection pressure. This dependency was also
observed by Crua [2000], Hiroyasu and Arai [1989, 1990], Huh et al. [1998],
Kennaird et al. [2002], Naber and Siebers [1996], and Yule and Filipovic [1991].
35
40
35
30
25
20
Density 12.1 kg m^-3 (2 MPa ICP)
15
10
5
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-29. Effect of ambient gas density on liquid spray penetration for cold intake; 160 MPa
injection pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3 fuelling.
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.8
Time (ms)
Figure 3-30. Effect of ambient gas density on liquid spray penetration for cold intake; 100 MPa
injection pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3 fuelling.
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-31. Effect of ambient gas density on liquid spray penetration for cold intake; 60 MPa injection
pressure; 0.2 mm single guided 3 hole VCO nozzle at 50 mm3 fuelling.
36
The penetration profile data gathered in the current test were compared to the post
breakup correlation proposed by Hiroyasu and Arai [1989, 1990] as described in the
following relation;
When t > tb
1
P 4
1
( Dn t ) 2
Lp = k
g
(3.1)
P 4
1
Lp = 2.15 ( Dn t ) 2
g
50
45
40
35
30
25
20
15
10
Experimental Data
5
0
0
0.2
0.4
Time (ms)
0.6
0.8
Figure 3-32. Comparison between correlation proposed by Hiroyasu and Arai [1989, 1990] and
experimental results for non-evaporating spray (cold intake); 160 MPa injection pressure; 6 MPa ICP;
36.3 kg m-3 ambient gas density; 0.2 mm single guided 3 hole VCO nozzle.
As can be observed from the above figure, the correlation proposed by Hiroyasu and
Arai [1989, 1990] tends to over estimate the penetration rate during the early stage of
injection. However, as the injection pressure decreases, the proposed correlation tends
to over estimate the penetration rate for the entire injection duration (Figure 3-33).
37
50
P 4
1
Lp = 2.15 ( Dn t ) 2
g
45
40
35
30
25
20
15
10
5
Experimetal Data
0
0
0.2
0.4
0.6
0.8
Time (ms)
Figure 3-33. Comparison between correlation proposed by Hiroyasu and Arai [1989, 1990] and
experimental results for non-evaporating spray (cold intake); 60 MPa injection pressure; 6 MPa ICP;
36.3 kg m-3 ambient gas density; 0.2 mm single guided 3 hole VCO nozzle.
As described in the literature review, Hiroyasu and Arai [1989, 1990] proposed two
correlations for spray penetration profile; a pre and post breakup correlation, where
during the pre breakup the authors assumed a linear relationship between Lp and t
exists. Thus, the authors further assume the breakup zone ends at the intersection of
post and pre breakup correlation giving a breakup zone length.
However, although Yule and Filipovic [1991] agree that there is a transition from t1 to
t1/2, they indicate that the transition is more gradual and without the clear break points
implied by Hiroyasu and Arai [1989, 1990]. This gradual transition without a clear
break point was also observed in this study. Therefore, the correlation proposed by
Yule and Filipovic [1991] was compared to the current spray penetration data, as
described in the following relation:
Lp = k
g
3 5
(Dn t ) 2 tanh t
tb
(3.2)
Again, an empirical constant of 2.15 instead of the proposed constant 3.8 by Yule and
Filipovic [1991] gave the best agreement with the experimental data obtained before
the transition to a steadier penetration length (Figure 3-34).
38
L p = 2.15
g
45
40
(Dn t )
35
t
tanh
tb
35
30
25
20
15
Experimental Data
10
5
0
0
0.2
0.4
Time (ms)
0.6
0.8
Figure 3-34. Comparison between correlation proposed by Yule and Filipovic [1991] and experimental
results for non-evaporating spray (cold intake); 160 MPa injection pressure; 6 MPa ICP; 36.3 kg m-3
ambient gas density; 0.2 mm single guided 3 hole VCO nozzle.
Equation 3.2 has the breakup time tb as an adjustable constant, so that, for an injection
pressure, the best fit between experimental data and Equation 3.2 provides a value for
tb (Yule and Filipovic [1991]).
As can be seen from figure 3-34, the above correlation for the initial stage of injection
is in excellent agreement, and overall, in good agreement with the experimental data
obtained. This agreement was consistent for all injection pressures as shown in Figure
3-35 for 60 MPa injection pressure.
50
45
40
L p = 2.15
g
35
30
(Dn t )
3 5
t
tanh
t b
25
20
15
correlation with k =2.15
10
5
Experimetal Data
0
0
0.2
0.4
0.6
0.8
Time (ms)
Figure 3-35. Comparison between correlation proposed by Yule and Filipovic [1991] and experimental
results for non-evaporating spray (cold intake); 60 MPa injection pressure; 6 MPa ICP; 36.3 kg m-3
ambient gas density; 0.2 mm single guided 3 hole VCO nozzle.
39
From experimental spray penetration data, and subsequently the penetration length
correlation (Equation 3.2), the breakup parameters can be derived (Yule and Filipovic
[1991]).
By curve fitting Equation 3.2 to the penetration data, a value for tb and Lb is obtained
for various in-cylinder gas densities and injection pressures (Figure 3-36 and 3-37).
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
10
15
20
25
30
35
40
Figure 3-36. The effect of in-cylinder gas density on breakup time for non-evaporating spray (cold
intake); 0.2 mm single guided 3 hole VCO nozzle.
20
18
16
14
12
10
8
160 MPa injection pressure
0
10
15
20
25
30
35
40
Figure 3-37. The effect of in-cylinder gas density on breakup length for non-evaporating spray (cold
intake); 0.2 mm single guided 3 hole VCO nozzle.
40
As can be observed from Figure 3-36, variation of in-cylinder gas density has an
insignificant effect on breakup time. However, the in-cylinder gas density has a
significant effect on breakup length especially at lower injection pressures (Figure 337).
The effect of gas density on breakup time and length was found to follow the trend
observed by Yule and Filipovic [1991], even though a direct comparison cannot be
made due to the alternative injection pressure and in-cylinder gas density range used
in this study (12.1 to 36.3 kg m-3, 60 to 160 MPa injection pressure as opposed to 14.2
to 60 kg m-3, 21 to 31 MPa injection pressure).
Figure 3-38 and 3-39 show the breakup time and length respectively as a function of
injection pressure for three different in-cylinder gas densities (a table of breakup time
and length for different injection and ambient gas pressures can be found in Appendix
A).
0.3
12.1 kg m^-3
24.2 kg m^-3
0.25
36.3 kg m^-3
0.2
0.15
0.1
0.05
0
40
60
80
100
120
140
160
180
Figure 3-38. Breakup time as a function of injection pressure for varied in-cylinder gas density; nonevaporating spray (cold intake); 0.2 mm single guided 3 hole VCO nozzle.
As can be observed from Figure 3-38 and 3-39, the injection pressure has a substantial
effect on the breakup time and length; as the injection pressure increases the breakup
time and length decrease. This effect is beneficial due to increased surface area of the
41
fuel in a relatively shorter time, assisting in a better mixture preparation. Again, this
trend was consistent with previous studies by Yule and Filipovic [1991].
18
12.1 kg m^-3
17
24.2 kg m^-3
16
36.3 kg m^-3
15
14
13
12
11
10
9
8
40
60
80
100
120
140
160
180
Figure 3-39. Breakup length as a function of injection pressure for varied in-cylinder gas density; nonevaporating spray (cold intake); 0.2 mm single guided 3 hole VCO nozzle.
3.3.8
The effect of increased ambient gas temperature (721 K at TDC) was clearly seen to
reduce the rate of penetration as the ambient gas density was increased. This was
clearly the result of droplets evaporating around the periphery of the penetrating spray
(Figure 3-40).
Nevertheless, the two distinct phases that were observed with the cold gas intake,
were also observed for elevated ambient gas temperature; that is, an almost linear
phase, followed by a rapid transition to a steadier penetration length fluctuating
around a slowly increasing average as shown in Figure 3-41 and 3-42 (additional
profiles can be found in Appendix A).
The rate of penetration during the initial stage of injection however was not
observed to be greatly influenced by the gas temperature, but rather the injection
pressure and the in-cylinder gas density. Further downstream however, the penetration
42
length decreases rapidly with increased evaporation of droplet clusters due to higher
gas temperatures.
Figure 3-40 and 3-43 highlight the influence of ambient gas temperature and the
reduction in penetration length at 0.6 ms after first sight of liquid fuel, from about 35
mm at 576 K at TDC to 25 mm at 721 K at TDC.
Cold intake (g = 36.3 kg m-3)
Figure 3-40. Comparison between non-evaporating and evaporating spray at 160 MPa injection
pressure; 6 MPa ICP; 50 mm3 fuelling; 0.2 mm single guided 3 hole VCO nozzle; 0.6 ms after first
sight of fuel; acquired with the Phantom camera.
50
45
40
35
30
25
20
Density 9.7 kg m^-3 (2 MPa ICP)
15
10
5
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-41. Effect of ambient gas density on liquid spray penetration for hot intake at 160 MPa
injection pressure; 0.2 mm single guided 3 hole VCO nozzle and 50 mm3 fuelling.
43
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-42. Effect of ambient gas density on liquid spray penetration for hot intake at 60 MPa
injection pressure; 0.2 mm single guided 3 hole VCO nozzle and 50 mm3 fuelling.
50
45
40
35
30
25
20
15
Density 36.3 kg m^-3 (6 MPa ICP cold intake)
10
5
0
0
0.2
0.4
0.6
0.8
1
1.2
Time (ms)
1.4
1.6
1.8
Figure 3-43. Effect of ambient gas temperature on liquid spray penetration length at 160 MPa injection
pressure; 0.2 mm single guided 3 hole VCO nozzle and 50 mm3 fuelling.
As can be seen from Figure 3-43, at about 0.26 ms after first sight of liquid fuel
(second phase of spray penetration for hot air intake), liquid penetration is largely
influenced by the ambient gas temperature rather than the ambient gas density. This
phenomenon was observed for all injection pressures and nozzles (single and 5 hole)
in the current test.
44
3.4
Conclusion of chapter 3
From the experimental data obtained via the high speed camera, a number of areas
have been studied. There are the influence of the spray isolator on multi-hole nozzles,
and the formation and breakup of Diesel fuel spray, for cold and hot ambient gas
intake at various in-cylinder densities and injection pressures, for single and multihole nozzles. The following will outline the main conclusions drawn from the
investigation.
The use of the spray isolator resulted in successful isolation of one Diesel spray from
a multi-hole nozzle, with minimal effect on the development of the other sprays. This
will allow successful measurement of droplet size and velocity when the PDA
technique is applied.
Hesitation during the initial stage of injection for a single hole nozzle was attributed
to the transversal movement of the needle. However, after this initial hesitation, the
subsequent spray penetration rate was unaffected.
The images obtained experimentally, have revealed the existence of a high density
portion (intact liquid core), where the spatial distribution of the liquid fuel is uniform
relatively close to the nozzle tip. This was observed to be independent of injection and
the in-cylinder pressures.
The effect of air entrainment was observed to increase detachment of larger clusters of
droplets around the periphery of the leading edge as well as stripping back smaller
clusters of droplets on the edge of the spray cone. The detachment and stripping of the
droplets were found to be directly dependent on injection and in-cylinder pressures.
45
The number of holes within a nozzle was found to have a significant effect on the rate
of penetration due to the drop in rail pressure. It is therefore not straightforward to
validate a multi-hole nozzle on the basis of a single hole nozzle, due to the limitations
sustained by each VCO configuration.
The spray penetration profiles for all the nozzles tested were dependent on in-cylinder
and injection pressure conditions.
A new empirical coefficient for the earlier proposed correlation by Hiroyasu and Arai
[1989, 1990] and Yule and Filipovic [1991] was derived from experimental
penetration data.
The correlation of Hiroyasu and Arai [1989, 1990] however was not applicable, as it
over estimated the penetration rate during the early stage of injection for high
injection pressure (160 MPa), and over estimated the penetration rate for the entire
injection duration for low injection pressure (60 MPa).
Yule and Filipovic [1991] correlation showed a very good agreement for all the
nozzles tested. Nevertheless, for the single hole nozzle, a time constant to counteract
the initial hesitation period was implemented.
The derivation of the new empirical coefficient for the earlier correlation proposed by
Yule and Filipovic [1991], allows breakup time and breakup length for liquid spray
penetration to be obtained. From data, the variation of ambient gas density showed
insignificant effect on breakup time, while a substantial effect on breakup length,
especially at low injection pressures was observed. However, variation of injection
pressure showed significant effect on breakup time and breakup length.
The effect of high ambient gas temperature was seen to clearly increase vaporisation
and reduce tip penetration. For the early stage of injection however, no change in tip
penetration was observed.
46
4.
4.1
Introduction
In order to substantiate the consistency of the volume flow for a given duration and
rail pressure for a given Diesel fuel injector, a series of rate tests using a Lucas rate
gauge was carried out. In addition, the implementation of the rate gauge allowed
direct comparison of a set of Delphi 7 hole VCO injectors with identical flow rates
and coefficient of discharge (Cd) from a batch.
The Delphi injectors used in the current test are solenoid type actuators. However,
these injectors are believed to be of a fast response type, competing currently with the
more known fast response piezoelectric types. In view of this, the Delphi injectors are
tested for multiple-injection strategy within a cycle. Using the rate gauge technique it
allows measurement of injected fuel for each consecutive injection and the assessment
of the dwell period between injections. Consequently, successful implementation and
assessment of the data will provide a test matrix that will be used for high-speed video
imaging and PDA technique, utilising a multiple-injection strategy.
4.2
The basic operating concept of the Lucas rate gauge technique is described below.
The fuel is injected into a long pipe containing the same working fluid under pressure.
This in turn creates a pressure wave which travels along the pipe at speed of the sound
in the fluid. The theory of pressure waves in liquids states that, at any instant in time,
the magnitude of the pressure wave is directly proportional to the flow velocity at a
given cross section in the pipe. Given a pipe of uniform cross sectional area, the flow
velocity, and hence pressure, is proportional to the rate of injection (Lucas 1995).
Fitted to the gauge near the injector nozzle outlet, a pair of strain gauges measure
variation in injection pressure and a second pair of strain gauges were used for
temperature compensation. The strain gauges were connected to a bridge amplifier
and the output was displayed on a storage oscilloscope. The rail pressure, injector
current, and the TTL signals were also recorded on the storage scope for analysis. The
47
variation in rail pressure, the delay in commencement of injection and the time
response for needle full lift were also recorded. However, due to the complexity of the
Delphi injector and the lack of a conventional control-plunger within the injector
body, no needle lift sensor could be integrated in to the system without changing the
flow-pattern. For all traces, an average of 100 cycles was recorded and compared with
instantaneous recordings to ensure repeatability.
In order to make a direct comparison for the rate of injection between the Delphi
injectors (4 in a batch), the injection duration for a given rail pressure was fixed, and
each injector was mounted on the Lucas rate gauge in turn. With injection pressures of
60, 100, 140, and 160 MPa, the injection duration varied from 0.3 ms to 1.7 ms. The
shortest duration of 0.3 ms was chosen due to the time taken by the needle to reach
full lift. This duration is approximately 0.25 ms. Figure 4-1 represents the variation in
rate of injection from all 4 Delphi injectors.
Rate 4
Current 1
0.5
0.5
Current 2
Current 3
0.4
Rate (V)
0.6
Rate 1
Rate 2
0.4
Current 4
TTL1
TTL 2
TTL 3
0.3
0.3
TTL 4
0.2
0.2
0.1
0.1
0
-0.1
0
0
-0.1
Current (mA)
0.6
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Time (ms)
0.8
0.9
1
-0.1
Figure 4-1 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO Delphi injectors;
60 MPa injection pressure; 0.41 ms injection duration.
As can be seen in Figure 4-1, the TTL and current durations are identical for all 4
injectors, while the response time to reach peak current is much shorter with injector
No 3, indicating faster response time when compared with the other current traces.
However, it was believed that the solenoid injector No 3 had a very low inductance.
This could be the result of a dramatic reduction in number of turns in the coil winding
due to a short-circuit, giving a very fast response time.
48
Also as shown in Figure 1, injector No 1 reaches peak current slightly faster than
injector No 2 and 4 by 0.05 ms. However, the opening delay (0.1 ms) for injector No
1, when compared with injector No 2 and 4, for all injection durations and fuel rail
pressures examined was constant, indicating possible differences in the inductance
and/or magnetic permeability of the solenoid core material.
As explained, the fuel rail pressure had no influence on the injection delay for the
Delphi injectors examined. However, previous experiments with Bosch injectors
(Chapter 3) and Henein et al. [2002] have revealed that the injection delay (for Bosch
injectors) is dependant on injection pressure.
From Figure 4-1, the displaced area under the injection rate curve for injector No 1 is
almost twice as much when compared with injector No 2 and 4, while injector No 3
exhibits no rate of injection. The trend for injector No 2 and 4 when compared with
No 1 was observed for all injection pressures at the lower end of injection duration.
After careful examination of the fuel rail pressure, the difference in the rate of
injection is believed to be the result of the longer time taken to reach peak current
(0.05 ms) needed to pull the needle to its maximum lift position and holding the
needle open. To test this hypothesis, the injection duration for injector No 2 was
increased by 0.05 ms and compared with injector No 1 as shown in Figure 4-2.
0.7
0.6
Rate 1
Rate 2
0.6
0.5
Current 1
Current 2
0.4
TTL1
Rate (V)
0.4
TTL 2
0.3
0.3
0.2
Current (mA)
0.5
0.2
0.1
0.1
0
0
-0.1
0
-0.1
0.1
0.2
0.3
0.4
0.5
Time (ms)
0.6
0.7
0.8
0.9
1
-0.1
Figure 4-2 Comparison of injector No 1 and No 2; 60 MPa injection pressure; 0.41 ms injection
duration for injector No1, and 0.46 ms injection duration for No2.
49
As can be seen from Figure 4-2, the delay between the two injected rates were the
same as before, while the integrated area under each curve was identical, giving the
same rate of injection. For longer injection durations however, there was almost no
difference in the rate of injection as the difference in time to reach peak current when
compared with the overall injection duration was insignificant (consequently, this will
allow the needle to reach its maximum opening lift position holding the needle open,
as shown in Figure 4-3).
0.8
0.8
0.7
0.7
0.6
0.6
Rate 1
Rate 2
0.5
Rate 4
Current 1
0.4
0.4
Current 2
0.3
0.3
Current 4
Current (mV)
Rate (V)
0.5
TTL1
0.2
0.2
TTL 2
TTL 4
0.1
0.1
0
-0.1
-0.1
0
0.1
0.3
0.5
0.7
0.9
1.1
1.3
1.5
1.7
1.9
2.1
2.3
2.5
2.7
Time (ms)
-0.1
Figure 4-3 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO Delphi injectors;
60 MPa injection pressure; 1.25 ms injection duration.
For injector No 3 however, the same trend was observed throughout; that is, no trace
of injection regardless of injection duration or rail pressure.
Further experiments into variation in the rate of injection from injector to injector
were carried out with splitting the injection event (multiple-injection). Again a fixed
duration was used allowing direct comparisons to be made as with the single injection
strategy (Figure 4-4). As expected, the TTL and current duration were identical for all
4 injectors, while the peak current response time varied as previously.
As can be seen in Figure 4-4, the area under rate trace No 4 for the first injection
duration was less than that of No 2 when compared with single injection event, while
trace No 1 was closer to trace No 2. However, for the second injection period, the rate
of injection is almost identical for No 2 and No 4 as in the single injection strategy.
50
1.2
1.2
Rate 1
Rate 2
Rate 3
Rate 4
Current 1
0.8
Current 2
0.8
Current 4
0.6
0.6
TTL1
TTL 2
0.4
0.4
TTL 3
Current (mA)
Rate (V)
Current 3
TTL 4
0.2
0
-0.1
0.2
0
0.1
0.3
0.5
-0.2
0.7
0.9
1.1
1.3
Time (ms)
1.5
1.7
1.9
2.1
2.3
-0.2
Figure 4-4 Comparison of injected rate, current, and TTL from a batch of 7 hole VCO Delphi injectors
with multiple injection strategy; 0.6 ms dwell time; 160 MPa injection pressure; 0.4 ms injection
duration.
As for injector No 3, the behaviour for the first injection duration was the same as the
single injection event. However, as Figure 4-4 shows, the injected rate trace for No 3
was clearly apparent (during the second injection period). This phenomenon was also
observed for triple injection event, i.e., no rate trace for the first injection event but
evident during the second and third injections. The behaviour of injector No 3 could
not be explained.
From the data collected, it was deemed necessary to calibrate each injector
individually. For this study, injector No 2 and 4 were chosen as they gave an identical
current match, followed by a close rate match.
With each injector mounted on the fuel calibration test-bed in turn, a series of pulse
lengths corresponding to an injection duration for a given rail pressure (60, 100, 140,
and 160 MPa) were sent from the injector controller and the times recorded.
By adjusting the injection duration for a given rail pressure the desired volume flow
(5, 10, 15, 20, 30, 40, and 50 mm3) was measured and recorded for an average of 100
cycles. With the volume flow rates established, the injectors were mounted on the
Lucas rate gauge as previously, and the injected fuel rates were recorded on the
storage scope (Figure 4-5).
51
26
Filtered rate
22
Rate of injection
18
14
10
6
2
-2 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Time (ms)
Figure 4-5 Comparison of injection rate with the filtered rate in mg/ms. 100 MPa injection pressure;
0.38 ms injection duration corresponding to 5 mm3 volume of fuel; Delphi injector No 2.
The data was then exported to an Excel worksheet, and the area under the rate curve
between the thresholds was matched and integrated. The level of the threshold chosen
was the minimum positive value of intensity after the last zero crossing point in
positive direction. For each setting, the calibration factor (ratio of volume and the area
under the curves) was calculated and the rate of injection in mg/ms per injection was
established (Figure 4-5). With the calibration factor known, the data was then applied
to multiple-injection strategies in order to establish the variation in the injected mass
for each consecutive injection.
With injector No 2 mounted on the Lucas rate gauge as before, the dwell period
(separation time between each pulse for multiple-injection) was varied in the range of
0.3 to 1 ms, in increments of 0.1ms. In addition, the injected volume was also varied
in the range of 5 to 20 mm3, with the rail pressures set at 60, 100,140, and 160 MPa.
Figures 4-6 and 4-7 represent the rate of injection, current and TTL signal for
multiple-injections, with the addition of a single injection event superimposed on the
same figures for comparison (once for the first and once for the second injection
event). Again in Figure 4-6, the TTL signal and the current durations are identical for
each consecutive injection, while the rate of injection is about 30% less for first
injection compared to the second injection event, being closer (+ 4%) to the single
injection target. This inconsistency was observed at all injection pressures, dwell
periods, and injection durations as shown in Figure 4-7.
52
16
0.7
Rate of injection
14
0.6
Rate (single)
Current
0.5
Current (single)
10
0.4
TTL
8
TTL (single)
0.3
6
0.2
Current (mA)
12
4
0.1
0
-0.2
0.2
0.4
0.6
0.8
-2
1.2
1.4
1.6
1.8
2.2
2.4
2.6
2.8
-0.1
Time (ms)
Figure 4-6 Differences in the rate of injection between single injection event (solid lines) and multipleinjection (broken lines). The data for the single injection has been superimposed on the above Figure
twice, with an offset to match the second injection event. Injection pressure at 60 MPa; 0.46 ms for
each injection duration; 0.5 ms dwell period.
25
0.4
TTL
TTL (single)
15
0.3
10
0.2
0.1
0
-0.2
-5
Current (mA)
20
0
0
3.2 3.4
-0.1
Figure 4-7 Differences in the rate of injection between single injection event (solid lines) and multipleinjection (broken lines). The data for the single injection has been superimposed on the above Figure
twice, with an offset to match the second injection event. Injection pressure at 60 MPa; 0.72 ms for
each injection duration; 0.5 ms dwell period.
Further investigation into the possible cause of this variation in the rate of injection
was carried out. The data has revealed the drop in rail pressure is more substantial for
the second injection than the first. However, the data also revealed that the rail
pressure at the start of the second injection was at a higher starting value compared to
the first injection (Figure 4-8). This is because as the needle closes, the fuel pressure
builds up and reaches a value greater than the starting value (start of injection).
53
Therefore, the reduction in the rate of injection for the first injection duration could be
the result of lower starting rail pressure.
70
45
Rail pressure
40
65
Rate of injection
P ressure (M P a)
30
55
25
50
20
45
15
40
10
35
30
-0.2
35
60
0
0.2
0.6
1.4
1.8
Time (ms)
2.2
2.6
Figure 4-8 Differences between fuel pressures in the rail prior to injection. Injection pressure at 60
MPa; 0.72 ms for each injection duration; 0.5 ms dwell period.
Further inspection of the data has shown a certain level of variation between current
traces for each consecutive injection. This variation was in the form of time taken to
reach peak current when compared with the single injection current traces (Figure 49). The difference in time was measured to be in the order of 0.025 ms. However, this
variation in time was not constant. Nevertheless, as explained previously, the smallest
fluctuation in duration to reach peak current could translate into a significant
fluctuation in injection duration when the command pulse was relatively short.
25
Current (multiple-injection)
20
Current (single)
Current (A)
TTL (multiple-injection)
15
TTL (single)
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
1.6
Figure 4-9 Variation between single and multiple-injection current traces. The data for the single
54
injection strategy has been superimposed on the above Figure twice, with an offset to match the second
injection event. Injection pressure at 60 MPa; 0.46 ms for each injection duration; 0.5 ms dwell period.
To verify the consistency of the injector control unit with subsequent influence on the
injector solenoid, the injector was replaced by a resistor and then the injector was
operated without any fuel pressure as shown in Figure 10.
As can be seen from Figure 4-10, the current trace (solid line) taken for the resistor in
place of the injector shows no variation from shot to shot indicating consistency of the
injector control unit. This phenomenon was also observed for triple injection event.
24
current (Injector)
20
current (Resistor)
16
Current (A)
TTL (Injector)
TTL (Resistor)
12
0
0
0.2
0.4
0.6
0.8
1.2
1.4
1.6
Time (ms)
Figure 4-10 Current trace for an injector (doted line) and a resistor (solid line). Injection duration 0.46
ms; dwell period 0.5 ms.
Current traces were also taken for the injector with no fuel pressure (Figure 4-10). The
inconsistency between the first and the second injection were repeated, and again the
same inconsistencies were observed for the triple injection event.
From the data collected in the current experiment, the only remaining possibility for
the variation in the rate of peak current, which has a direct influence on the rate of
injection, is the characteristics of the Delphi injector, when multiple-injection strategy
is applied
Dwell periods between 0.3 and 1ms in increments of 0.1 ms were examined. For
injection durations longer than 0.6 ms, regardless of the injection pressure (between
60 and 160 MPa), dwell periods less than 1 ms were found to be totally ineffective,
resulting in one continuous injection duration. Since this phenomenon occurred
55
independent of rail pressure, it can only be assumed that at relatively long injection
durations the solenoid becomes so charged that at the end of the first pulse, flux
dissipation does not occur instantaneously. However, due to lack of information on
time/frequency response and dynamic behaviour of the solenoid further investigation
is needed to correctly establish the characteristics of such injectors.
4.3
Conclusion of chapter 4
Using a Lucas rate gauge on a set of 7 hole VCO Delphi injector, the rate of injection
for single and multiple-injection strategy within a cycle was examined.
For single injection strategies, the rate of injection from injector No 1 was found to be
greater than injectors 2 and 4 for the same duration at low volume flows. This was
also evident from the current trace data, showing a faster response to reach peak
current with subsequently longer needle lift duration.
This hypothesis was further examined by increasing the injection duration by the
difference in time between the two peak current traces (0.05 ms). The result was an
identical rate of injection for injector No 2 compared to injector No 1.
The delay for injector No 1 at 0.325 ms and for injector No 2/4 at 0.425 remained
constant for all injection durations and fuel rail pressures.
The time taken to reach peak current, which was directly related to the time taken to
reach needle full lift, was critical for the rate of injection with short injection durations
(low volume flows). For longer injection durations, no significant difference in the
rate of injection was found. This is the result of relatively short needle lift period
compared to injection duration.
For multiple injection strategies the rate of injection was up to 30% less for the first
injection at low volume flows (5 mm3). However, for the second injection the rate of
injection was found to be closer to the target single injection strategy. This
inconsistency was observed for all dwell periods (between 0.3 and 1 ms), injection
pressures and injection durations.
The inconsistencies observed were believed to be the result of differences in the rail
pressure prior to injection, and fluctuations in the time to reach peak current.
56
For the former the drop in rail pressure was more substantial for the second injection
than the first. However, at the start of the second injection event (for multiple
injections) the rail pressure starts at a higher value compared to the first injection.
This will effectively increase the mass flow rate. For the latter, the variation in the rate
of injection for each consecutive injection was also seen in the current trace data in
the form of fluctuations in time to reach peak current. These fluctuations in time to
reach peak current were believed to be directly related to the rate of needle lift which
will translate directly to needle opening period. Coupling these two phenomena the
rate of injection could be severely influenced depending on the magnitude of each.
The consistency of the injector drive unit was established by replacing the Delphi
injector with a known resistor and examining the current trace data.
For a resistor in place of an injector, no variation in the current trace was found from
injection to injection. Also the variation in the current traces remained with no rail
pressure.
Dwell periods less than 0.5 ms were found to be ineffective when the injection
duration for the first split or the second split of a triple injection strategy exceeded 0.6
ms duration, regardless of the injection pressure. The results were one continuous
injection as if it were single injection strategy. This phenomenon was assumed to
occur due to non-instantaneous dissipation of the charge within the solenoid after
commanded injection signal elapsed.
57
5.
EXTENSION TO PHD
5.1
Introduction
This goal has been achieved to some degree by introducing multiple injections during
a cycle. The term multiple injections has a broad meaning. Pilot injection and split
injection are some of the examples of this strategy. Normally, pilot injection refers to
relatively small quantity of fuel injected early in the cycle before the main injection
(Su et al. [1996]). Split injection consists of a number of short burst injections with
varied dwell time between them (Figure 5-1). Unlike pilot injection, split injection
could have equal quantities of fuel or a variable amount of fuel, injected for each burst
of the injection phase. Nevertheless, studies by Lee et al. [2003], Tow et al. [1994],
and Pierpont et al. [1995] has shown that combined high injection pressures and
multiple injection are an effective way to improve Diesel combustion emissions.
Pilot
Main injection
Pilot injection
Dwell
1st Split
2nd Split
Split injection
Dwell period
58
5.1.1
Traditionally, for a single injection strategy, the calibration of mass flow rate is fairly
straight forward. For a given injection duration, the mean volume of fuel injected can
be measured, thus by fine tuning the injection duration, the desired mass flow rate is
accomplished. For multiple injection strategy however, the dwell time between each
injection plays an important part in the amount of mass injected for each consecutive
injection. If the dwell time is too short, the injector needle will not have enough time
to come to rest position, and injection occurs as one continuous injection. If too long,
it will be similar to a single injection strategy.
By employing a mass rate tube technique, the desired mass flow rate can be calibrated
for any combination of injected mass and dwell time. This information is vital in
determining the stoichiometric mixture ratio in an internal combustion engine, and
understanding the interactions between split injections.
59
5.1.2
By using a high speed camera and side-lighting and/or laser sheet setup, video
recordings of complete injection cycles with multiple injection strategy will provide
valuable information on the evolution of liquid spray penetration, and the spray cone
angle for each consecutive injected fuel mass. Information on structure and interaction
between each consecutive spray and possible air entrainment into the spray can be
deduced from images when high speed video imaging technique is implemented.
Also, the difference in mean injection velocity of the first and the second spray can
hold information on the velocity of air entrainment.
The dynamic behaviour of two split injections were also studied by Arai and Amagai
[1997] using high speed video recordings. The authors observed that the spray tip
velocity of the second spray was higher than the first spray. Also this tendency
increased as the dwell time decreased. The results were explained due to the flow field
formed by the first spray. However the dwell periods examined by the authors were
not extensive. Also the injection and in-cylinder pressures effects were not examined
by the authors. However, experiments planned in the forthcoming months will look
into the effects on varied dwell period, as well as high in-cylinder and injection
pressures with multiple injection strategy with the above technique.
5.1.3
In addition to the above, simultaneous droplet sizes and droplet velocities can be
measured at any given axial or radial position (along the spray axis) accurately with
the PDA technique. This information is paramount in establishing the aftermath of
interaction between injected sprays, where collision and coalescence changes the drop
size and velocity distributions.
This change in distribution of droplets with a multiple injection strategy was also
observed by Yoshizu et al. [1991], when PDA technique was implemented. The
authors found that the pilot injection spray had a more stratified size distribution,
meaning that the droplet sizes increased from the spray axis to some radial location,
and then from that location it decreased to the spray edge.
60
Laser power, photomultiplier voltage and signal validation level on data rate
acquisition has to be optimised and calibrated for measurements of dense
Diesel spray.
Fine tuning the probe volume should eliminate multiple droplet occupancy, as
well as avoiding too small a size probe volume which will result in improper
sizing of droplet particles.
The effect of Diesel fuel and the ambient gas refractive indices, as well as the
optical window (sapphire) refractive index has to be considered. These effects
will have a direct influence on the effective beam path that forms the probe
volume and the collection optics.
5.1.4
Modelling
This part of the thesis will be focused on linking the experimental data with available
phenomenological models, as well as development of the theory where applicable.
These models could be on spray penetration before and after breakup length (Sazhin
et al. [2001, 2003]).
In addition, the results of present and future experimental work could be validated
against the CFD predictions performed by modelling group at University of Brighton.
Also, the experimental results obtained in this work can serve as a benchmarking tool
for advanced spray modelling. A research on emissions such as soot produced by
combusting sprays is under investigation by Mr R Demory. Since the current work is
mainly focused on non-combusting sprays, it makes it complementary to that of spray
autoignition (Crua [2002]) and on-going work by Mr R Demory. In brief, the current
studies are seen as an integral part of the ongoing research at University of Brighton.
61
References
REFERENCES
Afzal H., Arcoumanis C., Gavaiser M., Kampanis N. Internal Flow in Diesel
Injector Nozzles-Modelling and Experiments. IMechE, S492/S2/99, 1999
Anezaki Y., Shirabe N., Kanehara K., Sato T. 3D Spray Measurement System for
high Density Field Using Laser Holography. SAE 2002-01-0739, 2002
Arai M., Amagai K. Experimental Study on a Diesel Spray of Multi-stage Injection.
International Symposium COMODIA 94, 1994
Arai M., Amagai K. Dynamic Behaviour of Multi-Stage Injection Diesel Spray. SAE
97004, 1997
Arai M., Tabata M., Hiroyasu H., Shimizu M. Disintegration Processes and Spray
Characterization of Fuel Jet Injected by a Diesel Nozzle. SAE 840275, 1984
Arcoumanis C., Flora H., Gavaises M., Badami M. Cavitation in Real-Size MultiHole Diesel Injector Nozzles. SAE 2000-01-1249, 2000
Arcoumanis C., Gavaises M., Flora H., Roth H. Visualization Of Cavitation in
Diesel Injectors. Mec.ind.(2001) 2, page 375-381
Arcoumanis C., Gavaises M., French B. Effect of Fuel Injection Processes on the
Structure of Diesel Sprays. SAE 970799, 1997
Badock C., Writh R., Fath A., Leipertz A. Investigation of Cavitation in Real Size
Diesel Injection Nozzles. Elsevier, International journal of Heat and Fluid Flow, Vol
20, PP 538-544, 1999
Bae C.H., Kang J. Diesel Spray Characterisation of Common-Rail VCO Nozzle
Injector. THIESEL 2000. Thermofluidynamic Processes in Diesel Engine
Borman G.L., Ragland K.W. Combustion Engineering. New York: McGraw-Hill;
1998.
Browne K.R., Partridge I.M., Greeves G. Fuel property Effect on Fuel\Air Mixing
in an Experimental Diesel Engine. SAE 860223
Chaves H., Knapp M., Kubitzek A. Experimental Study of Cavitation in the Nozzle
Hole Of Diesel Injectors Using Transparent Nozzles. SAE 950290, 1995
Chou W., Faeth G.M. Temporal Properties of Secondary drop Breakup in the Bag
Breakup regime. International Journal of multiphase Flow, Vol 24, PP 889-912, 1998
Crua C. Combustion Processes in a Diesel Engine. PhD thesis, December 2002
62
References
Dantec Instruction & Users guide. Publication no: 9040U1101, date: 15 April 1999.
Copyright 1999 by Dantec measurement technology A/S, P.O. Box 121, Tonsbakken
18, DK 2740 Skovlunde, Denmark, 2000
Dai Z., Faeth G.M. Temporal Properties of Secondary Breakup multimode Breakup
Regime. International Journal of multiphase Flow, Vol 27, PP 217-236, 2000
Dec E.J. A Conceptual Model of DI Diesel Combustion Based on Laser-Sheet
Imaging. SAE 970873, 1997
Dent J.C. A Basis For the Comparison of Various Experimental Methods for
Studying Spray Penetration. SAE 710571, 1971
Farrell P.V., Chang C.T., Su T.F. High Pressure Multiple Injection Spray
Characteristics. SAE 960860, 1996
Fox T.A., Stark J. Discharge Coefficients for miniature Fuel Injectors. G
01188,IMechE 1998
Hattori H., Narumiya K., Tsue M., Kadota T. Photographical analysis of Initial
Breakup Processes of Diesel spray. Thiesel 2002, Conference on Thermal- and fluid
Dynamic Process in Diesel Engines. 2002
Heimgartner C., Leipertz A. Investigation on the Primary Spray Breakup Close to
the Nozzle of a Common-Rail High Pressure Diesel Injection system. SAE 2000-011799. 2000
Henein N.A., Lai M-C., Singh I.P., Zhong L., Han J. Characteristics of a common
rail Diesel injection system under pilot and post injection modes. SAE 2002-01-0218,
2002
Hiroyasu H., Arai M., Tabata M. Empirical Equation for the Sauter Mean Diameter
of a Diesel spray. SAE 890646, 1989
Hiroyasu H., Arai M. Structure of Fuel Spray in Diesel Engines. SAE 900475, 1990
Hiroyasu H. Spray Breakup Mechanism From the Nozzle and its Applications.
Atomization and Sprays, Vol 10, PP 511-527, 2000
Hiroyasu H. The Structure of fuel Sprays and the Combustion processes in Diesel
Engines. ASME 1998, ICE-Vol 31-1, 1998 Fall Technical Conference, Paper No 98ICE-117
Hosoya H., Obokata T. Effect of Nozzle Configuration on Characteristics of SteadyState Diesel Spray. SAE 930593, 1993
Hsiang L., Faeth G.M. Drop Properties After secondary Breakup. International
Journal of multiphase Flow, Vol 19, Issue 5, PP 721-735, October 1993
63
References
Huh K.Y., Lee E., Koo J. Diesel Spray atomization model considering Nozzle Exit
Turbulence Conditions. Atomization and Sprays, Vol 8, PP 453-469, 1998
Huh K.Y., Lee E., Yang O.Y. An experimental study on initial behaviour of Diesel
fuel spray characteristics. ICLASS-91 Gaithersburg, MD, USA. July 1991
Karimi E.R. High-speed photography of fuel spray and combustion event in a
production Diesel engine and combustion bumb. IMech E, 1989 page 269-281
Kennaird D.A., Crua C., Lacoste J., Heikal M.R., Gold M.R., Jackson N.S. InCylinder Penetration and Breakup of Diesel Sprays Using a Common Rail Injection
System. SAE 2002-01-1626
Koo J., Martin J. Droplet Size and velocities in a Transient Diesel Fuel Spray. SAE
900397, 1990
Lading L., Wigley G., Buchhave P. Optical Diagnostics for Flow Processes.
Proceedings of a summer school on optical diagnostics for flow processes held
September 26-October 2, 1993, Roskilde, Denmark
Laoonual Y., Yule A.J., Walmsley S.J. Internal Fluid Flow and Spray Visualization
for a Large Scale Valve Covered Orifice (VCO) Injector. ILASS-Europe 2001, Zurich
2-6 September 2001
Le Gal P., Farrugia N., Greenhalgh D.A. Laser Sheet Dropsizing of dense Sprays.
Optics and Laser Technology, vol 31, PP 75-83, 1999
Lee C.S., Park S.W. An Experimental and Numerical Study on Fuel Atomization
Characteristics of High-Pressure Diesel Injection Sprays. Elsevier, Fuel, Page 24172423, 2002
Lee T., Reitz R.D. The Effect of Split Injection and Swirl on a HSDI Diesel Engine
Equipped With a Common Rail Injection system. SAE 2003-01-0349, 2003
Lee S.W., Tanaka D., Kusaka J., Daisho Y. Effects of Diesel Fuel Characteristics
on Spray and Combustion in a Diesel Engines. Elsevier, JSAE Review 23, Page 407414, 2002
Lefebvre A.H. Atomisation and Spray. Combustion: An International Series, Taylor
and Francis Publishers, 1989
Lucas, Rate of injection manual; High T/P rate gauge manual; issued 11/9/95.
Mohammadi A., Miwa K., Ishiyama T., Abe M. Investigation of Droplets and
Ambient Gas Interaction in a Diesel Spray Using a Nano-Spark Photography Method.
SAE 981073, 1998
Morgan R., Wray J., Kennaird D.A., Crua C., Heikal M.R. The Influence of
Injector Parameters on the Formation and Break-Up of a Diesel Spray. SAE 2001-010529, 2001
64
References
Mugele R., Evans H.D. Droplet Size Distribution in Spray. Ind. Eng. Chem.,Vol.43,
No. 6, pp. 1317-1324, 1951
Naber D., Siebers D.L. Effects of Gas Density and vaporization on Penetration and
Dispensation of diesel Spray. SAE 960034, 1996
Nurick W.H. Orifice Cavitation and its Effects on Spray Mixing. Journal of Fluids.
Engineering, December 1976
Pierpont D.A., Montgomery D.T., Reitz R D. Reducing Particle and NOx Using
Multiple Injection and EGR in a D.I. Diesel. SAE 950217, 1995
Pitcher G., Wigley G. Application of Phase Doppler Anemometry to Combustion of
Diesel Fuel Spray Measurements Technique, Precision and Data Interpretation. 6th
workshop on two-phase flow prediction, March 30-April 2, 1992 at Lehrstuhl fur
Stromungsmechanik Cauerstr. 4, 8520 Erlangen FGR
Reitz R.D., Diwakar R. Structure of High-Pressure Fuel sprays. SAE 870598, 1987
Sazhin S.S., Feng G., Heikal M.R. A model for fuel spray penetration. Elsevier, Fuel
80 (2001)2171-2180
Sazhin S.S., Crua C., Kennaird D., Heikal M.R. The initial stage of fuel spray
penetration. Elsevier, Fuel 82 (2003)875-885
Schmidt D.P., Corradini M.L. The Internal Flow of Diesel Fuel Injector Nozzles: A
Review. INT. J. Of Engine Res, 2001 2 (1), I0034104BN
Schmidt D.P., Rutland C.J., Corradini M.L. Cavitation in Two-Dimensional
Asymmetric Nozzles. SAE 1999-01-0518, 1999
Schugger C., Renz U. Experimental Investigation of the primary Breakup zone of
High Pressure Diesel Sprays from Multi-Orifice Nozzles. 9th International conference
on liquid atomization and spray system, ICLASS 2003, July 13-17 2003
Sher E. Handbook of Air Polution from internal combustion Engines, Pollutant
formation and Control. ACADEMIC PRESS, 1998
Shimizu M., Arai M., Hiroyasu H. Measurements of Breakup length in High Speed
Jet. Bulletin of JSME, Vol 27, No 230, August 1984
Soteriou C., Andrews R., Smith M. Direct Injection Sprays and the Effect of
Cavitation and Hydraulic Flip on Atomization. SAE 950080, 1995
Sovani S.D., Chou E., Sojka P.E., Gore J.P., Eckerle W.A., Crofts J.D. High
pressure Effervescent Atomization: Effect of Ambient Pressure on Spray Cone Angle.
Elsevier, Fuel, Page 427-435, 2000
65
References
Strakey P.A., Talley D.G. Phase Doppler Measurements in Dense Sprays. Also In
ILASS (AMER)98 Proc. Carnegie Mellon Uni. Short Course On Spray Technology.
Vol 2, 1998
Stumpp G., Ricco M. Common Rail-An Attractive fuel Injection System for
Passenger Car DI Diesel Engines. SAE 960870, 1996
Su T.F., Patterson A., Reitz R.D., Farrell P V. Experimental and Numerical Studies
of High Pressure Multiple Injection Sprays. SAE 960861, 1996
Tow T.C., Pierpont D A., Reitz R.D. Reducing Particulate and NOx Emissions by
Using Multiple Injection in a Heavy Duty D.I. Diesel Engine/ SAE 940897, 1994
Tsue M., Hattori H., Saito A., Kadota T. Planer Fluorescence Technique for
Visualization of a Diesel spray. SAE 922205, 1992
Wu Z., Huang Z., Hao L. An Image-Shifting Technique Based On Gray-Scale
Classification for Particle Image Velocimetry. Optics and Lasers in Engineering, Vol
38, PP 567-575, 2002
Yule A., Akhtar P., Shrimpton J.S., Wagner T. PDA Measurements of Fuel Effects
on Atomization and Spray structure from a Diesel Engine Injector. SAE 982544, 1998
Yule A.J., Filipovic I. On the breakup times and Length of Diesel Spray.
International journal of Heat and Fluid Flow, Vol 13, No 2, June 1991
Yule A.J., Salters D.G. On the distance required to atomize diesel spray injected
from orifice-type nozzle. IMechE 1995, Vol 209
Zhao H., Ladommatos N. Optical Diagnostics for In-Cylinder Mixture Formation
Measurements in IC Engines. Prog. Energy Combustion. Sci. Vol. 24, pp, 297-336,
1998
Zhou P., Wang Y., Roskilly A.P. An Investigation of Droplet Size and Distribution
in DI Diesel Sprays Using Laser PDA and Mathematical Regression Techniques.
Trans IMarE, Vol 112, part 3/4, PP 103-113, 2000
http://www.pages.drexel.edu/~weg22/edge.html, Laplacian edge detection, accessed
on 20/07/2004
http://homepages.inf.ed.ac.uk/rbf/HIPR2/log.html,
Laplacian
edge
detection,
accessed on 23/07/2004
http://www.kraftfahrzeugtechnikheute.de/k/en/stsrt/product.jsp?mfacKey=ds_28_einspd,
Bosch
web
sight,
accessed on 10/11/2003
66
Appendices
APPENDICES
67
Appendices
Case Number
Plenum Air
Water Temp
Heated Boost
Temp (K)
(K)
371.15
373.25
3.2
371.15
373.2
372.15
373.15
5.5
371.15
373.15
366.15
373
375.15
373.25
7.5
375.15
373.15
8.5
375.15
373
10
375.15
373.25
10.5
10
295.85
373.35
11
296.15
373.05
12
296.85
373.15
13
297.45
373.25
14
297.65
373.15
15
295.85
373
15
298.25
373.35
Table A-1. Test conditions for the assessment of compression ratio and Polytropic coefficient.
A- 1
Appendices
Single Hole
3 Hole
5 Hole
Apparent Needle Lift
Apparent Needle Closure 5 Hole
Apparent Needle Closure 3 Hole
Apparent Needle Closure Single Hole
100
80
60
40
20
0
-2
10
12
14
16
Time (ms)
Figure A-1. Rail pressure traces recorded by an AVL Indiset high speed data acquisition system at 60
MPa injection pressure, 6 MPa ICP, and 50 mm3 fuelling. Broken lines indicate apparent needle full
lift \ closure. The nozzles tested were all 0.2 mm single guided VCO.
16
14
Current, 5 Hole
12
10
8
6
4
2
0
-2
Time (ms)
Figure A-2. Injection current traces recorded by an AVL Indiset high speed data acquisition system at
60 MPa injection pressure, 50 mm3 fuelling.
A- 2
Appendices
120
100
80
60
Single Hole
3 Hole
40
5 Hole
20
0
-2
10
12
14
16
Time (ms)
Figure A-3. Rail pressure traces recorded by an AVL Indiset high speed data acquisition system at 100
MPa injection pressure, 6 MPa ICP, and 50 mm3 fuelling. The nozzles tested were all 0.2 mm single
guided VCO.
16
14
Current, 5 Hole
12
10
8
6
4
2
0
-2
Time (ms)
Figure A-4. Injection current traces recorded by an AVL Indiset high speed data acquisition system at
100 MPa injection pressure, 50 mm3 fuelling.
A- 3
Appendices
160
140
120
100
80
Single Hole
60
3 Hole
40
5 Hole
20
0
-2
10
12
14
16
Time (ms)
Figure A-5. Rail pressure traces recorded by an AVL Indiset high speed data acquisition system at 140
MPa injection pressure, 6 MPa ICP, and 50 mm3 fuelling. The nozzles tested were all 0.2 mm single
guided VCO.
16
14
Current, 5 Hole
12
10
8
6
4
2
0
-2
Time (ms)
Figure A-6. Injection current traces recorded by an AVL Indiset high speed data acquisition system at
140 MPa injection pressure, 50 mm3 fuelling.
A- 4
Appendices
180
160
140
120
100
Single Hole
3 Hole
80
5 Hole
Apparent Needle Full Lift
60
20
0
-2
10
Time (ms)
12
14
16
Figure A-7. Rail pressure traces recorded by an AVL Indiset high speed data acquisition system at 160
MPa injection pressure, 6 MPa ICP, and 50 mm3 fuelling. Broken lines indicate apparent needle full
lift \ closure. The nozzles tested were all 0.2 mm single guided VCO.
16
14
12
Current, 5 Hole
10
8
6
4
2
0
-2
Time (ms)
Figure A-8. Injection current traces recorded by an AVL Indiset high speed data acquisition system at
160 MPa injection pressure, 50 mm3 fuelling.
A- 5
Appendices
40
35
30
25
20
15
10
5
0
0
0.5
1.5
2.5
Time (ms)
Figure A-9. The effect of injection pressure on spray penetration for a 0.2 mm single hole single guided
VCO nozzle, at 2 MPa ICP (ambient gas density at 12.1 kg m-3) at cold intake.
50
45
40
35
30
25
60 MPa injection at 2MPa ICP
20
15
5
0
0
0.5
1.5
2.5
Time (ms)
Figure A-10. The effect of injection pressure on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at 4 MPa ICP (ambient gas density at 24.2 kg m-3) at cold intake.
A- 6
Appendices
45
40
35
30
25
20
60 MPa injection at 2MPa ICP
15
10
0
0
0.5
1.5
2.5
Time (ms)
Figure A-11. The effect of injection pressure on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at 6 MPa ICP (ambient gas density at 36.3 kg m-3) at cold intake.
(3-hole nozzle)
50
45
40
35
30
25
20
15
10
5
0
0
0.5
1.5
Time (ms)
Figure A-12. The effect of injection pressure on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at 2 MPa ICP (ambient gas density at 12.1 kg m-3) at cold intake.
A- 7
Appendices
45
40
35
30
25
20
60 MPa injection at 6 MPa ICP
15
100 MPa injection at 6 MPa ICP
10
0
0
0.5
1.5
Time (ms)
Figure A-13. The effect of injection pressure on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at 4 MPa ICP (ambient gas density at 24.2 kg m-3) at cold intake.
45
40
35
30
25
20
60 MPa injection at 6 MPa ICP
15
10
0
0
0.5
1.5
Time (ms)
Figure A-14. The effect of injection pressure on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at 6 MPa ICP (ambient gas density at 36.3 kg m-3) at cold intake.
A- 8
Appendices
(5-hole nozzle)
50
45
40
35
30
25
20
60 MPa injection at 6 MPa ICP
15
10
5
160 MPa injection at 6 MPa ICP
0
0
0.2
0.4
0.6
0.8
1.2
1.4
Time (ms)
Figure A-15. The effect of injection pressure on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at 2 MPa ICP (ambient gas density at 12.1 kg m-3) at cold intake.
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
1.2
1.4
Time (ms)
Figure A-16. The effect of injection pressure on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at 4 MPa ICP (ambient gas density at 24.2 kg m-3) at cold intake.
A- 9
Appendices
50
45
40
35
30
25
20
60 MPa injection at 6 MPa ICP
15
10
0
0
0.2
0.4
0.6
0.8
1.2
1.4
Time (ms)
Figure A-17. The effect of injection pressure on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at 6 MPa ICP (ambient gas density at 36.3 kg m-3) at cold intake.
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
Time (ms)
2.5
Figure A-18. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at cold intake, and 60 MPa injection pressure.
A- 10
Appendices
50
45
40
35
30
25
20
15
10
5
0
0
0.5
1.5
Time (ms)
2.5
Figure A-19. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at cold intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
2.5
Time (ms)
Figure A-20. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at cold intake, and 140 MPa injection pressure.
A- 11
Appendices
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
Time (ms)
2.5
Figure A-20. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at cold intake, and 160 MPa injection pressure.
(3-hole nozzle)
50
45
40
35
30
25
20
15
10
0
0
0.5
1
Time (ms)
1.5
Figure A-21. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at cold intake, and 60 MPa injection pressure.
A- 12
Appendices
50
45
40
35
30
25
20
15
10
5
0
0
0.5
1
Time (ms)
1.5
Figure A-22. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at cold intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
Density 12.1 kg m^3 (2 MPa ICP) 140 MPa injection-Cold
15
10
0
0
0.5
1
Time (ms)
1.5
Figure A-23. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at cold intake, and 140 MPa injection pressure.
A- 13
Appendices
50
45
40
35
30
25
20
Density 12.1 kg m^3 (2 MPa ICP) 160 MPa injection-Cold
15
10
0
0
0.5
1
Time (ms)
1.5
Figure A-24. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at cold intake, and 160 MPa injection pressure.
(5-hole nozzle)
50
45
40
35
30
25
20
Density 12.1 kg m^3 (2 MPa ICP) 60 MPa injection
15
10
Density 36.3 kg m^3 (6 MPa ICP) 60 MPa injection
5
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-25. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at cold intake, and 60 MPa injection pressure.
A- 14
Appendices
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-26. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at cold intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-27. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at cold intake, and 140 MPa injection pressure.
A- 15
Appendices
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-28. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at cold intake, and 160 MPa injection pressure.
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
Time (ms)
2.5
Figure A-29. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at hot intake, and 60 MPa injection pressure.
A- 16
Appendices
50
45
40
35
30
25
20
15
Density 9.7 kg m63 (2 MPa IC) 100 MPa injection-Hot
10
0
0
0.5
1.5
Time (ms)
2.5
Figure A-30. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at hot intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
2.5
Time (ms)
Figure A-31. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at hot intake, and 140 MPa injection pressure.
A- 17
Appendices
50
45
40
35
30
25
20
15
10
0
0
0.5
1.5
2.5
Time (ms)
Figure A-32. The effect of ambient gas density on spray penetration for a 0.2 mm single hole single
guided VCO nozzle, at hot intake, and 160 MPa injection pressure.
(3-hole nozzle)
50
45
40
35
30
25
20
15
10
0
0
0.5
1
Time (ms)
1.5
Figure A-33. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at hot intake, and 60 MPa injection pressure.
A- 18
Appendices
50
45
40
35
30
25
20
Density 9.7 kg m^3 (2 MPa ICP) 100 MPa injection-Hot
15
10
5
0
0
0.5
1
Time (ms)
1.5
Figure A-34. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at hot intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
15
10
5
0
0
0.5
1
Time (ms)
1.5
Figure A-35. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at hot intake, and 140 MPa injection pressure.
A- 19
Appendices
50
45
40
35
30
25
20
Density 9.7 kg m^3 (2 MPa ICP) 160 MPa injection-Hot
15
Density 19.3 kg m^3 (4 MPa ICP) 160 MPa injection-Hot
10
Density 29 kg m^3 (6 MPa ICP) 160 MPa injection-Hot
5
0
0
0.5
1
Time (ms)
1.5
Figure A-36. The effect of ambient gas density on spray penetration for a 0.2 mm 3-hole single guided
VCO nozzle, at hot intake, and 160 MPa injection pressure.
(5-hole nozzle)
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-38. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at hot intake, and 60 MPa injection pressure.
A- 20
Appendices
50
45
40
35
30
25
20
15
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-39. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at hot intake, and 100 MPa injection pressure.
50
45
40
35
30
25
20
15
Density 9.7 kg m^3 (2 MPa ICP) 140 MPa injection-Hot
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-40. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at hot intake, and 140 MPa injection pressure.
A- 21
Appendices
50
45
40
35
30
25
20
15
Density 9.7 kg m^3 (2 MPa ICP) 160 MPa injection-Hot
10
0
0
0.2
0.4
0.6
0.8
Time (ms)
1.2
1.4
Figure A-41. The effect of ambient gas density on spray penetration for a 0.2 mm 5-hole single guided
VCO nozzle, at hot intake, and 160 MPa injection pressure.
Injection
Pressure/Incylinder pressure
(MPa)
160/2
160/4
160/6
140/2
140/4
140/6
100/2
100/4
100/6
60/2
60/4
60/6
tb
(ms)
Lb
(mm)
0.09
0.08
0.08
0.115
0.13
0.13
0.15
0.18
0.15
0.25
0.24
0.25
13
10.5
10
14.5
13
11.5
15
13.7
11.5
17
14.3
12.9
Table A-1. Breakup time and breakup length for a 0.2 mm 3 hole single guided VCO nozzle, obtained
from curve fitting equation 4.2 to experimental data at different injection and in-cylinder gas pressures.
A- 22