 

 

tuba.

7..
£41 ‘f'ff‘vﬁri'ft‘ a
,—‘\ ‘\ n

”3.1.”.

. . ‘-

'«2‘. .. .‘u
a

f‘.’
.. -r 0'3"”

'.' ' I‘o v
;,--...._u..

'. ‘ -. I: v . 1‘

w ' m .~. .13."

_':";u_ I L ' ‘ ‘ in“; :‘J
I j ’

" I

I. __.
."“ .

‘ 75' 1—‘01‘
.1 31.- -

I a

. .
‘ I r I
... . I .J‘. TJ‘J‘W’;
"p-lhﬂ‘:’~.

55.5,
Ir.

5,3,3; 3.23;; .

 

A
ivy};
s” 1

‘r
~ <1 5'" '. ‘- m1 ,1;
. .1, ,

'."'u' (C. H.
I K A r;
I

..,_,_‘

v'n ﬁnk"

.

. u- M‘, a“,
31:; :1 3-1;" - .15 J,“ Sal", ,
“I": v 1. Juan-

-., .I -.‘I. «.45

 

w. 1;
"v ._,

kzggfiz'gw

“ in (I'd
'- 433,53“

‘ 1 J “with.
4 r—
,x 7.;
5,-.7‘,_i.._‘ g. A
M ‘ ‘

~.:..I-. .
iaI.. 'VL‘ ‘l

.‘x'

-‘
Newsgr, 2-,” w
' Ian».
"l'l

.~.
.‘y'Jt’.I."";-_.- I

”V
v u». .. . _
Jon-~1. A .., .__\cw
“trl-nuﬁn-n”

" -

I
“::-v .~

'1‘ _;:--j"3;2 , :1...)

‘ . .l-u...
r. "a‘

rat.
Mac

a.“ -. .u.

" W‘V‘Vﬂ (I: ﬁ‘h'v::‘_
ii” “n.4,. !
"1‘ w “mum.“
'13..‘ ...... V“
A..." ~ .. .‘. Mg. I «WI; ', "ﬂy
~-\..‘ ...... . r < .-- :g
.‘.,..,,»-H a :1
.

.-
u”

.u 0’.

’ ”£37,.
-..‘zt.I.‘.'..rv.,_u

. ”1., .,.

n. .J. 4
'.‘ .W'.:‘f3‘,.‘.::-I

ml '.I."'

n.
“rm...

5.“:
'T-ouo

an to qu. .Z';

 

 

1: [9.3159

 

Date

0—7 639

02/22/91

                                                                                                                           

293 00911 6926

This is to certify that the
thesis entitled

HIGH PRESSURE FUEL INJECTION STUDIES
IN A MOTORED ROTARY ENGINE ASSEMBLY

presented by

Todd B. Morita

has been accepted towards fulﬁllment
of the requirements for

Masters degree inﬂechanjcaJ_Engineering

 

Major professor

MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

*‘

f LIBRARY
Ilichigau State

University
\_ A

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES Mum on or before duo due.

DATE DUE DATE DUE DATE DUE

J L_
ll:

||._____ |
____= J{___|

——7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7—“ l

MSU lo An Affirmative Action/Equal Opportunity Inotltution
emails-p1

 

 

HIGH PRESSURE FUEL INJECTION STUDIES IN A MOTORED ROTARY
ENGINE ASSEMBLY

By

Todd B. Morita

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of requirements
for the degree of

MASTERS OF SCIENCE

Department of Mechanical Engineering

1991

ABSTRACT

HIGH PRESSURE FUEL INJECTION STUDIES
IN A MOTORED ROTARY ENGINE ASSEMBLY

By

Todd B. Morita

A high speed ﬂow visualization technique was utilized to study fuel spray
development and the fuel-air mixing process in a direct fuel injection motored rotary
engine assembly. Kerosene fuel was injected through a single hole nozzle at an angle of
60° from the horizontal axis into the combustion chamber. The nozzle was installed in a
Servojet fuel injection system that was mounted in the central housing of a rotary engine.
A 40 watt copper vapor laser was synchronized with a high speed rotating prism camera
to expose motion picture ﬁlms at 5000 frames per second (fps). The fuel spray structure
was studied at injection pressures of 44 and 70 MPa, and at nominal injection durations
of 1.2 and 2.8 ms. The engine shaft speed was 2000 rpm. A sequence of successive
frames was selected from high speed ﬁlms and used to analyze the fuel spray
characteristics. The flow ﬁeld and rotor motion which had contributing effects on the
fuel spray behavior were also considered. The spray tip velocities and penetration
distances for the fuel spray were quantiﬁed by analyzing the high speed ﬁlms and were
found to be signiﬁcantly influenced by the injection pressure. The penetration distances
from the high speed ﬁlms were also compared to analytical expressions developed for
high pressure fuel injection. Blowby past the apex seals is also brieﬂy discussed. Finally,
the mixing characteristics within the compression pocket were also observed in this
thesis.

ACKNOWLEDGMENTS

At this point, I would like to express my gratitude and thanks to several people
whose guidance was extremely important to this work. First, I would like to thank my
advisor Dr. C. Somerton for his technical assistance and helpful suggestion that directed
this research to its culmination; to Dr. H. Schock and Dr. E. Scott, the other committee
members, for their invaluable time and supportive comments. I must also acknowledge
Dr. F. Hamady, whose expertise in the flow visualization techniques made this work
possible. I give thanks to my co-workers and peers for making my time spent at the MSU
Engine Research Laboratory enjoyable. I must also express my thanks to my parents
Marv and Estelle Morita for their loving support during my studies.

My last expression of thanks must be directed toward NASA Lewis Research
Center for their support of this project.

iii

TABLE OF CONTENTS

Page
LIST OF TABLES .................................................................................................. vi
LIST OF FIGURES ................................................................................................. vii
LIST OF SYMBOLS ............................................................................................... ix
CHAPTER 1 - INTRODUCTION
1.1 Problem Statement .................................................................... 1
1.2 Literature Review ..................................................................... 2
CHAPTER 2 - EXPERIMENTAL EQUIPMENT AND PROCEDURES ................ 6
2.1 Experimental Equipment ............................................................. 6
2.1.1 Fuel Injection System ....................................................... 6
2.1.2 Rotary Engine Assembly .................................................. 10
2.1.3 Flow Visualization System ............................................... 16
2.2 Experimental Procedures ............................................................ 16
2.2.1 Nozzle Tip Calibrations .................................................... 18
2.2.2 Injection Timing ............................................................... 22
2.2.3 Flow Visualization ........................................................... 25
2.2.4 Penetration Distance and Velocity Measurements ............. 26
2.2.5 Spray Cone Angle Measurements ..................................... 28
2.2.6 Analytical Penetration Distance Calculations .................... 29
2.2.7 Analytical Spray Angle Calculations ................................ 30
2.2.8 Fuel-Air Mixing Observations .......................................... 30
CHAPTER 3 -'RESULTS AND DISCUSSION ..................................................... 32
3.1 Flow Visualization for Case (i) ................................................... 32
3.1.1 Penetration Distances and Velocity Measurements ............ 32
3.1.2 Calculated Penetration Distances for Case (i) .................... 37
3.1.3 Calculated Spray Cone Angle for Case (i) ......................... 38

iv

3.1.4 Modiﬁed Penetration Distance Calculations ...................... 39

3.1.5 Quantifying The Fuel-Air Mixture After Injection ............ 44

3.2 Flow Visualization for Case (ii) .................................................. 47

3.2.1 Penetration Distances and Velocity Measurements ............ 47

3.2.2 Calculated Penetration Distances for Case (ii) ................... 51

3.2.3 Calculated Spray Cone Angle for Case (ii) ........................ 52

3.2.4 Modiﬁed Penetration Distance Calculations ...................... 52

3.2.5 Quantifying The Fuel-Air Mixture After Injection ............ 56

3.3 Flow Visualization for Case (iii) ................................................. 58

3.3.1 High Speed Film Observations .......................................... 58

3.3.2 Quantifying The Fuel-Air Mixture After Injection ............ 60

3.4 Comparing Case (i) and Case (ii) Spray Characteristics .............. 60

3.5 Nozzle Tip Verification .............................................................. 62

3.6 Blowby Observations .................................................................. 64

CHAPTER 4 - SUMMARY AND CONCLUSIONS .............................................. 66
CHAPTER 5 — RECOMMENDATIONS ................................................................. 68
LIST OF REFERENCES ....................................................................................... 69

LIST OF TABLES

Page

Table 1 Rotary engine characteristics ............................................................... 15
Table 2 Penetration and spray tip velocity parameter for injection pressure

of 70 MPa ........................................................................................... 36
Table 3 Comparison of spray characteristics ..................................................... 39
Table 4 Modiﬁed penetration distances: Hiroyasu & Arai Expression .............. 42
Table 5 Modiﬁed penetration distances: jet mixing Expression ........................ 42
Table 6 Probability of observing liquid fuel in segmented areas ....................... 47
Table 7 Penetration and spray tip velocity parameters for injection pressure

of 44 MPa ........................................................................................... 50
Table 8 Comparison of spray characteristics ..................................................... 52
Table 9 Modiﬁed penetration distances: Hiroyasu & Arai Expression .............. 54
Table 10 Modiﬁed penetration distances: Jet Mixing Expression ........................ 54
Table 11 Probability of observing liquid fuel in segmented areas ....................... 56
Table 12 Probability of observing liquid fuel in segmented areas ....................... 61
Table 13 Comparing penetration distances for Case (i) and Case (ii) .................. 62
Table 14 Spray cone angle comparison ............................................................... 63

vi

Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
Figure 18
Figure 19
Figure 20
Figure 21
Figure 22

Figure 23
Figure 24
Figure 25
Figure 26
Figure 27

LIST OF FIGURES

Page
Injector nozzle showing angle of injection ............................................... 5
SE-4D control box for injection system .................................................... 7
Hall effect system ..................................................................................... 7
Fuel injection schematic ............................................................................ 9
Walking gear for rotor assembly ............................................................... 10
Fuel injector and mounting assembly for the central housing ................... 11
Nozzle seating in rotary engine housing ................................................... 12
Fuel injector tip with modiﬁed engine housing wall ................................. 12
Schematic of lubrication system ................................................................ 13
Lubrication of crankshaft and oil capture ................................................ 13
Laser access plexiglass window in the central housing .......................... 14 '
Sapphire side window for high speed camera access ............................. 15
Nylon and steel water tubes to the motor ............................................... 16
Schematic of flow visualization system .................................................. 17
The copper vapor laser system used in study ......................................... 17
Bench stand set-up for the nozzle tip calibration tests ............................ 19
Fuel injector mounted into static engine ................................................. 19
Positioning of high speed camera and reflecting mirror for static test 20
Calibration collecting bottle attached to injector nozzle ......................... 21
Analytical balance used in calibration test .............................................. 22
Rotor positions of 90° BTDC and TDC .................................................. 24
Oscilloscope out put of encoder and hall effect meter for timing
set-up ..................................................................................................... 24
Rotor position at injection ...................................................................... 25
Camera positioned to capture injection events in motored engine ........ 26
Penetration distance calculations ..... . ...................................................... 27
Measuring the spray cone angle ............................................................ 28

Segmented combustion chamber for observations of fuel-air mixing... 31

vii

Figure 28
Figure 29
Figure 30
Figure 31
Figure 32
Figure 33
Figure 34
Figure 35
Figure 36
Figure 37
Figure 38
Figure 39
Figure 40
Figure 41
Figure 42
Figure 43
Figure 44
Figure 45
Figure 46
Figure 47
Figure 48
Figure 49
Figure 50
Figure 51

Pressure trace without fuel injection, engine speed of 2000 rpm ............ 33
Pressure trace with fuel injection and engine speed of 2000 rpm ......... 33
Successive frames of injection event at a pressure of 70 MPa .............. 34
Penetration distance for selected events and event average ................. 35
Velocity distribution for selected events and event average .................. 36
Penetration distances for experimental and calculated values ............... 37
Linear assumption for chamber pressure ............................................... 40
Linear assumption for the injection event ............................................. 41
Penetration distances for standard and modiﬁed expressions ............... 43
Velocity comparisons of original and modiﬁed expressions ................. 44
Fuel-air mixing after injection for Case (i) ............................................ 46
Successive frames of injection event at a pressure of 44 MPa ............. 48
Penetration distances of selected events ............................................... 49
Spray tip velocities of selected events .................................................. 50
Penetration distances at an injection pressure of 44 MPa ..................... 51
Penetration distances for standard and modiﬁed expressions ............... 53
Velocity comparisons of original and modiﬁed expressions ................. 55
Fuel-air mixing after injection for Case (ii) ........................................... 57
Successive frames of one injection event at a pressure of 44 MPa ...... 59
Mixing observations for Case (iii) ......................................................... 62
Penetration distances for both Case (i) and Case (ii) ............................. 62
Injection into ambient conditions using 60° nozzle ............................... 63
Injection into ambient conditions using 45° nozzle ............................... 64
Blowby seen in the engine after the injection event ............................ 65

viii

LIST OF SYMBOLS

D - spray oriﬁce diameter (m)

D' - spray oriﬁce diameter in analytical spray cone angle calculations (mm)

AP - difference between injection pressure and combustion chamber pressure (MPa)
AP1 - scenario 1 differential pressure term for analytical expressions (MPa)

AP2 - scenario 2 differential pressure term for analytical expressions (MPa)

AP3 - scenario 3 differential pressure term for analytical expressions (MPa)

AP' - difference between injection pressure and combustion chamber pressure (Pa)

S - penetration distance (mm)
S1 - penetration distance used in explaining measuring techniques (mm)

S2 - penetration distance used in explaining measuring techniques (mm)
t - time (5)

tb - break-up length time (s)

T. - air temperature at injection (K)

p, - density of air (kg/m3)

pl - density of liquid fuel (kg/m3)

pa - dynamic viscosity of air (N s/mz)

9 - Jet Mixing spray cone angle in degrees

6' - Hiroyasu and Kadota spray cone angle in radians

ix

CHAPTER 1
INTRODUCTION
1.1 Problem Statement

The rotary combustion engine (RCE) just as other types of power plants (both
traditional and non-traditional), are evaluated based upon their ability to meet industry
standards and society requirements. Before the oil crisis of the early 1970's, the rotary
engine was hailed as the new automotive engine of the future. The ability of the engine to
produce smomh power at higher operating speeds along with its compact size made it
very attractive to the automotive industry. To realize the potential of the rotary engine,
the proponents of the engine understood that some of the inherent problems associated
with the design would have to be corrected. In particular, the fuel consumption and
emissions of the Wankel design needed to be improved, so that a rotor based power plant
would be more acceptable to the domestic auto market. However, before these
improvements could be realized, the oil crisis of 1973 and higher emission standards
effectively ended the rotary engine research to which many of the domestic and foreign
auto makers had earlier committed.

It is recognized that the performance of the RCE lies somewhere between the
piston reciprocating engine and the small gas turbine. The RCE has a larger speciﬁc fuel
consumption rate than that of a piston engine. However, the RCE fuel consumption is
signiﬁcantly less than the single small gas turbine engine. One of the most attractive
characteristics of the RCE is its ability to consistently maintain its power density while
using a variety of fuels.

Though the RCE has lost much of its appeal in the automotive industry
(excluding Mazda), its potential to drive light aircraft has not been overlooked. The
stratiﬁed rotary engine (SCRE) has received considerable attention in recent years as a
means to develop more efﬁcient power from the rotary design. Some applications
include the possible replacement of large electric motors as well as an auxiliary power
unit for large aircraft. _

Economic factors also play an important role in the development of the rotary
engine. It has been demonstrated by Deere & Co. that the SCRE can develop the
predicted power density of four break horsepower per cubic inch of displacement

1

(4BHP/ci) [1]. It has been estimated by Hamady et al. [2] that a thermal efﬁciency
increase of 25% is needed before the SCRE could be successfully introduced
commercially to replace current power sources. A reasonable starting point to increase
thermal efﬁciency of an internal combustion engine is to study the combustion process.
More speciﬁcally, an enhancement of the fuel-air mixing through the use of high
pressure fuel injection might provide exceptional results. In trying to better understand
the characteristics of fuel injection process and its effects on the fuel-air mixing within
the combustion chamber, it is hoped that this investigation will provide a sound
foundation to further enhance the SCRE design.

A critical area in the development of the internal combustion engine involves
appropriate control of the fuel distribution within the combustion chamber. An
understanding of the fuel injection process is essential to evaluating and implementing
strategies such as multi-stage injection or air blast assist atomization. Barber [3] deﬁned
the requirements of a fuel injector design as the following:

"The performance requirement of the injector in conjunction
with the rest of the system is to atomize the fuel ﬁnely and to
distribute it throughout the combustion chamber so to ensure
its complete combustion within the cycle time available, at
all operating speeds and loads. This must be done without
any deleterious effects such as dribble, secondary injections,
hole choking or excessive smoke, hydrocarbon or NO
emissions. There should be no premature failure due to
stress, overheating or gas entering the nozzle during the
closing phase."

The objective of the present study is to develop an understanding of the fuel-air mixing
process in a direct injection rotary engine through the use of high speed ﬁlm techniques
and analysis. By better understanding the fuel-air mixing process, it is hoped that this
study will assist in future designs of the direct injection rotary engine.

1.2 Literature Review

The advent of high speed, high pressure fuel injection systems has led to the
consideration of a large number of injection options. Accompanying these options have
been studies concerned with the performance of the injection system, both in and out of
the engine. Out-of-engine studies are those studies that are conducted in quiescent engine
situations or sometimes in ambient conditions. In engine studies indicate that actual

testing was conducted in the engine. The out of engine studies primarily focus on the
development of visualization techniques and the evaluation of injector performance.
Savey et al. [4] used stroboscopic ﬂash photography to visualize the injection
characteristics of the Servojet unit injector operating with water injection into quiescent
air. The photographic evidence used in conjunction with dynamic laser attenuation
measurements allowed the determination of spray penetration distance and spray cone
angle. A shadow photographic technique was used by Kato et al. [5] for the analysis of
non-evaporating fuel spray characteristics. The authors used a high pressure injection
system which they developed in cooperation with Diesel Kiki Co., Ltd. They injected
fuel into a high pressure, nitrogen-ﬁlled vessel and recorded fuel spray growth with a
high speed camera. Spray droplet diameter and equivalence ratio were analyzed with
shadow photographs. Katsura et a1. [6] photographed the impingement of a diesel spray
on a ﬂat wall in a high pressure chamber with both transmitted and scattered light. The
effect of ambient density was investigated and it was found that the growth of the spray
radius decreases with increasing ambient density. Minami et al. [7] used high speed
shadowed photography, holography and image analysis to evaluate the effect of injection
pressure and injection rate upon the characteristics of non-evaporating fuel spray. The
synchronization of a laser sheet with a high speed camera was ﬁrst employed, in this
application, by Hamady et al. [2]. These experiments were conducted by injecting
kerosene into atmospheric air and several event to event variations in the injection
process were observed that had not previously been noted.

Many of the out-of—engine studies used pressurized chambers with air flow to
simulate in-engine situations. Hiroyasu and Nishida [8,9] used high speed Schlieren
photographs to observe Spray vaporization and analyze the structures of sprays in diesel
engines. Varying quiescent chamber pressures and air densities, simultaneous droplet
sizes and velocities of diesel fuel spray were studied using still photography by Koo et al.
[10]. From the still photographs, frequency of surface waves on the liquid exiting from
the nozzle were characterized. A repetitive micro-flash used in conjunction with a drum
camera allowed Yoshikawa et al. [11] to observe spray behavior. They also conducted
in-engine studies and utilized emission measurements to optimize the injection
conditions. Heinze et a1. [12] performed quantitative, ensemble averaged, two-
dimensional fuel-air ratio maps in an injection jet by analyzing a series of single point
Raman measurements. By injecting pure n-heptane into a special high pressure and
temperature chamber, nearly 100,000 measurements were taken. From the data points a
computer ﬁlm was produced that illustrated the propagation of the inner spray, spray tip
and their influence on the fuel-air mixing. Spatial and temporal characteristics of diesel

spray injected into ambient conditions through a multi-hole nozzle was studied by
Arcoumanis et al. [13] using laser techniques. Results showed that the penetration and
spray tip velocities increased as the frequency of injection increased, while at the same
time, the injection volume was held constant.

Three studies are known to have performed visualization of the injection process
within a ﬁring engine. Werlberger et al. [14] used the endoscopic high speed combuStion
photography technique to study wall jet development, the role of swirl and the effect of
pilot injection on mixture formation and combustion. Shimada et al. [15] used direct
photography to study the non-vaporization part of the spray while at the same time, the
vaporized part was detected by means of shadow photography. In conjunction with the
fuel-air mixing, a high speed camera was utilized to study the actual combustion in a
direct injection diesel cylinder. High speed shadowed methods were utilized to observe
the combustion process in the cylinder in the study done by Minami [7].

Three-dimensional computational fluid dynamic simulations have been developed
to characterize ﬂow ﬁeld, fuel-air mixing and combustion phenomena in the RCE.
Chemically reactive flow, spray and combustion models were considered by Raju and
Willis [16,17] in a stratiﬁed charge rotary engine. Preliminary results of pressure,
temperature and velocity distributions as a function of crank angle were presented in
addition to fuel concentration and droplet trajectories. Numerical analysis presented by
Abraham et al. [18,19] of pressure measurements and fuel-air mixing contributed to
speciﬁc design changes and recommendations to optimize fuel economy and fuel injector
locations in the engine.

Using a two-dimensional numerical approach Shih et al. [20] have developed a
grid generation technique to study the interaction of fuel injected into the combustion
chamber with the flow ﬁeld and the fuel-air mixing during compression stroke. Their
results revealed valuable information on the flow patterns and the fuel-air mixing. Later
their effort was extended to include three-dimensional calculations in Steinthorsson et al.
[21] and Li et al. [22] by modifying the two dimensional code.

It is important to note that there have been no experimental studies concerned
with the in-engine visualization of the fuel injection process within the rotary engine. In
the present study, the laser sheet-high speed photography visualization technique of
Hamady et al. [23] was employed. Three primary parameters affecting the injection
process and subsequent fuel-air mixing have been identiﬁed: injection pressure, injection
timing and injection angle. All three parameters are considered in the present work. The
injection pressure is controlled by the rail pressure on the BKM Servojet Injection
System. A Hall effect meter was used to set the injection timing with respect to the

crank shaft and rotor position. The injection angle is the angle between the injection jet
and the axis of the injector assembly shown in Figure 1. Special nozzle tips have been
manufactured to achieve injection angles of 60° ,45° and 30°. The 60° nozzle used in
this study had a nozzle length to diameter ratio (ID) of 4, as seen in Figure 1. This
paper continues with a thorough description of the experimental apparatus and procedure.

The objectives for this study were to verify that high speed flow visualization
techniques could be used to observe the injection event. Second, observable changes in
the fuel injection characteristics when the operating parameters were varied was desired.
Third, the feasibility of the angled injection used to direct the spray up into the
compression chamber was to be determined. Finally, comparing experimental findings to
quiescent and theoretical expressions was sought after. The listed objectives will be
addressed in the flow visualization results. The study concludes with a summary and
some recommendations for future research directions.

 

 

 

 

 

 

Figure 1 Injector nozzle showing angle of injection

CHAPTER 2

EXPERIMENTAL EQUIPMENT AND PROCEDURES
2.1 Experimental Equipment

In the following section, the experimental equipment used in this study will be
presented. The fuel injection system as well as the rotary assembly and flow visualization
system will be discussed in detail.

2.1.1 Fuel Injection System
The Servojet Fuel Injection System consists of four (4) basic elements:

1. SE-4D electronic controller

2. Hall Effect Sensory System

3. Electronically regulated hydraulic system
4. Electronic Unit Injectors

The SE-4D controller box provides the capability of operating the fuel injection
system in the in-engine mode or bench test mode. The SE-4D allows for easy adjustment
of the injection timing, solenoid energize time, engine speed or calibration frequency and
rail pressure.

The operating mode for the fuel injection system is controlled by a variable
switch on the SE-4D. When running bench tests or out of engine studies, the calibration
mode is selected. Within this mode, the energize time (or the amount of time the injector
solenoid is on), ﬁring frequency and rail pressure can be manually adjusted. When ﬁring
the injector in a running engine, the load mode is selected. Under load mode, the ﬁring
frequency of the injector is determined by the crankshaft speed with all other parameters
that were accessible under the calibration mode still remaining accessible. Figure 2
shows the SE—4D along with the adjustable parameter switches. In order to utilize the
engine speed to ﬁre the injector, a Hall Effect Sensory System (HESS) must be utilized.

The HESS is comprised of a Hall Effect Mask (HEM) and a Hall Sensor Pickup
(HSP). The HEM rotates at crankshaft speed that utilizes one modiﬁed metal tab to give
a rotor position reference. The HSP is positioned to allow the tab in the HEM to pass
through its magnetic sensor. Once the metal tab passes through the HSP, an electrical

6

.. IIIIEI

700‘

 

Figure 2 SE-4D control box for injection system

 

Figure 3 Hall effect system

pulse is created and sent to the SE-4D microprocessor. The signal is analyzed and the
engine speed and injection firing rate is determined. Figure 3 shows the HEM and HSP
before installation.

Through the hydraulic system, fuel is supplied to the injector and excessive fuel is
bled off and returned to the fuel reservoir. Rail pressures are controlled by an electronic
pressure regulator which includes a safety relief valve, accumulator and return check
valve. Fuel delivery is controlled by the regulation of the rail pressure. It is the
communication between the electronic controller and a solenoid valve which positions a
spool to achieve a programmed or manually selected rail pressure which makes the
regulation of the rail pressure possible.

Some of the vital components of the hydraulic system will now be discussed. The
transfer pressure pump operates in a range of 1020 psi, and it must be able to supply the
maximum quantity of fuel utilized by the engine. The heat exchanger regulates the fuel
temperature before it reaches the injector, and the rail pressure pump has a maximum
Operating pressure of 1500 psi. Figure 4 shows a simple schematic drawing of the
Servojet Fuel Injection System.

The electronic unit injector consists of four elements:

1. Solenoid valve

2. Pressure intensiﬁer
3. Accumulator

4. Nozzle tip

The ﬁrst stage solenoid valve is an interface between the injector and electronic
controller. When the solenoid is energized, a high pressure plunger in the fuel injector
intensiﬁer is actuated which supplies the required amount of fuel into the accumulator.
The intensiﬁer is capable of raising the effective pressure in the nozzle body to 15 times
that of the rail pressure . When the solenoid is de-energized, the low pressure piston and
high pressure plunger retract, causing a pressure differential on the needle valve. During
the pressure imbalance, the needle is lifted which opens the nozzle tip oriﬁce to the fuel
stored in the accumulator. Injection will continue until such a time when the pressure
difference between the accumulator and the needle valve closing pressure no longer
exists or in other words, the pressure forces equilibrate. The needle valve closing
pressure is predetermined by a spring force.

Special nozzle tips were manufactured by BKM Inc. for this study. Standard
nozzle tip (tips that inject in a direction along the central axis) were modiﬁed to inject at

 

 

 

 

J [1:

ﬁ,_
_ Rail
Fuel ' Transler Heat
8 l Pum Exchanger Pressure
U
991/ p A Pump

Water In

 

 

 

 

 

 

 

 

 

V Water DUI a;

A
‘

Hall Ettect

Sensor

 

 

 

 

 

Controller "—
Box

 

 

 

ll

 

 

 

 

 

 

 

ll

 

 

 

Figure 4 Fuel injection schematic

 

10

angles of 30°, 45° and 60° from the central axis of the injector. The injector oriﬁce hole
for all three nozzles has a diameter of 0.217 mm leading to a nozzle length to diameter
ration (L/D) of 4. All three nozzle tips were tested under static conditions (the engine
was not motored) to verify their feasibility in this study. From these results, only the 60°
nozzle was utilized in this study. Further details of the veriﬁcations of the modiﬁed
nozzles are discussed in the Experimental Procedures (Section 2.2).

2.1.2 Rotary Engine Assembly

The basic operation of the RCE is fundamentally different than any other internal
combustion reciprocating engine. Power is transmitted to the crankshaft through torque
which is produced by the eccentricity of the rotor as the rotor gear revolves around the
stationary gear of the crankshaft. The position of these gears in the rotor engine can be
seen in Figure 5. The rotary engine is characterized as a high power density engine. This
attribute of the rotary engine is due to its unique method of operation. In the rotary
engine, the intake, compression, combustion and exhaust cycles occur concurrently
during the operation of the engine. The tricoidal shape of the engine housing along with
the three-lobed rotor form three distinct chambers where these four

 

 

O ,

Figure 5 Walking gear for rotor assembly

11

cycles occur. From this conﬁguration, the rotor engine is able to obtain three power
strokes for every one revolution of the rotor.

The rotary engine used in this study was a single rotor peripherally ported custom
fabricated assembly based on a Mazda 12A geometry [24]. Figure 6 shows the fuel
injector and the mounting assembly used in the central housing at the modiﬁed upper
spark plug hole. The schematic drawing in Figure 7 illustrates the necessary machining
that was done to the inside housing wall to accommodate the angled injection of the
nozzle. In Figure 8, a close up of the machined groove and injector tip can be seen. This
assembly has been extensively updated during the past year. This change has resulted in a
demonstrated increase in shaft speed up to 3400 rpm. Two major modiﬁcations have
allowed for this signiﬁcant increase in the engine operating speed.

One of the major modiﬁcations to the rotary assembly was a redesign of the
lubrication system. Filtered oil is supplied by a gear pump to the rotor assembly at a
regulated pressure range of 174-279 kPa (25-40 psi). Oil is then returned to a reservoir
by means of a vacuum pump which is regulated by a flow control valve operating a 41
kPa (12" Hg). This is a closed system that originates and terminates at the oil reservoir.
Two schematic diagrams of the lubrication system can be seen in Figure 9 and Figure 10.

 

Figure 6 Fuel injector and mounting assembly for the central housing

l2

 

 

 

 

 

Figure 8 Fuel injector tip with modiﬁed engine housing wall

13

 

 

 

, - , Oil return
To main I + Engine
bearings \ , bearings / ll /Vacuum pump
/ Filter
[1 A/

\mow control

 

Filter

 

Electric mOtor

 

 

 

    
  

Oil resevoir

 

\ Pressure relief f-fi j
Pump

F

 

Figure 9 Schematic of lubrication system

 

 

/ on in\

Rotor Crank shaft

 

 

 

 
   
 

%

 

 

 

 

 

 

 

    
 

%% \ Lip seal

Oil catch

Needle valve -—>
1\ Oil retur/

 

 

Bearings

 

 

 

Figure 10 Lubrication of crankshaft and oil capture

 

 

14

The second major modiﬁcation was to the side housings. A thin layer of chrome oxide
was applied to the inside housing walls. The chrome oxide provides a smoother interface
between the inside housing walls and the rotor side seals which allows the engine to run
at higher speeds.

Production engine two-piece apex, side and oil seals were used in this work. More
details on the seals and their installation will be discussed later in this section. Optical
access was provided by custom fabricated sapphire windows that were mounted in the
end housing. A plexiglass window was ﬁtted in the top of the central housing chamber to
allow for the access of the laser light needed for flow visualization. The plexiglass and
sapphire windows are shown in Figure 11 and Figure 12.

One of the difﬁculties of the rotary engine is the sealing issue. In order for the
engine to operate correctly, there must be minimal leakage between the three chambers.
The chambers are isolated by using six side seals, three on each side of the rotor that ride
in spring-loaded grooves sealing the area between the side housing and the side of the
rotor. Apex seals are located in the apex of the three lobes. As in the case of the side seal,
the apex seals also ride in spring-loaded groves. The purpose of these seals are to prevent
interaction between the chambers at the rotor face and central housing along the lobe
apex where the two rotor ﬂanks meet. Finally, button seals, transition pieces that provide

 

Figure 11 Laser access plexiglass window in the central housing

 

Figure 12 Sapphire side window for high speed camera access

sealing between the side seals and the apex seals, also ride in spring-loaded groves.

The engine was powered using a 7.5 kW (10.0 hp) constant torque electric motor.
The motor is cooled by water delivered to the engine by nylon tubing attached to a water
facet. The water is delivered from the nylon tubing to the engine by steel tubing as shown
in Figure 13.

The RCE test rig used in this study was a peripherally ported Mazda 12A central
housing. Characteristics of the engine are given in Table 1.

 

 

l 1 R n in ti i
Major axis (y) 120 mm
Minor axis (x) 90 mm
Generating radius (R) 105 mm
Eccentricity (e) 15 mm
Intake port opening (IPO) 267°
Intake port closing (IPC) 350°
Exhaust port opening (EPO) 12°
Exhaust port closing (EPC) 66°
Engine speed 3400 rpm

 

 

 

16

 

Figure 13 Nylon and steel water tubes to the motor

2.1.3 Flow Visualization System

The high speed flow visualization system used in this study is shown
schematically in Figure 14. It consists of a 40 watt copper vapor laser (CVL) Model 451,
a high speed camera Nac E-lO/EE, mirrors and cylindrical lenses. The CVL is a gas
discharge device of 40 watts of average output power. It emits 4000-6000 pulses per
second (pps) with a pulse energy of 8 ml in the green-yellow visible spectrum and a
pulse duration of 30 ns. Figure 15 shows the CVL system used in this study. The camera
is equipped with a trigger pulse generator and. an optical pickup that triggered the CVL at
5 kHz for this study. The triggering system also made it possible to synchronize light
pulses from the CVL with the Nac high speed camera.

2.2 Experimental Procedures

In this section the procedures utilized to perform this study will be discussed. The
nozzle tip calibration, injection timing and flow visualization procedures will be
introduced in the ﬁrst three subsections. The techniques used to measure the penetration
distance and spray cone angle are discussed in the next two subsections. Finally, in the

l7

 

 

 

 

 

 

Figure 14 Schematic of ﬂow visualization system

 

 

 

Figure 15 The copper vapor laser system used in study

 

18

last three subsections, the analytical expressions used to predict spray characteristics and
the fuel-air mixing observations are discussed.

2.2.1 Nozzle Tip Calibrations

Before the special nozzle tips could be utilized in this study, three tests had to be
conducted to verify that the nozzles were functioning properly. First, the nozzles had to
be tested to verify that were capable of operating without extensive nozzle leakage
(failure of the nozzle needle to seat properly in the sack chamber). Second, the angle of
injection had to be examined for satisfactory orientation within the chamber. It was very
important that the fuel spray be directed into the chamber with minimal obstruction
(from the rotor and housing wall) over the largest possible crank angle range. Lastly, the
fuel delivered per injection was important information needed for possible future
reference of the fuel-air ratio characteristics of the test conﬁguration. Information related
to the standard 0.217 mm nozzle was available; however, since these angled nozzles were
unique to this study, no calibration information was available on their delivery
characteristics.

The ﬁrst test captured on high speed 35 mm still ﬁlm was the fuel injection event
of the three nozzles. From these stills, it was possible to justify the nozzles performance
in terms of a consistent fuel jet formation. Nozzles that leak tend to form inconsistent jet
sprays and large fuel droplet which effects the performance of an engine. A 4 watt argon—
ion laser manufactured by Coherent was utilized to provide lighting for a Nikon 35 mm
camera. The laser was mounted on a rigid traverse table (model 9900-1). The laser and
traverse table are actually part of a laser Doppler velocimeter system; however, only the
illuminating power of the laser was of interest in this test. The fuel injection unit was
secured to a bench stand. A series of mirrors were used to direct the laser beam to the
bench stand where the laser beam was spread into a sheet of light by cylindrical glass
rod. Figure 16 shows the experimental set up for the nozzle tip spray formation test. The
camera was triggered externally by the SE-4D control box. Leads were taken from the
SE-4D injector on-off switch and connected to the cameras remote adapter. When the
injector on switch was engaged, the camera was automatically triggered to take a picture.

The second test captured the fuel jet spray within a modiﬁed rotary engine on
high speed ﬁlm. Special plexiglass side housings were machined which allowed the high
speed camera to access the entire chamber where the fuel injection occurred. For this
particular test, the engine was not motored, and the rotor was position to a crank angle of
approximately 70° before top dead center (BTDC). A plexiglass window was installed in

19

 

Figure 16 Bench stand set-up for the nozzle tip calibration tests

' "’33., ..,_

 

Figure 17 Fuel injector mounted into static engine

T/yvt/ 740/ 761 (7’92/
/€’ac«<, 74/4 ﬂag 470/ 401/446. '9

 

Figure 18 Positioning of high speed camera and reﬂecting mirror for static test

the central housing of the engine (much like the housing described in the earlier section)
that allowed the laser beam to access the chamber for illumination purposes. The test
engine along with the mounted fuel injector in the modiﬁed upper spark plug hole can be
seen in Figure 17. The Nac E-lO/EE high speed camera along with the 40 watt CVL
were used to conduct this test and can be seen in Figure 18. The frame speed of the
camera was 5000 fps and the CVL pulsed at 5 kHz. The triggering system made it
possible to synchronize light pulses from the CVL to the camera. The operating mode on
the SE-4D control box was set to the calibration mode. This calibration mode was chosen
because the engine was not motored and the control of the ﬁring frequency was desired.
The ﬁring frequency was set at 20 Hz, while the energized time and rail pressure
parameters were set at 10 ms and 650 psi respectfully. After testing the 60° nozzle and
reviewing the high speed ﬁlm, further tests of the 45° and 30° were not pursued. The
results and details of this test are discussed in a later section.

In the third test of the calibration procedure, an attempt was made to calculate the
injectors performance using the 60° nozzle tip. A test matrix was constructed to guide
this test. The rail pressure was varied using settings of 500, 700, 900, 1000, 1100, and
1200 psi. The frequency of injection was varied between 20, 30 and 50 Hz. Finally, the
energize time was varied between 9, 10, 11, 12, 13 and 14 ms. These parameters were

21

chosen to mirror the test parameters of the BKM Inc [25]. calibration study of the
standard 0.217 mm nozzle tip. Special nozzle adapters were constructed to allow easy
collection of the injected fuel by a plastic container. The nozzle tip adapter was angled to
60° to minimize the restriction of fuel during the collection process. The fuel injector
was mounted in the-bench stand rotated 180° from the mounting position it assumes in
the engine. The purpose of this mounting procedure was to ease the fuel gathering
process by utilizing the gravitational force. Nylon tubing transferred the fuel from the
adapter to the plastic container. The fuel injector and collection system can be seen in
Figure 19.

An analytical balance, accurate to a 0.1 mg, was used to measure the plastic
container and the collected fuel after each run and is shown in Figure 20. A total number
of injections (nominally 2000) were weighed and recorded after each run. When the total
weight of the collecting container and fuel approached the maximum rated weight limit
(150 gms), the fuel was returned to the fuel reservoir. The plastic collecting bottle was
then weighed again and the test was continued. The mass of fuel utilized per injection
was simply calculated by taking the mass of fuel collected during the runs, and dividing
it by the recorded number of injections.

 

Figure 19 Calibration collecting bottle attached to injector nozzle

22

 

Figure 20 Analytical balance used in calibration test

2.2.2 Injection Timing

The importance of accurately setting the injection timing for the high speed ﬁlms
used in this study can not be overly emphasized. Proper timing settings were needed to
ensure that a comprehensive study of the fuel injection characteristics could be
completed. Many times in the engineering ﬁeld, the techniques used to accomplish the
work is new and unfamiliar to the researcher. The cost of engineering research is very
expensive and progress is often measured against time. In this study, the production of
one high speed ﬁlm was an expensive process. Mistakes in the timing of the fuel
injection could have resulted in ﬁlms that were useless to this study. Injection events that
occurred too early in the compression cycle might have yielded results in which the fuel
did not have the proper means to vaporize. Firing the injector too late in the compression
cycle might have resulted in a total blockage of the injection event by the rotor.

The timing procedure for synchronizing the injection event with the rotor position
utilized both an angular encoder system and the injector Hall effect system. The
following steps contain the procedure to setting up the injection timing.

23

The angular velocity is determined from the test rpm that will be used (typically in
units of deg/ms).

From the BKM injector documentation [25], the injection delay time can be
determined. The delay time is a function of the rail pressure.

Multiplying the injection delay time by the angular velocity found in step 1 will
give the angle off-set (the angle adjustment to the rotor to account for the injection
delay time).

One apex of the rotor should be aligned with the center window frame screw of the
housing. This is 90° BTDC for the crankshaft position and is relatively easy to
locate. For convenience, the apex position of TDC is referenced to 0°. The 90°
reference angle assumes that the desired crank angle for injection is less than 90°.
If for example the desired crank angle of injection is greater than 90°, then a larger
crank angle for reference is suggested. The reasoning behind the crank angle
reference selection becomes clear in later steps. Figure 21a-21b illustrate the 90°
BTDC reference angle and the TDC position of the rotor respectively.

The rotor position during the injection process must be determined at or before this
step. The rotor position is selected to the test speciﬁcations. To account for the
injection delay time inherent to the mechanics of the fuel injector, the off-set angle
is added to the crank angle of injection.

Once the ﬁnal rotor position (rotor position for injection plus the off-set angle) is
determined, subtract this from the 90° reference angle and advance the crankshaft
toward TDC by this result.

Adjust the encoder shaft while at the same time monitoring the electronic output of
the encoder with a volt meter. The leads of the volt meter must be attached to the
encoder channel that outputs one pulse per revolution of the encoder. When the
encoder responds with a pulse, the encoder shaft can then be coupled to the rotor
crankshaft.

The final step is to synchronize the Hall effect censoring system to the encoder.
The encoder and injector logic signal (on the SE-4D control box) are monitored
with an oscilloscope. Note, the engine must be running at the same speed that the
angular velocity was calculated at. The injector logic signal is a 5 volt square
wave pulse transmitted to the injector actuator by the SE-4D. This 5 volt square
wave corresponds to the energize time of the injector. The voltage spike from the
encoder must line up with the trailing edge of the 5 volt square wave. To
accomplish synchronization, the Hall effect mask is adjusted until both signals
coincide. Figure 22 illustrates the correct orientation of both the encoder and
injector logic signal for synchronization.

24

 

 

 

 

 

 

 

 

 

 

(a) (b)

 

 

Figure 21 Rotor positions of 90° BTDC and TDC

 

Channel 1

Encoder Pulse

   

Solenoid 0n

 

Channel 2
Beginning of - / - Beginnlng of
Energize Time Fuel Injection

 

 

 

Figure 22 Oscillosc0pe output of encoder and hall effect meter for timing set-up

25

2.2.3 Flow Visualization

The CVL was conﬁgured to pulse at SkHz with a power range of 35-38 watts.
The pulse energy was 8 m] in the green-yellow visible spectrum. The pulse duration was
30 ns. The Nac high speed camera was Operated at 5000 frames per second (fps). Two

hundred feet of high speed negative ﬁlm was utilized for each case. Three different cases
were considered for this study:

Case (i) Injection Pressure of 70 MPa (11,875 psi), Injection Tinting of 70° BTDC
Case (ii) Injection Pressure of 44 MPa (6352 psi), Injection Timing of 65° BTDC

Case (iii) Injection Pressure of 44 MPa, Injection Timing of 55° BTDC

Each case was run with an engine speed of 2000 rpms. At this speed, the crank shaft and
the rotor rotational velocities were approximately 12°/ms and 4°/ms. Figure 23 illustrates
the rotor position at injection for each case.

The beam from the CVL was converted to a light sheet by cylindrical lenses and
directed to the engine housing through a series of highly polished mirrors. Once the
proper orientation of the laser beam was obtained, the synchronization between the

 

 

 

 

 

Figure 23 Rotor position at injection

26

high speed camera and the CVL was checked. By reﬂecting some of the laser light to the
timing gears of the Nac camera while running at speed, the action of the timing tabs
could be observed. If the tabs appeared to align with the tinting mark during the
synchronized test, then the synchronization between the camera and CVL system was
successful. At this point, the Nac camera was positioned and focused to capture the
injection events through the sapphire window. In Figure 24, the camera, engine housing
and the ﬁnal reﬂecting mirror are shown. After the injection system reached the proper
rail pressure, the injector was ﬁred while at the same time the camera was started.

2.2.4 Penetration Distance and Velocity Measurements

In processing the high speed ﬁlms to obtain fuel spray penetration distances, a
Nac DF-16C analysis projector along with a caliper measuring instrument accurate to 25
um were utilized. In the ﬁlms, there existed areas of well deﬁned fuel distribution,
surrounded by a cloud of scattered mist. The leading portion, or apex, of the well deﬁned
fuel distribution was used as the datum point. The distance from this point to the tip of

 

Figure 24 Camera positioned to capture injection events in motored engine

27

 

Figure 25 Penetration distance calculations

28

the injector was carefully measured using the caliper for each frame of the injection
event. Velocity measurements were obtained directly from the penetration distances by
simply dividing the penetration distance by the frame time. A ﬁlm speed of 5000 fps
corresponds to 0.2 ms per frame.The procedure for determining the penetration distance
can be seen in Figure 25a-25b. Penetration distance and velocity plots verses time were
constructed from the data obtained.

2.2.5 Spray Cone Angle Measurements

The spray cone angle measurements also utilized the Nac DF-16C analysis
projector and caliper measuring device along with, a standard protractor. The spray cone
angle is deﬁned as the angle formed by two straight lines drawn from the injector tip to
the outer edge of the spray distribution at a distance of 60xD', where D' is the spray
oriﬁce diameter [26]. In Figure 26, the technique of determining the spray cone angle
from the high speed ﬁlm is illustrated.

 

Figure 26 Measuring the spray cone angle

29
2.2.6 Analytical Penetration Distance Calculations

Two expressions were utilized in predicting the penetration distances of the
injection spray for the three cases. The ﬁrst expression used was the jet mixing theory
[27]. The penetration distance were calculated using the following theoretical expression:

S = 3. 01[(AP lpl )0"s D* U” (295/T. )0'25 (1)

where AP is the difference between the injection pressure and chamber pressure. The
density of the air in the chamber at injection is p.. The injection oriﬁce diameter is
deﬁned as D. Time is given by t. Finally, T. is the air temperature inside the chamber at
injection.

A previous fuel injection study by Hiroyasu and Arai [8] utilized non-visual
means to quantify the structure of fuel sprays injected into quiescent air. Quiescent spray
tip penetration distances were calculated using the following expression [8]:

For 0 < t < tb

s = 0.39(2AP/ p,)°-5*t (2)
For t> tb

S = 2.95(AP/ p.)°'25(D*t)0'5 (3)

where tb is deﬁned as the break-up length time and is given by

tb = 28. 65[(p, *D) / (p, .. AP)°"] (4)

The parameter pl is the density of the liquid fuel.

The theoretical and quiescent penetration distances were calculated and plotted
verses time. Comparisons between the analytical calculations and the experimental
results were conducted and are discussed later in the Results section of this paper.

30

2.2.7 Analytical Spray Angle Calculations

As in the case of the previous section, the analytical spray angle calculations
utilized two different expressions. Theoretical spray angles were calculated [27] using
the expression

tan6=0.13[1+p./pl] (5)

The second expression utilized to predict spray cone angles was developed by Hiroyasu
and Kadota [28] and is given by

6’ = 0.025[(pa*AP’*D2 / #3] (6)

where 9’ is in radians and AP’ is the difference between injection pressure and chamber
pressure in units of MPa.

2.2.8 Fuel-Air Mixing Observations

An attempt to quantify the fuel-air mixing after the termination of the fuel
injection event within the combustion chamber was made for all three high speed ﬁlms.
The ultimate goal in this procedure was to obtain some insight on how the fuel-air
mixture behaved during the compression process once the injection event had ceased.
The compression chamber was segmented into 'areas' of observation to make the process
easier. A total of six areas were used to describe the fuel-air mixing concentration. A
qualifier of 1 was given to the area if it appeared to contain any fuel, while an area was
labeled 0 if it appeared to absent of fuel. Note that there was no attempt to designate
differences in fuel concentration within an area or between areas. Areas were lettered 'A'
through 'F' where 'A' was closest to the rotor apex and 'F' was located just above the
injector hole. Figure 27 illustrates the segmented combustion chamber along with the
multiple silhouettes of the rotor edge (designated as rotor advancement in ﬁgure). The
multiple silhouettes of the rotor edge represent the rotor advancement between fuel-air
concentration measurements.

For Case (i) and Case (ii), the observation time covers a period of 2.6 ms with
ﬁve different intervals at 0.0, 0.8, 1.4, 2.0 and 2.6 ms. For Case (iii), the observation
time lasted only 1.4 ms because of the late injection timing. A total of 10 injection events

31

were observed for each case. Once the areas were assigned a zero (0) or a one (1), the
probability of ﬁnding fuel in each area at each observation time was calculated.
Probabilities were calculated by taking the total fuel observations (intervals assigned 1)
and dividing by ten (10). Results for this particular part of the study are discussed in a
later section.

 

    
  
 

Engine Housing

 

Rotor Area Segment

 

Rotor Advancement

 

 

 

 

 

 

 

Figure 27 Segmented combustion chamber for observations of fuel-air mixing

CHAPTER 3
RESULTS AND DISCUSSION
3.1 Flow Visualization for Case (i)

In the following sections, the measured and analytical results for the case having
an injection pressure of 70 MPa at a timing of 70° BTDC will be discussed.

3.1.1 Penetration Distances and Velocity Measurements

This section presents fuel spray characterization of kerosene fuel injected through
a single hole nozzle having a L/D ratio equal to 4 and a spray oriﬁce diameter of 0.217
mm. Fuel was injected into the combustion chamber toward the end of the compression
stroke of a motored rotary engine at a crankshaft speed of 2000 rpm. At this speed, the
crank shaft and the rotor rotational velocities were approximately 12°/ms and 4°/ms.

The engine pressure at injection was approximately 0.3 MPa. Pressure traces with
and without fuel injection are shown in Figures 28 and 29. Figure 28 shows a maximum
pressure of 1.39 MPa that occurs at top-dead-center (TDC). The maximum pressure in
Figure 29 is 1.31 MPa, which also occurred at TDC. It is noted that at the end of intake,
the pressure was assumed to be atmospheric. The lower pressure observed in the case
with fuel injection of the liquid kerosene by the hot air in the compression chamber.
During the evaporation process, the combustion chamber temperature was reduced. This
also facilitated a drop in pressure within the chamber.

For case (i), the injection delay time was approximately 2.2 ms [25]. Using this
2.2 ms injeCtion delay, a crank offset angle of 88° was used. The offset angle describes
the compensation (in degrees) used to account for the injection delay time. The offset
angle sets the proper rotor position at the start of injection. The injection event was
observed within a time interval of 1.6 ms, so that injection occurred during a rotor
position of 70° to 64° BTDC. From prior calibration testing, the estimated mass of fuel
introduced per injection was 18 mg per injection. At the start of injection, Figure 30a
shows a cone shaped spray formed at the nozzle exit. Figures 30b and 29c show that as
the fuel jet moved outward from the nozzle tip, the liquid core of the jet traveled a ﬁnite
distance, called the break-up length, before dissipating. After reaching the break-up
length, the spray tip penetration velocity rapidly decreases. This spray behavior is a

32

33

 

 

 

 

 

 

PressureTreee
O O O
P
f
O
I
O O O O 6
O
u
r
0
O O O O O
M
p
I
O O O O O
O O O O O 6
O O 0
01cm

 

Figure 28 Pressure trace without fuel injection, engine speed of 2000 rpm

 

 

 

 

Pressure Trace
O O O
p
f
O
. O O O 6 O
O
u
l‘
0
O O O O O
M
p
I
O O O O O
O O O O O O
O O O
0 cu.
Cycles

 

 

 

Figure 29 Pressure trace with fuel injection and engine speed of 2000 rpm

34

(c) 65.65 BTDC ‘ _
-;t=0.6 ms _ "

'(‘e)‘6o.2°,éTDo j" "" (t)57.75‘BrDC
-f-t=1.0'ms" t=1.2ms

 

Figure 30 Successive frames of injection event at a pressure of 70 MPa

35

consequence of the aerodynamic resistance between the fuel and air charge in the
combustion chamber, which usually leads to enhancement of the fuel-air mixing
process. As the injection event proceeds, the fuel jet continues to penetrate the air charge,
but at a much lower rate. At the outer surface of the jet, the mixing region grows in
breadth, and the diameter of the liquid core will decrease in radius until it vanishes and
the mixing region dominates the whole jet. Figures 30d-30f demonstrate these features of
the fuel jet divergence and the signiﬁcant decrease in velocity.

The penetration distances were measured and plotted for selected injection events
using the high speed ﬁlm of case (i) and can be seen in Figure 31. This ﬁgure illustrates
typical penetration distances found in all injection events for this experiment. Notice that
between 0.4 and 0.6 ms, the slopes of these curves suddenly decrease, quantifying the
visual observation of liquid core break-up. At approximately 1.6 ms, the curves reach an
asymptote of 100 mm. At this point, it was observed that the forward motion was
negligible an no further measurements were taken. From Figure 30, a data spread of :t 20
mm in the penetration distance at 1.6 ms is observed. This was due to cyclic variations in
the fuel injection system. The average penetration distances were also calculated for each
time increment and are shown in this ﬁgure, as well as in Table 2.

 

 

p
e
n
e X Event 16
t
r U Event 17
t " Event21
g 0 Event23
n - HSF Ave.

 

 

 

33

 

 

 

 

 

Time (me)

 

Figure 31 Penetration distance for selected events and event average

36

In Figure 32, the spray tip velocities for case (i) are shown. Figure 32 shows a
maximum velocity of approximately 180 m/s for selected injection events at 0.4 ms
(second frame of the capture event). All of the events shown in Figure 32 exhibit a

 

 

 

 

 

 

 

 

 

I” T ----------------------------------
,‘*\ I I I I I I
I” -I-- - - - a, I-\\' - -I- - - - -I - - - - - - - - I- - - - - I- - - - -I
/ I \‘ I I I I I I
6 I I I I I I
1‘0 "" " ' J, 0‘ -I\ - -I ..... I - - - - I - - - - I ..... I ..... I
v ‘ ‘ I I I I I I
O I ‘
_U_ - - _ _I ..... I _ - - - 1 _ - _ _I ..... I _____ I
I 1m , I ‘\ I I I I I I 0M”
0 'l . I I I I I
01m.v_‘__._-\:. __________________ . _____ .UEvenIZ‘I
I / ‘ I ‘ I I I I I AMI”
t m 0. - - - I- - .\ I- -‘- - n- - - - I - - - - I ..... I ..... I
v I' . - , .\- . . - ~ ' XEvomao
' I \‘ ' K K. I I I I
m '_ - - - I- - _ - I _ _ -l _, - - _ - _ ._ g - - - _I ______ —I"6FAV..
’ w ' I V I ~ \I I I I
. ' I I I \\‘ I I
----- » -------- ------
' A
20 -------------- ' - - - - I - - — ------ ‘ 3
I \
I I
O I r r I I 11 41
0.2 0.4 0.6 0.8 1 12 1.4 1.6
Tlrne(me)

 

Figure 32 Velocity distribution for selected events and event average

 

 

 

T12Pn ri -. I. 91'; 'Ilei Itre, " I ° ° 'I I ‘ .r‘ I I '._
Time (ms) Penetration (mm) Velocity (m/s)
0.2 19.8 98.9
0.4 49.8 150.0
0.6 63.7 69.7
0.8 75.2 57.3
1.0 84.1 44.7
1.2 90.8 33.4
1.4 96.2 27.3
1.6 99.5 16.4

 

 

 

 

decrease in the spray tip velocity after 0.6 ms which is a consequence of the liquid core
break-up due to the aerodynamic interactions between the charge air and the fuel jet. The
average initial spray tip velocity was calculated to be 98.9 m/s, followed by an average

37

penetration rate of 150 m/s at 0.4 ms after initial capture of the injection event. In
addition to the curves in Figure 32, Table 2 provides the averaged spray tip velocities
calculated at each time increment for a time span of 1.6 ms.

3.1.2 Calculated Penetration Distances for Case (i)

The quiescent air penetration distances were calculated for case (i) using
equations (1), (2) and (3). Figure 33 compares calculated penetration distances with the
average experimental values and the values for a single, typical injection event in the
study. In the early stages of the injection process, there is fair agreement among the
curves, particularly with respect to the shape of the curves. It is important to note that if
the curves for the averaged and selected event were to be shifted to the left, excellent
agreement with the calculated values would be achieved for the early stage of injection.
This curve shifting is arguably correct in light of the cycle-to-cycle variability in the
initiation of injection. Very few of the photographed injection events began exactly at the
start of a ﬁlm frame; rather, they began sometime during the interval before the frame

 

 

1m.° -1 ooooooooo I ---------------------- II - -’ /8
I I I I I I /K. . '
I I I I I ’ ,
P I I I I I /‘ ' ‘ . I
. 1WD ., ..... I - - - - I- - - - I. - - - .I ...... ' .......
n /' '
. I I I I , o v
' . ' ‘ - ' ' ' ' - HSF Ave.
" ”.0 .n— - -I - - I- - c b - o- / . - o I- o u - l- - o - cl - - - -
. I / I I I I ﬁ Him. 8 W
I I I I I
j . ' , , , , 0 Jet Mix.
0 60'0 -‘ ..... I - - - i' - - I. - - - .0 - - - - I- - - - I ..... I - - - -
r U Event 37
n I I I I

 

 

 

33

 

 

 

 

I I I I I
/ I I I I I
I I I I I
00 L. i . ‘r l i l i
o 0.2 04 o e 0.3 1 12 14 1.6
Time(me)

 

Figure 33 Penetration distances for experimental and calculated values

38

was exposed. At later stages in the event, starting around 0.4 to 0.6 ms, the shapes of the
curves become much less similar. The quiescent curve appears to be a straight line, while
the experimental curves of the present work become exponential and begin to approach
an asymptote. This difference may well be explained by the fact that during the in-engine
injection process, the air pressure within the combustion chamber increases, thus having
a retarding effect on the forward momentum of the fuel spray. This pressure increase is
not accounted for in the expressions of Hiroyasu and Arai [8] nor the jet mixing theory
[27], so that penetration distances calculated by their expressions continues at the same
slope. At the same time the pressure increases inside the chamber, the injection pressure
is decreasing throughout the injection event. The resulting pressure term for both
analytical expressions decreases in magnitude thus leading to smaller predicted
penetration distances.

As a side note, the importance of cyclic variability in the injection process serves
no other purpose than to detract from the engines performance. Watanabe and Hamai et
al. [29] determined that the cyclic combustion variations in the rotary engine were
signiﬁcantly influenced by the geometric conﬁguration and placement of the trailing side
spark plug. Temperatures, mixture composition an gas velocities were considered in their
study. In light of their ﬁndings that showed the variations in ﬂame propagation going
from the leading side spark plug to the trailing side spark plug led to combustion
variations, the variations in the fuel-air mixing could possibly further enhance the
combustion variations.

3.1.3 Calculated Spray Cone Angle for Case (i)

The calculated Spray cone angles for case (i) were calculated using Equations (5)
and (6) discussed in detail earlier in the thesis. The theoretical spray cone angle gave a
result of 7.43° (included half angle). The second expression developed by Hiroyasu and
Kadota [28] estimated the spray cone angle for an injection pressure of 70 MPa to be
l6.l° This value was very close to the spray cone angle of the high speed ﬁlm that was
measured to be 16°: 1°. From these results, it can be seen that the expressions used to
estimate the spray cone angle are quite good in their predictions for this study.
Considering the theoretical expression and its result, 7.43° for the included half angle or
14.83° for the whole angle, its use can provide quick results at a slight expense of
accuracy. Although the Hiroyasu and Kadota [28] expression is more demanding than the
theoretical expression in its calculations, the results provided by Hiroyasu and Kadota
[28] were much closer to the measured result in this case. It appears from these

39

observations that quick results for the spray cone angle can be found using the theoretical
the jet mixing expression [27] but a more accurate result can be obtained with the
quiescent expression of Hiroyasu and Kadota [28]. Simulation studies might utilize these
analytical expressions for the spray angle in their code because of their relatively simple
calculations and good results. In Table 3, the quiescent, theoretical and experimental
averaged results for penetration distances, spray jet velocities and spray cone angles are

 

 

 

 

 

 

 

given.
Tab] m ari n f h ' i
Penetration Penetration Spray Cone Angle
Distance (mm) Velocity (m/s)
Time 0.4 ms 1.2 ms 0.4 ms 1.2 ms -
Experiment Ave. 49.8 90.8 150.0 33.4 16°
(70MPa)
Hiroyasu and Arai 61.2 105.9 145.0 46.0 -
(70MPa)
Hiroyasu and - - - - l6.1°
Kadota (70MPa)
Jet Mixing Theory 58.9 102.1 86.7 44.5 14.86°
(70MPa)

 

 

 

 

 

3.1.4 Modiﬁed Penetration Distance Calculations

As mentioned before in section 3.1.3, the cycle-to-cycle variations in the fuel
injection event appear to contribute to the deviations found between the experimental
spray penetration distance results and the theoretical and quiescent results. Unlike the
experimental results, the theoretical and quiescent calculations when plotted verses
injection time did not appear approach an asymptote value. This characteristic ‘of the
calculated values was attributed to the pressure differential terms in the analytical
expressions. The pressure terms were not assumed to vary with time in the original
calculations. In an attempt to force the results form the analytical expressions toward a
similar asymptote value found in the experimental results, the pressure terms were
recalculated to account for the dynamic pressure change in the fuel injection event as
well as in the compression process. New pressure terms were obtained by assuming three
different scenarios: (1) assuming injection pressure to be constant while varying the
chamber pressure over the compression cycle, (2) assuming both the injection pressure

40

and chamber pressure change during the compression cycle, and ﬁnally, (3) assuming
that only the injection pressure changed during the cycle while holding the chamber
pressure constant. While scenario (1) failed to provide any signiﬁcant changes in the
theoretical and quiescent penetration distances, scenario (2) and (3) showed signiﬁcant
decrease in the penetration distance. Results from the three scenarios indicate that the
penetration distances that were analytically calculate were not greatly affected by the
dynamic pressure change of the charged air in the compression chamber. When taken in
combination, that is the dynamic pressure change in the charged air and the injection
pressure change during the event, the penetration distance showed little difference from
the scenario that considered only injection pressure changes.

In determining the pressure differential terms, assumptions concerning the engine
pressure trace and the injection pressure trace were made. It was assumed that the engine
pressure trace and injector trace were linear with crankshaft displacement during the
injection event. Utilizing this assumption, it was a simple matter to calculate the pressure
terms for the analytical expressions by subtracting the injection pressure by the chamber
pressure at every 0.2 ms interval of the injection event. Linear interpolation was used to
calculate the instantaneous pressure for both the engine and injector trace at each

 

1.31 MPa
Pressure trace

MPa

-— Linear assumption

.. Injection range

Chamber Pres.
1
l

 

 

0.1 MP8 I I I I I l J
I l l l I l

0‘ 90‘ 180'

(BBC) Crankangle (TDC)

 

 

 

Figure 34 Linear assumption for chamber pressure

41

 

    
    

 

 

 

Peak
(6
‘L__ Injection trace
2
:5 Linear assumption
of“
C
.2
8
I»
§-_

Closing i # i I I l ]
Time (ms)

 

 

Figure 35 Linear assumption for the injection event

interval. The injection pressure was interpolated between the maximum injection
pressure and the closing pressure of the nozzle spring that was indicated by a calibration
reference manual [25]. The pressure trace interpolation used 90° BTDC and TDC as the
end datum points. Though the accuracy of the pressure trace interpolation was
questionable, especially at the extreme crank angles of the range, the penetration distance
calculations were limited to ranges well within the end bounds. Figures 34 and 35
illustrate the linear assumptions made for the compression process in the engine and the
injection event. These ﬁgures do not represent an exact replication of the actual traces
and are only meant to illustrate the linear assumptions that were made.

In the ﬁnal analysis, varying the pressure differential term throughout the
injection event, as opposed to utilizing a constant, resulted in theoretical and quiescent
responses that mirrored the averaged experimental penetration distance better. Other
observations included that the engine pressure during the combustion cycle had little
effect upon the penetration results. This observation indicated that the engine pressure
calculations could be neglected without sacriﬁce to the penetration results. The approach
toward an asymptotic value by the jet mixing [27] and Hiroyasu - Arai [8] values, which
reﬂected the tendencies of the experimental results, occurred when scenario (3) was
used. Table 4 presents the results of the Hiroyasu - Arai [8] penetration distances using
the three scenarios.

u i In at

42

'!-.. v.2!-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o s F. . I I' ' 0
Time Scenario 1 Scenario 2 Scenario 3 Original

(ms) APl Dist. AP2 Dist. AP3 Dist. AP Dist.

-------------- MPa mm MPa mm MPa mm MPa mm
0.0 -- 0.0 -- 0.0 -- 0.0 -- 0.0
0.2 69.6 32.2 66.4 31.4 66.4 31.4 69.6 32.2
0.4 69.56 61.2 62.8 59.5 62.9 59.5 69.6 61.2
0.6 69.53 74.9 59.3 71.8 59.4 71.9 69.6 74.9
0.8 69.5 86.5 55.9 81.8 56.0 81.8 69.6 86.5
1.0 69.47 96.7 52.4 89.9 52.5 90.0 69.6 96.6
1.2 69.44 105.9 48.9 96.8 49.1 96.9 69.6 105.9
1.4 69.4 114.4 45.3 102.6 45.5 102.7 69.6 114.4
1.6 69.36 122.2 41.9 107.6 42.1 107.7 69.6 122.4

Time Scenario 1 Scenario 2 Scenario 3 Original

(ms) APl Dist. AP2 Dist. AP3 Dist. AP Dist.

-------------- MPa mm MPa mm MPa mm MPa mm
0.0 -- 0.0 -- 0.0 -- 0.0 -- 0.0
0.2 69.6 41.6 66.4 41.1 66.4 41.1 69.6 41.6
0.4 69.56 58.9 62.8 57.3 62.9 57.4 69.6 58.9
0.6 69.53 72.1 59.3 69.2 59.4 69.3 69.6 72.1
0.8 69.5 83.3 55.9 78.8 56.0 78.8 69.6 83.3
1.0 69.47 93.1 52.4 86.7 52.5 86.7 69.6 93.2
1.2 69.44 102.0 48.9 93.3 49.1 93.4 69.6 102.1
1.4 69.4 110.1 45.3 98.9 45.5 99.0 69.6 110.2
1.6 69.36 117.7 41.9 103.7 42.1 103.8 69.6 117.9

 

 

 

 

 

 

 

43

In Table 4, results show that the injection pressure dominates the chamber
pressure term and has signiﬁcantly more effect on the penetration distance. Table 5
contains the results of the jet mixing penetration distances [27] for the three scenarios.
The terms APl , AP2 and AP3 are the pressure differential terms for the three scenarios.

In Figure 36, the original and third scenario penetration distances are plotted for
the theoretical and quiescent expressions. The experimental average is also plotted in
Figure 36 and is labeled as HSF Ave. From the plot, it can be seen that the modiﬁed
expressions track the experimental average very well. As the graph indicates, the
consideration of dynamic pressure changes can enhance the predictive power of the
analytical models for the penetration distance.

 

”or

 

 

 

 

p

O

n

o X Hire. Orig.

I

I 0 Mod. Hire.

: - HSFAvo.

g A Jot Mix. Orig.
" Cl Mod. Jet Mix.
m

m

 

 

 

Time (ms)

 

 

 

Figure 36 Penetration distances for standard and modiﬁed expressions

Velocities calculated from the original and modiﬁed penetration expressions
provide more insight on the expressions ability to accurately predict the penetration
distances. Figure 37 exhibits the velocities. Figure 37 contains plots of the velocities
calculated from the penetration distances for both the original and modiﬁed version of
the analytical expression. The average velocity for all the events is also shown in
Figure37. Both velocities for the original and modiﬁed version of the Hiroyasu and Arai
[8] track one another very closely through 0.6 ms. The velocities for the jet mixing

44

 

250- ..............................
I I I I I I I I

 

- i-Ire.&Anl Vet.

' . X Mod. Hm; AraiVol.
. [3 mm Avo.Voi.

- A Jocht.Vol.

 

 

X ModtJotMixVel.

 

0‘3 <~—oo—o<

 

 

 

u I l I I
O 02 04 05 03 1 12 1A 13
Time (ms)

 

 

 

Figure 37 Velocity comparisons of original and modiﬁed expressions

expression are also very similar to one another during the ﬁrst 0.6 ms. After 0.6 ms, the
modiﬁed versions of the analytical expressions begin to follow the same paths and begin
to deviate from the original velocities. Between 0.6 and 1.4 ms, the modiﬁed velocities
are very close to the averaged velocities. The original versions of the analytical
expressions never fall below 40 m/s. The ﬁnal velocity calculated for the averaged
velocity for the experiment was approximately 16 m/s, far below the original predicted
velocity by the analytical expressions. The modiﬁed expressions however, predict the
final velocity to be approximately 25 m/s. The modiﬁed velocities are still higher than
the actual experimental velocity, but are close to the predictions of the original
expressions.

3.1.5 Quantifying The Fuel-Air Mixture After Injection

The ability to characterize the fuel-air mixture in the combustion chamber before
ignition plays an important role in the physical designs of the combustion chamber and
the fuel delivery system. Ikegami et a1. [30] studied the effects of combustion chamber
shape and the role of pressurized injection in diesel engines. In their study, it was

45

determined that a combination of proper injection design and good chamber geometry
enhances the mixing process. In this study, a simplistic approach was taken to quantify
the fuel-air mixture in the combustion chamber at the end of the injection process.

The instant the fuel injection event terminated, the combustion chamber was
sectioned into six distinct areas. Figure 38a—38e illustrate the sectioning technique
applied. Each photo in Figure 38 represents one measuring time interval. The time
intervals for case (i) were taken every 0.6 ms, over a time span of 2.6 ms. From the
photos in Figure 38, it is observed that some areas are apparently free of fuel (lean),
while other areas tend to contain the fuel (rich). In Figure 38a, fuel is found in area 'A',
'B', ‘C‘, 'D', 'E' and 'F'. This photograph represents the instant after injection had
terminated. Figure 38b shows the fuel distribution 0.8 ms after the injection termination
and again, all areas appear to contain some fuel. After 1.4 ms, Figure 38c contains
observable fuel in areas 'A' - ‘E'; however, area 'F‘ is absent of fuel. The observations
taken from the photo in Figure 38c indicate that the forward momentum of the fuel
carries it away from the injector tip and out of the immediate area where the spark
ignition occurs. Figure 38d has a similar distribution that was found in Figure 38c after
2.0 ms. Finally, after 2.6 ms, Figure 38e shows that in areas 'A' - 'C' and possibly 'D'
there exists some fuel, while areas 'E' and 'F' are still empty of fuel.

Quantifying the fuel-air mixture in great detail was not the goal of this study. By
labeling the sectioned area with a 1 if it contained fuel or with a 0 if no fuel was
observed, simple probabilities of fuel existence within the area was possible. The
probability distribution for ﬁnding fuel within an area is presented in tabular form in
Table 6. Recent studies by Shimada [15] have shown that shadowed photography can
provide insight on the evaporation of fuel into the charged air that direct photography
may not provide. The question of considering the vaporized fuel and its possible
contribution to the 'lean - rich' description of the fuel-air mixture can not be completely
answered by this study. It is safe to assume that vaporization of the fuel does. occur
during and after injection and that higher probabilities of ﬁnding fuel (liquid or vapor) in
an area would increase if vaporization can be included.

46

(b) 40.5°BTDC \A (c) 38.1°BTDCi
. 1:08 ms * I ' 1:14-4ms i ‘

,V.

( ) meme
; utt=2.6 ms "

 

Figure 38 Fuel-air mixing after injection for Case (i)

47

 

 

 

 

 

 

 

Time (ms) A B C D E F
0.0 1 1 1 1 1 1
0.8 l 1 1 1 1 0.7
1.4 1 1 1 1 0.3 0
2.0 l 1 1 l 0 0
2.6 0.5 1 1 1 0 0

 

 

 

 

Area 'A' in Table 6 shows the presence of fuel at every point in the observation
period except for the last observation time where fuel is found half the time. Areas '8' -
'D' contain fuel throughout the observation period. The areas closest to the injector tip ,
'E‘ and 'F' have fuel at most the ﬁrst 2.0 ms after injection. For area 'B', the probability of
having fuel is only 0.3 at 1.4 ms and 0.1 at 2.0 ms. In area 'F‘ , the probability of fuel
existing has a 70% chance. It can be concluded from the observations of the last areas
that the forward momentum carried most of the fuel into farther segments.

3.2 Flow Visualization for Case (ii)

Experimental measurements and analytical predictions are discussed in the
following subsections. Case (ii) had a timing of 65° BTDC and an injection pressure of
44 MPa.

3.2.1 Penetration Distances and Velocity Measurements

For this section, all the engine operating parameters remained the same as in Case
(i). The injection pressure and injection timing were adjusted however. The estimated
amount of fuel introduced into the combustion chamber was approximately 10 mg per
injection. The average duration time of injection was 1.2 ms. The injection delay time
was determined to be 2.8 ms [25]. Using this delay time, a 11° angle off-set was
established.

Figure 39 presents the fuel spray development for Case (ii). In this ﬁgure, the
spray development process of the fuel jet is analogous to those described in Case (i). The
fuel in Figure 39 makes contact with the rotor pocket before it reaches the engine
housing which constrains the fuel to ﬂow tangentially along the rotor surface as shown in

48

(c) 602° BTDC ' , (d) 57.8° BTDC
~t=0.6 ms . . _ t=0.8 ms

 

. ,_ . \ : -...
(e) 55.4°BTDC ‘ q \ ’(t)53.0°BTDC

‘ t=i.0 ms 1:1.2 ms

Figure 39 Successive frames of injection event at a pressure of 44 MPa

49

Figures 39a-39f. Figure 39 also shows that the penetration distance and the Spray cone
angle are dependent upon the injection pressure, which in turn affects the rate of air
entrainment into the spray. These physical aspects are all related to the aerodynamic
interaction at the liquid/gas interface.

Penetration distances were also calculated for the Case (ii). Figure 40 shows the
penetration distance curves. these results are analogous to the penetration curve
characteristics for Case (i) mentioned earlier. The plots for the penetration distances
encompass the entire injection event unlike the plots for Case (i). Each curve has an
initial positive penetration slope, followed by successive decreasing positive slopes. For
Case (ii), the sudden change in the slopes that occur between 0.4 and 0.6 ms indicate
liquid core break-up. This decrease in the break-up time is to be expected; at 44 MPa, the
liquid core will have less momentum than at 70 MPa, and dissipation of this momentum
by the charged air will take less time. All the curves in Figure 40 asymptotically
approach a penetration value between 65 mm and 75 mm. The average penetration
distance at each time increment over a 1.2 ms period can also be found in Table 7. Unlike
the ﬁrst case in this study which had an observation range over 1.6 ms, the observation
range for this case lasted only 1.2 ms for two reasons. First, the injection event lasted for
only 1.2 ms. Second, the forward momentum after 1.2 ms was barely measurable by

visual means.

 

 

p

o

n

e X Eventas
1

r 0 Event 37
t C} Eventaa
:1 A Event39
n - HSF Ave.

 

 

 

 

 

 

 

 

Time (ms)

 

Figure 40 Penetration distances of selected events

50

Figure 41 presents the spray tip velocity of selected injection events for Case (ii).
In Figure 41, all the injection events show an increase in velocity between 0.2 ms and 0.4
ms. After 0.4 ms, all events exhibit the same characteristics of a deceleration pattern that
is sustained through the end of the injection event. At the lower injection pressures, there
are corresponding lower spray tip velocities. The average velocity over 1.2 ms was
calculated to be 57.9 m/s. The initial velocity was estimated to be 81.8 m/s, followed by
an average velocity of 111.8 m/s at 0.4 ms. Table 7 lists the averaged spray tip velocities
at each time increment for a time span of 1.2 ms.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

V
7 1 Cl EvontZO
3 A Evom21
: O Evont23
y X was
IT - HSFAvo.
O I I 1 1 I
02 0.4 0.6 0.8 1 1.2
Timo(rno)
Figure 41 Spray tip velocities of selected events
T.ble PMW I: °v1'ors I-. Im‘ r fI' ' 1010 ' .r‘ I H U..'.
Time (ms) Penetration (mm) Velocity (m/s)
0.2 16.4 81.8
0.4 38.7 111.8
0.6 51.3 62.7
0.8 59.4 40.8
1.0 65.1 28.3
1.2 69.4 21.7

 

 

 

 

51

3.2.2 Calculated Penetration Distances for Case (ii)

The quiescent air penetration distances were calculated again using equations (1),
(2) and (3). The results of the calculations are very similar to the ﬁndings in Case (i).
Figure 42 presents the penetration distances for the analytical expression, the average of
all the events and a single typical event. In the early stages of the injection process, there
is some agreement between the typical event and the expressions. The average of all the
events is smaller in magnitude throughout the whole observation period. Again it is felt
that the discrepancies between the averaged penetration distance and the other values are
heavily inﬂuenced by the cyclic variations in the injection process. The analytical
expressions contain two distinct slopes during the 1.2 ms. The ﬁrst slope is contained in
the ﬁrst 0.4 ms. The penetration slopes and values of both the Hiroyasu-Arai [8] and the
jet mixing theory [27] are almost identical over the entire range. Again as in Case (i), the
penetration values of the two expressions do not appear to approach an asymptote. The
penetration values obtained from the high speed ﬁlm do however show effects of the
increase in chamber pressure, and do gravitate towards a possible asymptote.

 

 

.......... A Hire. 0 Arai
0 JET Mix.

' — AVERAGE
40.00 - ' , ‘ ---------------------- El Event 8

30-"IW-003.‘
8
8
J

 

 

 

33

 

 

 

duh- -
d

 

 

 

Figure 42 Penetration distances at an injection pressure of 44 MPa

52

3.2.3 Calculated Spray Cone Angle for Case (ii)

Using Equations (5) and (6), the spray cone angles were calculated. The spray
cone angles for this case were slightly smaller than those found in the previous case.

Tl m'nrh

 

 

 

 

 

 

 

 

 

 

 

Penetration Penetration Spray Cone Angle
Distance (mm) Velocity (tn/s)
Time 0.4 ms 1.2 ms 0.4 ms 1.2 ms
Experiment Ave. 38.7 69.4 111.8 33.4 15°
(44MPa)
Hiroyasu and Arai 50.8 94.1 127.0 41.0 -
(44MPa)
Hiroyasu and - - - - 14.3°
Kadota (44MPa)
Jet Mixing Theory 52.3 90.6 76.5 39.5 14.86°
(44MPa)

 

Table 8 contains a summary of the calculated spray cone angle results and also
the penetration and velocity results. The results show that at lower injection pressures,
the spray cone angle tends to decrease.

The jet mixing theory [27] produced the same result of 7.43° for the included half
angle, or 14.86° for the whole angle. The results were exactly the same due to the
dependance of the jet mixing equation on only the densities of the air and fuel. The
density of air was assumed to be the same for both Case (1) and Case (ii) since both
utilized the same compression ratio. The density of the fuel was assumed to be constant
over the injection period.

3.2.4 Modiﬁed Penetration Distance Calculations

The dynamic pressure changes within the engine were again analyzed to
determine their effect on the calculated penetration distances. Once again the three
scenarios discussed in an earlier section were considered. In the ﬁrst scenario, it was
assumed that the chamber pressure varied with time only. In the second scenario, it was
assumed both the chamber and injection pressure varied in time. Finally, in the last

53

scenario, only changes in the injection pressure were considered. It was not surprising to
ﬁnd that again as in Case (i), the calculated penetration distances did not respond to the
small changes in the chamber pressure. Only when the fuel injection pressure changes
were considered did the penetration curves react differently. The same procedure for
determining the differential pressure term in Case (i) was implemented. Once again,
when the pressure term in the Hiroyasu-Arai [8] and the jet mixing expression [27] was
varied as a function of the injection pressure, the resulting penetration curves ﬂattened
signiﬁcantly. Even with the modiﬁed pressure term however, the curves failed to reach
the average penetration curve of the high speed ﬁlm as seen in Figure 43. After 0.4 ms,
the modiﬁed curves have nearly the same slope as the averaged curve. From 0.0 ms to
0.4 ms, the modiﬁed curves penetrate farther into the compression chamber in the same
amount of time as the averaged curve.

As in Case (i), the question that remains to be answered is how well can the
modiﬁed expression predict the penetration of the fuel. It appears that the general shape
of the modiﬁed penetration curves mirrors the averaged penetration curve very closely.
This observation begs another question of how the possible effect of cyclic variation in

 

 

 

 

 

   

 

 

 

 

1w HF -------------------------------
I I I /
go. ...................... . ....... . ..... 2
z . /
I
w .i ...................... I - v .' : - - , -- '-
P /' . / .
. I I a /‘ ' - . / I
n 70 .................... - .........
o XHiroprig.
I ............ ,- - . ' ..............
, 6° 7 . / OMod.Him.
': 50 «r --------- , . ' ----- ' ---------- —HSFAve.
i . .
o 40 4 __________________________ Am MIX. Ong.
" DMoIIJoIMIx.
x - ---------------------------------
m
m I
m up. - - o/ ----------------------------
10‘ """" .-""r """ I '''' .""".
o L 1 i i ﬁ‘ 4. i
0 02 04 06 0.8 1 12
TIrno(ms)

 

Figure 43 Penetration distances for standard and modiﬁed expressions

54

the fuel injector response might result in conservative ﬁlm measurements. It has not been
shown in this study that cyclic variation of the fuel injection event are signiﬁcant enough
to pull down the penetration curve observed in the 0.0 to 0.4 ms range. It is quite possible
that the analytical expressions (original and modiﬁed) are over zealous in their
predictions. The fact that the modiﬁed curves and the averaged curve cover
approximately the same amount of distance between 0.2 ms and 1.2 ms tends to suggest
to this author that cyclic variations are the greatest cause in the discrepancies found
between the modiﬁed analytical and ﬁlm penetration results. Tables 9 and 10 contain the
results of the three scenario for both the Hiroyasu-Arai [8] and the jet mixing theory
[27].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T l iﬁ i n ' ° 'r
Time Scenario 1 Scenario 2 Scenario 3 Original
(ms) AP1 Dist. AP2 Dist. AP3 Dist. AP Dist.
-------------- MPa mm MPa mm MPa mm MPa mm
0.0 -- 0.0 -- 0.0 -- 0.0 -- 0.0
0.2 43.6 25.3 40.5 24.5 40.5 24.5 43.6 25.4
0.4 43.57 50.7 37.6 47.2 37.6 47.2 43.6 50.8
0.6 43.54 66.4 34.67 62.8 34.7 62.8 43.6 66.5
0.8 43.51 76.7 31.74 71.0 31.8 71.0 43.6 76.8
1.0 43.48 85.7 28.81 77.4 28.9 77.5 43.6 85.9
1.2 43.45 93.9 25.87 82.6 26.0 82.7 43.6 94.1
Time Scenario 1 Scenario 2 Scenario 3 Original
(ms) APl Dist. AP2 Dist. AP3 Dist. AP Dist.
-------------- MPa mm MPa mm MPa mm MPa mm
0.0 -- 0.0 -- 0.0 -- 0.0 -- 0.0
0.2 43.6 36.9 40.5 36.3 40.5 36.3 43.6 37.0
0.4 43.57 52.3 37.6 50.4 37.6 50.4 43.6 52.3
0.6 43.54 64.0 34.67 60.5 34.7 60.5 43.6 64.1
0.8 43.51 73.9 31.74 68.4 31.8 68.4 43.6 74.0
1.0 43.48 82.6 28.81 74.6 28.9 74.7 43.6 82.7
1.2 43.45 90.5 25.87 79.6 26.0 . 79.7 43.6 90.6

 

 

 

 

 

 

55

The results found in Tables 9 and 10 and observing their plots in Figure 43 implies that
simulation programs might utilize the analytical expressions with some success when
attempting to predict the penetration distances of fuel jet sprays.

As in Case (i), the determination of the velocities calculated from the original and
modiﬁed penetration distances might provide more information about the ability of the
expressions to accurately predict penetration distances. Favorable comparisons to the
average velocity plot of the high speed ﬁlm would indicate good predictive power.
Figure 44 provides the velocity plots of the modiﬁed and original penetration
expressions. Also included in Figure 44 is the average spray tip velocity measured from
the high speed ﬁlm. The original calculated velocities for the Hiroyasu-Arai [8] and the
jet mixing [27] expression are very high when compared to the measured velocity at 0.2
ms. The original velocity and modiﬁed velocity for the Hiroyasu and Arai [8] track one
another during the ﬁrst 0.8 ms. This observation of tracking is also true for the jet mixing
velocities. At 0.8 ms, the original and modiﬁed velocities for both expression separate
where the original velocities assume a high prediction line while the modiﬁed velocities
fall almost exactly on the average velocity line. At 1.2 ms, the original velocity lines
have a value of approximately 40 m/s. The modiﬁed velocity curves have a value of

 

 

 

. 0 ModJotMixVoi.

 

v

0

I X Hire. a Aral Vol.

0 -

c . Cl Mod. Him. a Arai Vol.

: - Cue (i) Ave. Vol.
I

V . A Jot Mix. Vol.

III

I

8

 

 

 

 

 

 

 

Figure 44 Velocity comparisons of original and modiﬁed expressions

56

25 m/s which is very near the average velocity of 20 m/s at 1.2 ms. These results indicate
that the modiﬁed versions of the analytical expressions are better at predicting the
penetrating distance than the original versions, especially well into the injection event.

3.2.5 Quantifying The Fuel-Air Mixture After Injection

The combustion chamber was again sectioned into 6 areas labeled 'A' - 'F‘. Due to
the later injection timing of 65° BTDC and a smaller injection pressure, the mixing
results exhibit some slight differences from the observations in Case (i). Figure 45a-45f
presents a typical mixing event observed in this case. Figure 45a contains fuel in all the
area except 'A' at the instant the injection event terminated. In Figure 45b, 0.8 ms after
termination, area 'E' is the only segment without fuel. A combination of the fuels forward
momentum and the rotor advancement which decreases the area of observation explains
the emergence of fuel in area 'A'. At 1.4 ms, the fuel appears to moved out of the two
areas closest to the injector tip and is concentrated in areas 'A' — 'D' which can be
observed in Figure 45c. Further investigation of the mixing process shows little
difference from what was observed at 1.4 ms. Figure 45d shows that most of the fuel is
concentrated in areas '3' - 'C‘. Fuel is observable in areas 'A' and 'D', but the fuel-air
mixture is much leaner in these areas. Figure 45c shows much of the same results as
Figure 45d, but also might show some fuel in are 'E'.

The results presented in this section are based solely on the observations obtained
from the high speed ﬁlm and the 35 mm photos of selected frames. The granular
resolution of the high speed ﬁlm was transferred to the 35 mm stills which made the
identiﬁcation of areas with a lean mixture was very difﬁcult. Table 11 presents the
probability of ﬁnding fuel in each area during each time interval.

.Il' l'IIII I II rv' ' I.-. .' 'I 2' '1,°!o._'-

 

 

 

Time (ms) A B C D E F
0.0 0 1 l 1 1 1
0.8 0.6 1 l 1 1 0.9
1.4 0.9 1 1 l 0.3 0.2
2.0 l 1 l 1 O 0
2.6 1 1 1 0 8 0 0

 

 

 

 

 

 

 

 

57

(c) 45.7mm“:

 

Figure 45 Fuel-air mixing after injection for Case (ii)

58

The probability of ﬁnding fuel in area 'A' increases as time increases. Forward
momentum of the fuel along with the decrease in volume due to compression would
explain this distribution for area 'A'. Typically, the middle areas ‘8' - 'D' will contain fuel
during the post-injection and pre-combustion period. The ﬁnal two areas, 'E‘ and 'F‘ will
contain some fuel during the ﬁrst three observation stages, but forward momentum will
ultimately carry the fuel out of the areas. By 0.2 ms into the observation period, area 'E'
will only have a 30% chance of containing fuel, while area 'F‘ will contain fuel only
about 20% of the time.

3.3 Flow Visualization for Case (iii)

Mixing observations are presented in the following subsections. Case (iii) has an
injection pressure of 44 MPa and a timing of 55° BTDC

3.3.1 High Speed Film Observations

Figure 46 shows the fuel spray development for Case (iii). The early interaction
of the fuel jet with the rotor surface, is indicated in Figures 46b-46e. Figures 46f-46i
illustrate the deflection of the fuel jet from the r0tor pocket toward the engine housing.
The fuel-air mixing in his case was relatively weak as compared to the results in the prior
cases. Due to the early interaction between the fuel spray and the rotor, no qualiﬁcation
of the spray characteristics were made other than the fuel-air mixing. One observation
that can be made is the area that the jet consumes is much less than case (i) and (ii). This
observation can be related to the physical amount of fuel introduced per injection is
nearly half the mass seen in Case (i). In terms of Case (ii) which also injected at 44 MPa,
the impingement on the rotor prohibits the spreading of the jet during the injection
process. As seen in the velocity plots for the ﬁrst two cases, higher injection pressures
lead to higher spray tip velocities. Higher velocities have greater shear forces at the radial
edge of the liquid core which results in a greater removal rate of liquid dr0plets from the
liquid core at the radial edge. These observations would imply that the optimal fuel-air
mixture is very likely to occur at high injection pressures within the combustion
chamber.

59

(a) 55 0° BTDC - . (b)52f5°BT’oo (c) 50.16’Broc
t=0gms ' t. , . ‘ 7 t=0.6 ms

(6343-666;be (e') 45.2°'BTDC
"t=0.8 ms : ’ t=1.o ms -

(§)'40T33_B‘TDCT
t=1.4 ms ,'

 

Figure 46 Successive frames of one injection event at a pressure of 44 MPa

60

3.3.2 Quantifying The Fuel-Air Mixture After Injection

Figure 47 illustrates the difﬁculties encountered when an attempt was made to
quantify the mixing region. The observable volume was much smaller than those found
in Case (i) and Case (ii). Figure 47a shows fuel existing in areas 'B', 'C' and 'D' at 0.0 ms
after the injection event had ceased. At 0.8 ms, fuel can be found in areas B', 'C, 'D' and
'E' as shown in Figure 47b. The possibility of fuel existence in area 'A' might also be
argued for due to the thin sheet of light near the apex. It determined that this ﬂash in area
'A' could probably be attributed to light reﬂection at the Plexiglass window. The high
speed ﬁlm clearly indicates that the flash is a consequence of the laser light and
plexiglass window interaction. In the process of transferring information from the high
speed ﬁlm to the 35 mm stills, the resolution of the small areas is not retained 100%.
Figure 47c also produces the same results found in Figure 47b but also may include some
fuel in 'F‘. The observation range only contained three steps for this case due to the
conﬁned working volume. The probabilities of ﬁnding fuel in the six areas were again
calculated from the data taken from the high speed film. Table 12 contains the
probability results for this case.

‘5
- (cféSlf-Bfod

' “37.9°3Toc
' '-" j..->i.¢:-"I'§‘::‘i' =O14 m5

   

Figure 47 Mixing observations for Case (iii)

61

 

 

 

l ili n
Time (ms) A B C D E F
0.0 0 1 1 1 0.1 0.5
0.8 0.1 1 1 1 0.3 0
1.4 0 0.9 1 1 0 0.1

 

 

 

 

 

 

 

 

From the table, it can be seen that there is not a high probability of ﬁnding fuel in
area 'A' at any time. This indicates that the fuel impingement on the rotor which results in
a severe forward momentum loss. The internal areas 'B', C' and 'D' almost always contain
fuel throughout the observation period. It is in these areas that the fuel appears to be most
concentrated. The two areas closest to the injector tip, 'E' and 'F' contain fuel only 10% to
50% of the time after the injection event has terminated.

3.4 Comparing Case (i) and Case(ii) Spray Characteristics

The similarities and differences in the spray characteristics of the two cases will
be addressed in this section. It has already been stated that the physical combustidn
chamber volume consumed by the fuel spray during the injection event appears to be
greater in Case (i) as illustrated in Figures 30 and 39 of the previous two sections. This
result is to be expected due to the greater amount of fuel being introduced in Case (i).

Overall, the penetration distance, spray tip velocity and the spray cone angle were
smaller in magnitude in Case (ii) for all observations. Although there is a 5° difference in
the injection timing (70° BTDC for Case (i) and 65° BTDC for Case (ii)), it very
unlikely that the timing could cause these differences between the cases. To further this
idea, the penetration distance calculated from the analytical expressions were clearly very
responsive to injection pressure changes rather than the compression pressure changes
within the chamber during the injection event. So it appears after analyzing both cases
that not only does the injection pressure play an important role in the fuel spray
characteristics during the injection event, it also has a global effect upon the spray
characteristics from event to event. Table 13 contains the penetration distances for both
cases.

191'.

Oll'ejl' 0'1'!-. II

62

 

 

 

 

 

 

 

Hiroyasu & Jet Mixing Experimental
Arai(mm) (mm) Ave. (mm)
Time Case Case Case Case Case Case
ms 0) (ii) (i) Iii) (i) (ii)
0.0 0.0 0.0 0.0 0.0 0.0 0.0
0.2 31.4 24.5 41.1 36.3 19.8 16.4
0.4 59.5 47.2 57 .4 50.4 49.8 38.7
0.6 71.9 62.8 69.3 60.5 63.7 51.3
0.8 81.8 71.0 78.8 68.4 75.2 59.4
1.0 90.0 77.5 86.7 74.7 84.1 65.1
1.2 96.9 82.7 93.4 79.7 90.8 69.4
1.4 102.7 - 99.0 - 96.2 -
1.6 107.7 - 103.8 - 99.5 -

 

From Table 13, it can be seen that the penetration distance decreases as the injection
pressure decreases over the ﬁrst 1.2 ms. The results for the Hiroyasu-Arai [8] and the jet
mixing [27] expressions were obtained by assuming the pressure differential term ﬁts the
third scenario. The modiﬁed theoretical estimation does a better job of approximating the
penetration distance for both cases. Figures 48 illustrate the penetration distances for both

cases.

 

SCI-"I‘I'OODC'U

 

 

 

X Mod. Him.
X Case (1) Ave.
A Mod. Jot Mix.
' 0 Mod. leo.
"' Cm (1i) Avg.

 

' Dmuwx.

 

 

 

 

0.8 0.8
Time (ms)

 

 

Figure 48 Penetration distances for both Case (i) and Case (ii)

 

63

Table 14 shows that the spray cone angle also decreases with a decrease in injection

 

 

 

pressure.
WW
Hiroyasu & Kadota Jet Mixing Experiment
Case (i) 16.1° 14.86° 16.0°
Case (ii) 14.3° 14.86° 15.0°

 

 

 

 

 

3.5 Nozzle Tip Verification

The nozzle tip used throughout this study was part of a special set manufactured
by BKM Inc. Orders were placed to have nozzle tips that re-directed the spray at
orientations of 60°, 45° and 30° from the nozzle axis. Upon receiving these tips, it was
necessary to observe the spray formations induced by these tips in order to verify their
operational status. The ﬁrst nozzle tested was the 60° tip. It proved to operate very
successfully at all injection pressure loads. Figure 49 illustrates the jet formation of the

 

Figure 49 Injection into ambient conditions using 60° nozzle

64

60° nozzle ﬁred into ambient conditions. The 45° tip was then tested with less than
satisfying results. It appeared from the spray formulations captured on high speed 35 mm
ﬁlm that tip had a leaking problem. Large fuel droplets and inconsistent spray patterns
were observed during the tips testing. Figure 50 shows a 45° injection event also
conducted in ambient conditions. The slow exposer time tends to exaggerate the large
fuel droplets that can be seen streaking away from the main jet formation; however, no
events of the 60° test produced these same results. Previous studies have indicated that
the nozzle hole shape has a great inﬂuence on the jet formation [31]. The nozzle hole
geometry of this particular 45° tip could be the reason for its inconsistent performance.
The 30° nozzle was not tested due to the fact that its angled injection would not lend
itself to easy ﬂow visualization studies in the motored engine. It was determined in the
static test that at all injection timings would impinge on the rotor at a 30° angle of
injection.

 

Figure 50 Injection into ambient conditions using 45° nozzle

3.6 Blowby Observations

Recent studies have determined that there is existence of blowby in the motored
rotary engine without fuel injection. The blowby jet which is created when separation

65

between the apex and the engine housing occurs has been observed in high speed ﬁlms
and measured with Laser Dopler Velocity (LDV) techniques [1,32]. When observing the
high speed ﬁlms for each case, blowby can be seen during and after the fuel injection
event. The airﬂow past the apex seal was seeded with the fuel particles from the fuel
injection event; however, exactly how much inﬂuence did the injection event have on the
blowby can not be said. Measuring the blowby during fuel injection with LDV
techniques might provide some insight to the question of how the fuel injection effects
the blowby in the engine. Figure 51 illustrates the blowby seen past the apex seal after an
injection event.

 

Figure 51 Blowby seen in the engine after the injection event

CHAPTER 4

SUMMARY AND CONCLUSIONS

Some characteristics of the behavior of a high pressure fuel injected into the

combustion chamber of a rotary engine have been described by analyzing high speed
ﬁlms of the process. In addition to the qualitative description of the injection process, the

ﬁlms provide quantitative information such as spray tip penetration velocities and spray

angles. Comparison with current fuel injection models gives additional insight on the

characteristics of the fuel spray. The following remarks can be made:

1.

Spray tip penetration distances and velocities are signiﬁcantly inﬂuenced by
injection pressure. Both increase with an increase in the injection pressure.

At these relatively high injection pressures (44 and 70 MPa), the combustion
chamber air ﬂow seems to have little inﬂuence on the spray characterization as
conﬁrmed by bOth visual observations and comparisons with quiescent air models.
After the injection process is completed, the fuel-air distribution is signiﬁcantly
affected by the combustion chamber air ﬂow.

The Hiroyasu and Arai [8] and the jet mixing [27] models describing spray
characteristics for fuel injection into quiescent air can be used to predict injection
penetration in simulation studies, provided the air pressure differential term in the
expression is allowed to vary with time.

At higher injection pressures, the jet mixing area is larger due to the higher
velocities and greater shear stresses at the radial edge of the liquid core. These

higher velocities tend to enhance the break-up of the fuel spray.

Liquid core break-up length appears to increase with an increase in the injection
pressure.

Late fuel injection at high pressures can lead to direct impingement of fuel on the
rotor which results in a reduction of fuel-air mixing.

66

10.

11.

67

Experiments have shown the Hiroyasu and Kadota [28] spray cone angle
expression to be very adequate at predicting spray cone angles while varying the
injection pressure.

Pressure traces might be utilized in numerical simulations to calibrate evaporation
rates during the injection process.

Inclusion of a pressure differential term in the quiescent Hiroyasu-Arai model [8]
and the theoretical jet mixing equation [27] can provide better estimations of the
fuel penetration distance.

For these experiments, it was observed that there existed areas of rich and lean
fuel-air mixture.

Blowby can have a major inﬂuence on the local air-fuel ratio existing in the
combustion chamber prior to ignition.

CHAPTER 5
RECOMMENDATIONS

The observations and ideas discussed in this study are not intended portray, nor
do they approach a complete understanding of direct fuel injection within a rotary
engine. Research is a continuous process that never stops. It is hoped that this study
might provide some foundation for further work towards a better understanding of the
combustion process in the rotary engine. Future work that might build upon the ﬁndings
of this study will now be recommended.

1. It appears from the results in this study that analytical expressions might be used
to predict penetration distances of the fuel spray with some success. Further
investigations might reveal how the cyclic characteristics of the fuel injection system
when captured on synchronized ﬁlm lead to differences in the experimental and
calculated results.

2. A complete mapping between the spray tip velocity and fuel injection pressure
might be accomplished by using LDV techniques over a variety of operating ranges of
the engine.

3. The mixing process within the combustion chamber after injection was brieﬂy
discussed in this study for all three cases. Future studies might concenu'ate on the mixing
characteristic of the combustion chamber using other optical recording techniques that
can access the close quarters of the combustion chamber. Improving the combustion
process will clearly depend heavily upon the understanding of the mixing process.

4. The inﬂuence that the fuel injection has upon the leakage past the apex seals

should be studied in order to accommodate its effects in simulation modeling of the
rotary engine.

68

REFERENCES

 

LIST OF REFERENCES

l. Chouinard, E., "Experimental Investigation of the Flow Field in a Motored
Rotary Engine Assembly," Thesis for the Degree of M.S. at Michigan State University,
1990.

2. Hamady, F.J., Kosterman, J., Chouinard, E., Somerton, C., Schock, H., Chun, K.
and Hicks, Y., "Stratiﬁed Charge Rotary Engine Studies at the MSU Engine Research
Laboratory," SAE Paper 890331. Also in SAE Trans., Vol 98, Sec. 3, 1989.

3. Haddad. S.D..Aslxan£s.d_1_>msel_ﬁngmmund_0mtanm Ellis Horwood
Limited, New York, NY., 1988.

4. Savey, C.W., Beck, J .N., Dobovisek, Z. and Gevert, K., "Injection Characteristics
of High Pressure Accumulator Type Injector," SAE Paper 890266, 1989.

5. Kato, T., Tsujimura, K., Shintani, M., Minami, T. and Yamaguchi, I.,"Spray
Characteristics and Combustion Improvement of DJ. Diesel Engine with High Pressure
Fuel Injection," SAE Paper 890265, 1989.

6. Katsura, N., Saito, M., Senda, J. and Fujimoto, H., "Characteristics of a Diesel
Spray Impinging on a Flat Wall," SAE Paper 890264, 1989. 1

7. Minami, T., Yamaguchi, I., Shintani, M., Tsujimura, K. and Suzuki, T., "Analysis
of Fuel Spray Characteristics and Combustion Phenomena under High Pressure Fuel

Injection," SAE Paper 900438, 1990.

8. Hiroyasu, H. and Arai, M., "Structures of Fuel Sprays in Diesel Engines," SAE
Paper 900475, 1990.

69

70

9. Hiroyasu, H. and Nishida, K., "Fuel Spray Trajectory and dispersion in a DJ.
Diesel Combustion Chamber," SAE Paper 890462, 1989.

10. Koo, J-Y. and Martin, J.K., "Droplet Sizes and Velocities in a Transient Diesel
Fuel Spray," SAE Paper 900397, 1990.

l 1. Yoshikawa, S., Furusawa, R., Arai, M. and Hiroyasu, H., "Optimizing Spray
Behavior to Improve Engine Performance and to Reduce Exhaust Emissions in a Small
D.I. Diesel Engine," SAE Paper 890463, 1989.

12. Heinze, T. and Schmidt, T., "Fuel-Air Ratios in a Spray, Determined between
Injection and Autoignition by Pulsed Spontaneous Raman Spectroscopy," SAE Paper
892102, 1989.

13. Arcoumanis, C., Cossali, E., Paal, G. and Whitelaw, J .H., "Transient
characteristics of Multi-Hole Diesel Sprays," SAE Paper 900480, 1990.

14. Werlberger, P. and Cartellieri, W., "Fuel Injection and Combustion Phenomena in
a High Speed DI Diesel Engine Observed by Means of Endoscopic High Speed
Photography," SAE Paper 870097, 1987.

15. Shimada, T., Shoji, T. and Takeda, Y., "The Effect of Fuel Injection Pressure on
Diesel Engine Performance," SAE Paper 891919, 1989.

16. Raju, M.S. and Willis, E.A., "Analysis of Rotary Combustion Processes Based on
Unsteady, Three Dimensional Computations." AIAA-90-0643, 1990.

17. Raju, M.S. and Willis, E.A., "Computational Experience with a 3-D Rotary
Engine combustion Model," presented at the 1990 Joint AIAA/FAA.

18. Abraham, J ., Wey, M.J. and Bracco, F.V., "Pressure Non-Uniformity and Mixing
Characteristics in Stratiﬁed-Charge rotary Engine combustion," SAE Paper 880624,
1988.

71

19. Abraham, J. and Bracco, F.V., "Comparisons of Computed and Measured
Pressure in a Premixed Charge Natural Gas Fueled Rotary Engine," Rotary Engine
Design: Analysis and Development, SAE SP-768, SAE, 1989, pp. 117-131.

20. Shih, T. I-P., Schock, H]. and Ramos, J.I., "Fuel-Air Mixing and Combustion in
a Two-Dimensional Wankel Engine," SAE Paper 870408, 1987.

21. Steinthorsson, E., Shih, T. I-P., Schock, HI. and Stegeman, J ., "Calculation of
the Unsteady, Three-Dimensional Flow Field Inside a Motored Wankel Engine," SAE
Paper 880625, 1988.

22. Li, T., Steinthorsson, E., Shih, T. I-P. and Nguyen, H.L., "Modeling and
Simulation of Wankel Engine Flow Fields," SAE Paper 900029, 1990.

23. Hamady, H.J., Stuecken T. and Schock, H.J., "Airflow Visualization and LDA
Measurements in a Motored Rotary Engine Assembly Part 1: Airﬂow Visualization,"
SAE Paper 900030, 1990. Also to appear in SAE trans, 1990.

24. Mazda Motor Co., "Mazda RX-7 Workshop Manual," Repair Manual, 1984.

25. BKM, Inc., F-429, "Calibration Test Report Servojet Injector for NASA Lewis
Research Center."

26. Lefebvre, A.H., 'Properties of Sprays," International Conferences on Mechanics
of Two-Phase Flows, 1989.

27. Abramovich, G.N., W, MIT Press, Cambridge, Mass.,
1963.

28. Hiroyasu, H. and Kadota, T., "Fuel Droplet Size Distribution in a Diesel
Combustion Chamber. " SAE trans, Paper 74017, 1974.

29. Watanabe, S. and Hamai, Y., "Investigation of Cyclic Combustion Variation in
Rotary Engine," SAE Paper 890327, 1989.

72

30. Ikegami, M., Fukuda, M., Yshihara, Y. and Kaneko, J., "Combustion Chamber
Shape and Pressurized Injection in High-Speed Direct-Injection Diesel Engines," SAE
Paper 900440, 1990.

31. Ohm, T.F., Senser, D.W. and Lefebvre, A.H., "Geometrical Effects on the
Breakup of Hi gh-Velocity Jets," ASME Paper, Winter Anual Meeting, 1990.

32. Chouinard, E., Hamady, E]. and Schock, H.J., "Airflow Visualization and LDA
Measurements in a Motored Rotary Engine Assembly Part 2: LDA Measurements," SAE
Paper 900031, 1990. Also to appear in SAE trans., 1990.

"llllllll‘lll"Elli