-
x. a... ,..
:14) w.

. nun,
..wm,

{a ..r
,i

kmmu «up
9 ......aaﬁawm
uﬁnr ‘

.Irﬁ.

51;?!

, an“;
. am,
is
4

1‘ .3115 A:
‘ nAﬂaiu‘wﬂ...
{a ...a

. 1.1.
.1735 .IV}
1 ¢ 2.

 

......!.:c‘ ...-.-.. ...,
.a s ...i iii: ... 5min ..nmgﬂ ... .

.... .. ‘2; . ‘ 1:... , . 3i; .5! 3... ....

. : : ....

...},

 

 

Jada—1:. . .....x, . 1...“: .
, 4.. Anal... Lt.

l I. I1 ’I ' ll

2'

I
II, It.“

I'

 

THESE;

2“ 1111111111

LIBRARY
Michigan State
Unlverslty

 

111111111111111111111111111111111111

 

                  

 

 

 

This is to certify that the

thesis entitled
EFFECT OF SPARK PLUG GROUND
ELECTRODE GEOMETRY 0N EXHAUST
EMISSIONS OF SMALL UTILITY ENGINES
presented by
Mahmood Ahmed Akhter Rahi

has been accepted towards fulﬁllment
of the requirements for

_Ji..S._ degree in Mechanical. Engineering

   
 

Major professor

Date 5//3/?‘9'
ﬁ

0-7639 Msuaan Aim—mu- . - ".1 m" ' lmtimtion

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

we mm“

EFFECT OF SPARK PLUG GROUND ELECTRODE GEOMETRY ON
EXHAUST EMISSIONS OF SMALL UTILITY ENGINES

By

Mahmood Ahmed Akhter Rahi

A THESIS

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
For the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

1998

ABSTRACT

EFFECT OF SPARK PLUG GROUND ELECTRODE GEOMETRY ON
EXHAUST EMISSIONS OF SMALL UTILITY ENGINES

By

Mahmood Ahmed Akhter Rahi

With increasingly stringent regulations placed on exhaust emissions of small
utility engines, more opportunities for reduction in emissions are being explored. This
thesis presents the results of a study in which the inﬂuence of spark plug electrode shape
on exhaust emissions of three small utility engines is analyzed. The spark plugs studied
were Champion models of standard electrode shape and spark plugs with a crown-shaped

electrode, believed by the manufacturer (Pyrotek) to give improved combustion.

The experimental results indicate that the Pyrotek spark plug causes an
appreciable reduction in hydrocarbon emissions with a moderate increase in emissions of
oxides of nitrogen. The Pyrotek plug was more sensitive to changes in the angular
position of the ground electrode, spark gap, engine speed, and to minor changes (i.e.
notches) to the electrode’s geometry. The reduction in hydrocarbon emissions is
attributed to a faster burning rate and ﬂame front propagation. The Pyrotek plug is also
thought to favor the formation of more spherical ﬂame ﬁonts than the baseline spark plug

and to inﬂuence the bulk charge motion in the vicinity of the spark gap.

This thesis is dedicated to my parents

Bashir Ahmed Akhter (late)
&
Salima Akhter

For their love, kindness and prayers for my success

iii

ACKNOWLEDGMENT

I would like to pay special thanks to a few important people:

First, I would like to extend my sincere gratitude to Dr. Giles Brereton, my Major
Professor, for being my adviser, and for his support and patient guidance throughout my
thesis work and studies.

Dr Harold Schock, for being on my committee, and for his moral support and
guidance; Dr Indrek Wichman, for being on my committee and for his helpful attitude
and guidance in understanding the peculiar nature of the problem; Dr Keunchel Lee and
Dr Bassam Ramadan for their valuable comments; Lawrence Dalimonte for his help in
setting up the engines and data collection; Tom Stueken, for his help in test rig
modiﬁcation and professional photography. I would also like to thank my fellows at
MSUBRL, for their day to day guidance and their wonderful company; Matt Foster, Jon
Darrow, Mark Novak, Mikhail Ejakov and Hans Hascher.

Finally, my wife Rashda, for her patience and support throughout my Masters

Program, and my mother without whose prayers, this day would not have been possible.

iv

TABLE OF CONTENTS

TABLE OF CONTENTS ................................................................................................ v

LIST OF TABLES .......................................................................................................... x

LIST OF FIGURES ........................................................................................................ xi

NOMENCLATURE ..................................................................................................... xiii
CHAPTER 1

INTRODUCTION ........................................................................................................... l

1.1 Overview ............................................................................................................. 1

1.2 Formation of Noxious Exhaust Emissions ............................................................ 3

1.2.1 Carbon Monoxide ..................................................................................... 3

1.2.2 Unburned Hydrocarbons .......................................................................... 3

1.2.3 Oxides of Nitrogen ................................................................................... 5

1.3 Literature Review ................................................................................................ 6

1.4 Objective ............................................................................................................. 9
CHAPTER 2

THEORETICAL EXPERIMENTAL BACKGROUND ON IGNITION AND FLAME

DEVELOPMENT ......................................................................................................... 10

2.1 Ignition by Sparks .............................................................................................. 10

2.2 Simpliﬁed Ignition Analysis .............................................................................. 11

2.2.1 Laminar Flame Speed ............................................................................. 13

2.2.1.1 Factors affecting Flame Speed .......................................................... l6

2.2.1.1.1 Pressure ................................................................................ 16

22.1.1.2 Temperature .......................................................................... 16

2.2.1.2 Relation between Flame Speed and Critical Radius ..................... 17

2.2.1.3 Heat Loss Effect on the Flame Speed .......................................... 19

2.2.1.4 Flame Stretch Effect .................................................................... 22

2.2.2 Turbulent Flame Speed .......................................................................... 22

2.3 Burn Duration .................................................................................................... 24

2.4 Effect of Flame Kernel Growth .......................................................................... 26
CHAPTER 3

EXPERIMENTAL APPARATUS AND PROCEDURE ................................................ 27

3.1 experimental Setup ............................................................................................ 27

3.1.1 Test Equipment Calibration .................................................................... 29

3.1.2 Recorded Data ........................................................................................ 30

3.2 Test Procedure ................................................................................................... 30

3.2.1 Air-Fuel Ratio ........................................................................................ 32

3.2.2 Two Stroke Chainsaw Engine ................................................................. 32

3.2.3 Four Stroke Utility Engines ................................................................... 33
CHAPTER 4

TEST RESULTS ........................................................................................................... 35

4.1 Overview ........................................................................................................... 35

4.2 Two Stroke Homelite 2 Chainsaw Engine .......................................................... 36

4.3 Four Stroke Tecumseh Sidevalve (VLV60) Utility Engine ................................. 38

4.3.1 Effect of Ground Electrode Geometry .................................................... 39

4.3.2 Effect of Ground Electrode Orientation .................................................. 41

vi

4.3.3 Inﬂuence of Spark Gap ........................................................................... 44

4.4 Four Stroke Tecumseh Overhead Valve (OHHSO) Utility Engine ...................... 45
4.4.1 Effect of Ground Electrode Geometry .................................................... 46
4.4.2 Effect of Center Electrode Geometry ...................................................... 47
4.4.3 Effect of Engine Speed on Engine Emissions ......................................... 50
CHAPTER 5
DISCUSSION OF RESULTS ....................................................................................... 52
5.1 Effect of Ground Electrode Geometry ................................................................ 52
5.2 Effect of Ground Electrode Orientation .............................................................. 55
5.3 Effect of Center Electrode Geometry ................................................................ 60
5.4 Effect of Spark Gap ........................................................................................... 61
5.5 Effect of Engine RPM on Comparative Emissions of Two Spark Plugs .............. 62
CHAPTER 6
CONCLUSION ............................................................................................................. 64
Appendix A
FACTORS EFFECTING BURN DURATION .............................................................. 66
A.1 Engine Speed .................................................................................................... 66
A2 Equivalence Ratio ............................................................................................. 66
A.3 Flow Field Effects ............................................................................................. 67
AA Effect of Swirl and Tumble ............................................................................... 67
Appendix B
FACTORS EFFECTING THE FLAME KERNEL GROWTH ...................................... 69
B.1 Ignition Energy and Spark Duration .................................................................. 69
8.2 Effect of Electrical Energy on Flame Growth .................................................... 7O

vii

83 Heat Losses to Spark Plug Electrodes ............................................................... 70

B.4 Flame Kernel Displacement .............................................................................. 70
B.5 Characteristics of Charge Motion ...................................................................... 71
8.6 Effect of Convection Velocity ........................................................................... 71
Appendix C
ENGINE DATA ............................................................................................................ 73
Appendix D
OVERVIEW OF EMISSION PANEL AND DYNAMOMETER .................................. 75
DJ Dynarnometer ................................................................................................... 75
D2 Emission Measurements .................................................................................... 75
D.2.1C0 and C02 Measurements ....................................................................... 75
D.2.1HC Measurements .................................................................................... 76
D.2.2N0Jr Measurements .................................................................................... 76
Appendix E
EXPERIMENTAL RESULTS IN TABULATED DATA FORM .................................. 77
El Ground Electrode Geometry .............................................................................. 77
E. l .1 Two Stroke Chainsaw Engine ................................................................... 77
E.1.2 Four Stroke Tecumseh Side valve (VLV60) Utility Engine ....................... 78
E. l .3 Four Stroke Tecumseh Overhead valve (OHHSO) Utility Engine .............. 79
E.2 Effect of Ground Electrode Orientation ............................................................. 80
E3 Effect of Center Electrode Geometry ................................................................ 82
E4 Inﬂuence of Spark Gap on Emissions ................................................................ 83
E5 Effect of Engine Speed ..................................................................................... 83
APPENDIX F
Effect of Use of Indexing Washers on Compression Ratio ............................................. 85

viii

LIST OF REFERENCES ............................................................................................... 86

ix

LIST OF TABLES

TABLE 1: Typical Levels of Exhaust Emissions for Automotive SI engine ..................... 1
TABLE 2: Tables of Parameters for Equation 24 ........................................................... 19
TABLE 3: Two Stroke Chainsaw Engine Data .............................................................. 73
TABLE 4: Four Stroke Tecumseh Side Valve (VLV60) Engine Data ............................ 73
TABLE 5: Four Stroke Tecumseh Overhead Valve (CHI-160) Engine Data ................... 74

TABLE 6: Effect of Three Different Spark Plug on Two Strokes Engine Characteristics77
TABLE 7: Effect of Ground Electrode Geometry on Four Stroke Emissions ................. 78
TABLE 8: Effect of Ground electrode Geometry on Overhead Valve Engine ................ 79
TABLE 9: Effect of Ground Electrode Geometry (Pyrotek) Spark Plug at Part Load ..... 80

TABLE 10: Effect of Ground Electrode Geometry (Baseline) Spark Plug at Part Load.. 80

TABLE 1 1: Effect of Ground Electrode Geometry (Pyrotek) at Idle .............................. 81
TABLE 12: Effect of Ground Electrode Geometry (Baseline) at Idle ............................. 81
TABLE 13: Effect of Center Electrode Geometry (V-Notch) on Emissions ................... 82
TABLE 14: Influence of Different Spark Gaps on Emissions ........................................ 83

TABLE 15: Effect of Engine RPM on Exhaust Emissions with Pyrotek Spark Plug ...... 83

TABLE 16: Effect of Engine RPM on Exhaust Emissions with Baseline Spark Plug ..... 84

LIST OF FIGURES

FIGURE 1 : Trends in Hydrocarbon Emissions with Variation in Engine Parameters ..... 4
FIGURE 2 : Heat Balance Diagram .............................................................................. 12
FIGURE 3 : Lay Out of Test Equipment ....................................................................... 28
FIGURE 4 : Output Data File Indicating Measured and Calculated Quantities .............. 31
FIGURE 5 : HC vs % Throttle Setting for Three Spark Plugs ....................................... 37
FIGURE 6 : Torque vs Three Spark Plugs .................................................................... 37
FIGURE 7 :NOx vs Three Spark Plugs ........................................................................ 38
FIGURE 8 :HC Emissions vs % Load ........................................................................ 40
FIGURE 9 : AFR vs % Load ........................................................................................ 40
FIGURE 10: NOx vs % Load ........................................................................................ 41
FIGURE 11: Spark Plug 0° Orientation ......................................................................... 42
FIGURE 12: Layout of Combustion Chamber ............................................................... 42
FIGURE 13: Orientation Angle vs HC Emissions at Part Load ...................................... 43
FIGURE 14: Orientation Angle vs Hydrocarbon Emissions at Idle ................................ 44
FIGURE 15: Spark Gap vs HC Emissions ..................................................................... 45
FIGURE 16: HC vs % Load .......................................................................................... 47
FIGURE 17: NOx Emissions vs % Load ....................................................................... 47
FIGURE 18: Spark Patterns for Two Spark Plugs .......................................................... 48
FIGURE 19: Center Electrode Geometry and Spark Pattern .......................................... 49

xi

FIGURE 20: HC Emissions vs % Load ......................................................................... 50

FIGURE 21: NOx Emissions vs % Load ....................................................................... 50
FIGURE 22: Hydrocarbon Emissions vs Engine RPM ................................................... 51
FIGURE 23: Ground Electrode Geometry ..................................................................... 55
FIGURE 24: Ground Electrode Base ............................................................................. 55
FIGURE 25: Cartoons of the Inﬂuence of Plug Orientation on Flame Front Shape ........ 58
FIGURE 2: Center Electrode Geometry ......................................................................... 61

xii

NOMENCLATURE

En ish S bols

m Ms

A Area

A 1 Flame surface area

Ac Contact area between the flame front and electrode
Bm,B¢ Parameters of equation (24)

b Combustion chamber bore

C Capacitance

c Speed of sound

C,- Constant speciﬁc to the engine
E Energy discharged

Ea Activation energy

Minimum ignition energy

xiii

EELS

F arad

m/ sec

Joules

J/Kmole

hc Enthalpy of combustion or enthalpy

h Hi ght of the combustion chamber

K Stretch factor

10 Integral scale or Taylor macro scale

m' Mass ﬂow rate

m .. Volumetric mass production rate

Ma Markstein Number

Q’” Volumetric energy generation rate
Radius or specific gas constant

Rm, Critical radius

Rf Flame kernel radius

S1 Laminar ﬂame speed

S, Turbulent ﬂame speed

T Temperature

u’ Turbulence intensity

V Voltage

V Volume

W Volumetric mass consumption rate

Y R Mass fraction of reactants

Greek Smbols

0: Thermal diffusivity

,6 Pressure exponent

xiv

J/K g

m
kg/ sec
Kg/s-m3
W/m3

m or J/Kg-K
m

m

m/s

m/s

$391

Plus»

Acronyms

CO

C 02
COV
HC
IMEP
N0Jr

OHV

Laminar ﬂame thickness

Equivalence ratio

Parameter defined in Equation (23)
Temperature exponent

Thermal conductivity

Crank angle

Crank angle at start of combustion

Total combustion duration in crank angle
Density

Unburned gas density

Burned gas density

Unburned gas density at the time of ignition

Characteristic burn time

Air Fuel Ratio

Carbon Monoxide

Carbon Dioxide

Coefficient of Variance
Hydrocarbon

Indicated Mean Effective Pressure
Oxides of Nitrogen

Overhead Valve

XV

W/m-K
Degrees
Degrees
Degrees
Kg/m3
Kg/m3
Kg/m3
Kg/m3

Sec

Kgair/Kg Fuel

SI

Spark ignition

Top Dead Center

Unburned hydrocarbons
Vector Lightweight Vertical

Wide Open Throttle

xvi

CHAPTER I

INTRODUCTION

1.1 Overview

The performance levels required of gasoline engines have been rising with each
passing year. Recently, attention has been focused on factors improving performance and
emissions in small utility Spark Ignition (SI) engines, with trends towards the use of
catalytic converters, leaner air/fuel ratios, etc. In the case of the smaller utility engines,
which tend to operate at rich air-fuel ratios, the exhaust-gas emissions remain a major
cause for concern. For reference purposes, typical values of SI exhaust emissions from an

automotive engine are given at Table 1[1].

Table 1: Typical Levels of Exhaust Emissions for Automotive SI Engines

 

 

 

 

 

 

 

 

Pollutant Emissions(kg/ 1 000 liters)
Aldehydes 0.5
Carbon monoxide (CO) 276
Hydrocarbons (HC) 24
Oxides of Nitrogen (NOx) l4
Particulate 1.4
Organic Acids 0.5
Sulfur oxides 1.1

 

 

 

 

In engine combustion, the ﬁrst important stage of the combustion process is the
ignition. In many investigations conducted to date, attempts have been made to improve
combustion in the engines by enhancing ignition. Using more energetic sparks in the
engine’s ignition system can reduce exhaust emissions and fuel consumption.
Enhancement of ignitability depends on a large number of parameters such as: spark
energy, mixture composition, initial pressure and temperature, oxidation kinetics and

bulk ﬂuid motion in the vicinity of spark gap.

In this study, the performance of the Pyrotek spark plug (with a crown-shaped
ground electrode geometry) relative to a Champion spark plug (baseline) was compared
in terms of exhaust emissions. The Champion spark plug is representative of all baseline
spark plugs because the geometry of the ground electrode is essentially identical to all the
other brands. The comparison was made by carrying out a series of laboratory
experiments, in which engines operated at the same controlled conditions were tested

ﬁrst with one spark plug, then the other.

The effectiveness of each spark plug was measured in several engines in terms of
changes in hydrocarbon and N0, emissions. In general, increases in N0, emissions
correspond to increases in burning rate as a consequence of a larger initial ﬂame kernel. It
has been observed experimentally that larger initial ﬂame kernels cause higher burning

rates, torque and N0Jr emissions, and lower the hydrocarbon emissions [20].

The hydrocarbon emissions are mainly due to ﬂame quenching in crevices. A

faster reaction causes higher ﬂame temperatures with higher N0, emissions. However,

higher heat losses may lower the relative N0, emissions. The mechanism of formation of

these noxious emissions is discussed in the following section.

1.2 Formation of Noxious Exhaust Emissions

1.2.1 Carbon Monoxide

At the high temperatures and pressures during the combustion process, signiﬁcant
quantities of C0 form, even when there is sufﬁcient oxygen for complete combustion to
occur. This C0 formation is due to a dissociation reaction, which can lead to equilibrium
compositions of the products of combustion having signiﬁcant CO at high temperatures.
The majority of the rate equations involved in the C0 formation and oxidation to C02

proceed sufficiently fast that equilibrium occurs within a few crank-angle degrees.

Once the exhaust valve opens, measurements indicate [1] that the rapid fall in
pressure and temperature, due to exhaust blow down, may cause some departure from
equilibrium. However, the reaction rates for the above process are fast enough that, even
here, equilibrium is maintained. As the mixture becomes richer, the CO concentration
increases at a progressively faster rate and rapidly becomes quite signiﬁcant due to
insufﬁcient oxygen to complete the combustion. Therefore, ﬁnal C0 formation is

predominantly controlled by the equivalence ratio.

1.2.2 Unburned Hydrocarbons

Most of the initial unburned hydrocarbons exist in the quench zone. The quench

zones are deﬁned as the regions where the ﬂame can not be supported. Crevice regions

such as the space above the upper piston ring and within the spark plug (area between
insulator nose and internal seal) are the important quench areas, but some quench areas
exist adjacent to walls of the combustion chamber, on account of wall heat transfer and
their inhibition of ﬂame-front propagation. In addition, unburned hydrocarbons result
from absorption and desorption of unburned fuel from engine oil at the cylinder wall

during a single cycle.

03C

 

V

 

Increasing compression ratio

OIC

\

Increasing engine speed

 

 

Figure 1: Trends in Hydrocarbon Emissions with Variation in Engine Parameters [l]

Unburned hydrocarbons from within the quench regions in engines are expelled

during the exhaust process. The lowest levels of unburned hydrocarbons tend to occur at

an equivalence ratio slightly less than one. The major engine variables affecting the
unburned hydrocarbons in SI engines are equivalence ratio, compression ratio, engine

speed, and spark timings.

A higher compression ratio increases unburned hydrocarbons because, with a
smaller combustion space at Top Dead Center (T DC), the quench zone comprises a larger
proportion of the total volume. Unburned hydrocarbons decrease with increasing engine
speed because the correspondingly higher turbulence levels promote faster combustion
and decrease the size of the quench zone. In the case of a retarded spark, the unbtn'ned
hydrocarbons decrease because more combustion occurs after TDC, when the
combustion volume is enlarging and the quench zone is proportionally smaller.
Moreover, retarding the spark results in exhaust gases at higher temperatures so that more

oxidation of HC can take place in the exhaust system.

1.2.3 Oxides of Nitrogen

The formation of oxides of nitrogen is a highly temperature dependent
phenomenon. It occurs because equilibrium concentrations of various N0, compounds
form when oxygen and nitrogen are mixed at high temperatures such as 2000°K to
3000°K. However, the reaction rates are slow relative to engine-combustion time scales
and equilibrium is not fully attained in the time available under most engine conditions.
This applies, to some extent, to the forward reaction but is particularly true of the

backward reaction.

0+ N2 <=> N0+N
(1)

The NO, is effectively from for a long period after it is exhausted from the
engine, giving it time to react with other substances to form photochemical smog. The
rate of NO, formation is also coupled to the levels of tm'bulence in the ﬂow. At peak
combustion temperatures, small changes in temperature correspond to large differences in
the amount of N0, formed. This is predominantly due to the observation that combustion

is faster at higher turbulence levels and reaches higher peak temperatmes[ l ].
1.2 Literature review

It is a well-known observation that the geometric details of the spark plug
electrodes affect ignition and early ﬂame grth processes. There has been very little
quantiﬁcation of this observation. In a spark ignition engine, one is interested in the
effectiveness of the spark plug in igniting a mixture and ensuring a faster burn rate.
Many studies have been carried out to gauge the mixture ignitabilities as measures of
spark-plug performance. One of the methods used by Daniel and Scilzo [2] was to
measure the Coefﬁcient of Variance (COV) of Indicated Mean Effective Pressure (IMEP)
at idle and lean fuel conditions. Nishio, Oshima, Kyongdoung and Heywood [3] have
used Air to Fuel (A/F) ratio lean limit as the lean limit comparison test, to determine the
leanest mixture that can be ignited by a spark plug. Hood [4] used the counting method to
count lean limit misﬁres during a certain interval of time, in order to determine which
spark plug shows a better performance in terms of igniteability. In addition, maximum
Exhaust Gas Recirculation (EGR) tolerance has also been used as a parameter to measure

the ignitability by the spark plug.

Piston motions in engines impose constraints on ignition and ﬂame propagation
via pressure and temperature in the end gas on the one hand and charge motion and
turbulence on the other. Maly, Saggau, Wagner and Ziegler [11], while evaluating the
prospects of ignition enhancement, concluded that these constraints may be overcome by
improved ignition devices, which provide the largest possible expansion velocities and

initial sizes of the ﬂame kernel.

Brereton, Bertrand and Macklem, [13] while evaluating the effects of changing
humidity and temperature on the emissions of small utility engines, noticed the strong
inﬂuence of varying humidity on H C emissions. Hadjiconstantinou, Kyoungdoug and
Heywood [5] correlated the ﬂame propagation characteristics with hydrocarbon
emissions. They concluded that mean engine hydrocarbon emission levels increased
when the air to fuel ratio increased above its stoichiometric value because of decreasing
HC oxidation. Furthermore, they noticed that for stoichiometric mixtures, there was no
correlation between the individual cycle bm'n rate and HC emissions because changes in

the ﬂame speed and exhaust gas temperature cancelled each other.

Mantel [21] showed that in the presence of a mean velocity ﬁeld, the development
and propagation of the ﬂame depended largely on the ﬂow ﬁeld induced by the ground
electrode in the vicinity of and particularly in the wake of the spark gap. Bianco, Cheng
and Heywood [6] characterized the behavior of the ﬂame kernel by an expansion speed
which describes its growth rate, and a convection velocity, which describes its overall

movement.

Herweg, Begleris, Zettlitz, and Zeilger [8] used high speed Schlieren photography
in an optically accessible side chamber and found that the ﬂame kernel is changed by the
turbulence at ﬂame radii of 0.5 to 1 mm, depending on turbulence intensity. The
inﬂuence of turbulence on the turbulent burning velocity during the ﬂame kernel
formation is less compared to the main combustion period. Hall [9] noticed that
turbulence of higher intensity and smaller scale enhances the rate of ﬂame kernel growth,
even for relatively young ﬂames (having a diameter of up to 1 cm), and for mixtm'es

having an equivalence ratio as lean as 0.7.

A number of studies have been conducted on the effect of ground electrode
orientation. Anderson and Asik [25] found experimentally that the orientation of the
conventional spark plug ground electrode, with respect to the mean swirl velocity, can
have a quantiﬁable effect on combustion stability and average burn time. Positioning the
ground electrode upstream of the oncoming swirl ﬂow resulted in improved ﬂame
initiation over a downstream, or 180 degree rotated position. Zeigler, Schaudt, and
Herweg [26] found improved ﬂarne initiation with the ground electrode positioned

dowstream of a uniform in a ﬂow combustion chamber.

Witze [7] investigated the effect of the spark plug location on the burning rates
with a disc shaped combustion chamber. He deduced that, in general, ignition at the
center of the chamber was preferable. While evaluating the effect of spark duration on
combustion, Nakai, Nakagawa, Hamai, and Sone [12] found that lengthening the spark
duration helped to promote ﬂame kernel growth. This was attributed to the continued
supply of spark energy to the ﬂame kernel as it is being formed during the initial stage of

ignition.

Heat losses from the spark-ignited ﬂame kernel to the electrode play an important
role in SI engine combustion. Pischinger and Heywood [24] had deduced from
experiments using high speed photography that the heat losses from the electrodes can
promote cycle to cycle variations in the initial ﬂame kernel growth. The heat losses vary
from cycle to cycle, due to substantial cycle to cycle variation in the contact area between
the ﬂame kernel and the electrodes. Anbarasu and Abata [10] investigated the effect of
heat loss to the spark plug during the early ﬂame growth by modeling ﬂame initiation and
ﬂame propagation around the spark plug. They found that the higher ﬂame travel
velocities decreased the heat loss rate and this effect becomes less signiﬁcant as the ﬂame

travel velocities are increased.

1.3 Objectives

Efforts to reduce the exhaust emissions from small utility engines, which
contribute a signiﬁcant proportion to the increase in atmospheric pollution, are presently
in introductory states. This study contributes to the ongoing efforts to reduce small engine
emissions by evaluating the effect of different ground electrode geometries on exhaust

emissions. The objectives of this thesis are:
1. To assess the comparative effectiveness of the Pyrotek spark plugs relative to
conventional (Champion) spark plugs by measuring exhaust emissions

produced by several small utility engines in controlled experiments.

2. To improve understanding of how and why spark-plug geometry affects the

exhaust-gas emissions of such engines.

CHAPTER 2

THEORETICAL I EXPERIMENTAL BACKGROUND ON IGNITION AND

FLAME DEVELOPMENT

2.1 Ignition by Spark

Ignition of combustible mixtures can be accomplished by creating a spark by
electrical discharge between two electrodes placed in the mixture. The spark is caused
either by capacitance discharge (which is relatively fast z 0.01usec) or by inductance
discharge. A capacitance discharge is produced by quickly discharging a condenser,
whereas an inductance discharge is produced by opening a circuit, which involves

transformers, ignition coils and magnetos.

If C1 is the capacitance of a condenser, and V1 and V2 are the voltages on the

condensers respectively before and after the spark, the energy discharged is given by [30]
1 2 2
E = 5 C. ( V. - V2 )

(2)

The process of ignition of premixed charges is envisaged as occurring in the
following manner: the passage of the spark raises the temperature of a small, roughly

spherical, volume of air (referred to henceforth as the spark kernel) to sufﬁciently high

10

value to initiate rapid evaporation of any fuel drops contained within this volume. The
subsequent reaction rates and mixing times of air and ﬁre] are assumed inﬁnitesimally
fast, so any fuel vapor created within the spark kernel is instantly transformed into
combustion products at the (assumed) adiabatic ﬂame temperature. If the rate of heat
release by combustion exceeds the rate of heat loss by thermal conduction at the surface
of the inﬂamed volume, then the spark kernel grows in size to ﬁll the entire combustion
volume. If, however, the rate of heat release is lower than the rate of heat loss, the
temperature within the spark kernel falls steadily until fuel evaporation on the unburned

side of the ﬂame front ceases altogether [32].

The size of the spark kernel is of crucial importance, since the rate of heat loss at
the kernel surface is just balanced by the rate of heat release, due to the instantaneous
combustion of fuel vapor, throughout its volume. This concept leads to the deﬁnition of
“quenching distance” as the critical diameter that the enﬂamed volume must attain to
propagate unaided. We also deﬁne the minimum ignition energy as the amount of energy

required from an external source to attain this critical size.

2.2.1 Simpliﬁed Ignition Analysis

According to the Williams’ second criterion for ignition and quenching, the rate
of liberation of heat by chemical reaction inside the volume must approximately balance
the rate of heat loss from the volume by thermal conduction. We can apply this criterion
to the ﬂame kernel, which represents the incipient propagation ﬂame created by a point
spark. We deﬁne a critical gas-volume radius such that a ﬂame will not propagate if the

actual radius is smaller than the critical value and also assume that the minimum energy

11

supplied by the spark is the energy required to heat the critical volume from its initial
state to the ﬂame temperature. To determine the critical radius (Rwy), we equate the rate
of heat liberated by the reaction rate to the heat lost to the cold gas by conduction, as

shown in the Figure 2.

 

Heat conducted

 

 

Heat
generated

 

Figure 2: Heat Balance Diagram

Q," V = Qcond
(3)
where Q"”is the volumetric heat release rate and
Q5": _ mgrAhc
Also (4)
V = (4/3)7tR3.
Similarly from the Fourier’s law: (5)
Qcand: «(4&371') )d/dr| Reﬁt.
(6)

Here, mp'm is the mass ﬂow rate (kg/m3—s) of fuel, and ,1 is the thermal conductivity of

the burned gas. Hence, equation (3) can be written as:

3”?

mm Hare/3) =-242di3..d%1,l....-
(7)

We have expressed the volume and surface area of the sphere in terms of critical
radius in equation (7). The temperature gradient at the gas boundary can be evaluated by
determining the temperature distribution of the gas beyond the sphere (Rem: r S 00) with

the boundary conditions: Th,“ = T2, and To, =T.. , This leads to the result

.1.) a... = {LR-L)
at!

(3)

Substituting (8) in (7) and simplifying we get
"" -rh;.~"Ah.

(9)

The equation (9) indicates that critical radius strongly depends on the therrno-chemical

properties of the mixture.

2.2.1 Laminar Flame Speed

Spark ignition engines with well-designed intake systems are often good
examples of an application of premixed combustion. The fuel-air mixture is produced by
a carburetor or fuel injection system. Even though the fuel is introduced as a liquid,
because it is highly volatile it has time to vaporize and thoroughly mix with the air,
before the mixture is ignited by the spark. In some engines, however, the use of short
intake systems and close-coupled carburetors inhibits thorough mixing, and combustion

may not be completely premixed.

l3

The laminar burning speed is deﬁned as the velocity relative to the ﬂame ﬁ'ont,
with which unburned gas moves into the front and is transformed to combustion products
under laminar ﬂow conditions. The laminar burning velocity (5;), at pressures and
temperatures typical of unburned gas mixtures in engines, is usually measured in
spherical closed vessels by propagating a laminar ﬂame radially outward ﬁom the vessel

center.

According to the Mallard Le-Chatelier ﬂame theory[40], the premixed ﬂame is
treated as a travelling wave. The propagation process is viewed as a balance between heat
production via chemical reaction and the associated convective and conductive
redistribution mechanisms. Consider a ﬂame cross section joining unburned upstream gas
and downstream burned gas. In the upstream gas, there must be a balance between

downstream convection and upstream diffusion of heat if the ﬂame propagation is steady.

 

Hence
T-
pC.&(T.—7;)=A("5T‘)
or (10)
__1__£_ﬁ.
pCpS, S, :52"

(11)

The ﬂame thickness 6 is usually of the order of 0.1 cm. The mass burning rate can be

given as

14

(12)

where [W] = mass/voI—sec. In other words, the mass ﬂux into the ﬂame equals the
volumetric mass consumption rate W multiplied by the ﬂame thickness 6 over which the

consumption occurs. Substitution of

5"5’7W

 

gives (13)
ng = AW: 512 = ’1 .—
Cp pC, p
or (14)
S, = aORR,

(15)

where a = A / p Cp is the thermal diffusivity and RR = W /p is the reaction rate. We can
evaluate a in the unburned gas, W at the ﬂame temperature, and p (in RR) at the

unburned state. We may write

Therefore, S; can be rewritten as, (16)

15

/ .. 3%.
.g,=: EZZ:£_____,
Pu

The laminar ﬂame speed from the above equation indicates a strong dependence on the

(17)

thermal diffusivity and the reaction rate. The Mallard Le-Chatelier model is useful in
estimating the order of the magnitude of ﬂame speeds, but, in reality, the reaction zone is

much more complicated than the Mallard Le-Chatelier model indicates.

2.2.1.1 Factors Affecting Flame Speed

2.2.1.1.1 Pressure.

From equation (17), since a ocp '1 and YR ac p ,

we obtain 5, ac [p " . p ""1”? =p W2“. (18)

But pocP,sothat s,ocP("/2"’ (19)

By varying the overall pressure, the order of the overall reaction n can be deduced from

ﬂame speed measurements.

2.2.1.1.2 Temperature.

From the equation (17), since RR oc e '5’"; , this means that S, cc e 'mkrf. Hence

the laminar ﬂame speed S1 is extremely sensitive to the ﬂame temperatm'e Tf

l6

2.2.1.2 Relation between Flame Speed and Critical Radius

Using a series of assumptions [l4], and assuming a balance between conduction
and convection heat transfer for spherical volume, the critical radius can be related to the

laminar ﬂame speed S; as

(20)

The minimum ignition energy, Eign, can be determined if we assume that the energy
added by the spark heats the critical volume to burned gas temperature and can be written

as

Thus (21)

7t
Eran = 43pr3'C (Ti: - 7:.)

 

on! p
(22)
Substituting (19) into (21), we get
C _ 3
W(—”)( 712)
Rb T. S,
(23)

The energy supplied by the ignition systems of SI engines is always greater than the
minimum ignition energy required for the successful ignition of the mixture. The initial

ﬂame grows under the inﬂuence of the following parameters: energy supplied through the

17

spark, heat loss to the electrodes, the effect of strong curvature due to the small radius,

and the turbulence in the ﬂow ﬁeld.

Meghalchi and Keck [31] experimentally determined the laminar ﬂame speeds for
various fuel-air mixtures over a range of temperature and pressures typical of conditions
associated with reciprocating internal combustion engines. They calculated the burning

velocities in a bomb from implied burn rates under the following assumptions:
0 The unburned gas is initially at rest and has uniform temperature and composition.

0 The thickness of the reaction zone is negligible, and gas within the bomb consists of a
burned fraction x at local thermodynamic and chemical equilibrium, and unburned
fraction l-x at local thermodynamic equilibrium, but with ﬁxed chemical

composition.
0 The pressure is independent of position and a ﬁmction of time only.

0 The reaction front is smooth and spherical.

5: = 51.0(72/ To)"(p/po)ﬂ,
(24)

where

To = 298 k
p0 = I atrn
T u = temperature of the unbm'ned mixture

P = pressure of the unburned mixture

18

S [,0 = constant for a given fuel, equivalence ratio, burned gas dilution

fraction

for iso-octanes, these constants can be represented by

a=2.18—0.8 (45-1), (25)
p = - 0.16 + 0.22(¢- 1), (26)
S 1.0 = 3., + BM M. (27)

where ¢,,, is the equivalence ratio at which S [,0 is maximum with value B,,,.

Table 2: Table Of Parameters For Equation 24

 

 

Fuel ¢.. Bm(cm/sec=) B, (cm/sec)
Iayqxnane 1J3 263 -ai7*
Gasoline 1.21 30.5 -54.9

 

 

 

 

 

 

 

2.2.1.3 Heat Loss Effect on the Flame Speed

The effect of heat losses to the electrodes on the laminar ﬂame speed can be

modeled in the following way:

AS: = Smart - Simonadia

(28)

19

OI'

Qcomb = he dnydt = hcpuSlAf

'Thus (29)

 

C

S ._
[m puAfh

(30)

The heat transfer from the ﬂame kernel to the electrodes will be convective as
well as conductive. Radiation processes are neglected as radiation from ﬂames is mainly
due to infrared emissions from C02 and H20 bands [15]. These infrared emissions can
not be absorbed by the unburned gas as, C02 and H20 are not available ahead of the
ﬂame front [15]. In addition, heat loss can be estimated using the lumped heat transfer

coefﬁcient ‘h’, and is given by

(31)

where, A, is the contact area between the ﬂame and electrodes, T1, is the gas temperature,
and T1 is the electrode temperature. The effect of this heat transfer on the laminar ﬂame

speedcanbewrittenas

(32)

20

This indicates that with the increase in the heat losses from the ﬂame kernel, the
laminar ﬂame speed decreases. This reduction in the laminar ﬂame speed strongly
depends on the contact area fraction fc, deﬁned as the ratio of the contact area Ac to ﬂame

surface area Af. (a sphere of radius r; ).

(33)

This heat loss from the ﬂame kernel affects the ﬂame kernel growth through two
mechanisms:
0 It decreases the kernel temperature leading to a relative contraction of the kernel, due
to its lower density.

0 It transfers heat ﬁom the ﬂame ﬁ'ont, thereby decreasing the burning velocity.

In order to have a signiﬁcant impact on the ﬂame growth, the heat loss from the
initial ﬂame kernel has to be comparable to the energy gain rate from combustion. This

ratio can be estimated by using a laminar burning speed S].

g...__ MAY; -T.)

Qeamb puAfSlhc

(34)

The ratio in equation (34) has to be greater than one, in order to have a successful

ﬂame initiation.

21

2.2.1.4 Flame Stretch Effect

The effects of ﬂame stretch are to reduce the area available for ﬂame propagation
and also to increase the heat losses from the stretch surface of the unburned mixture [29].
The strong curvature of a spherical ﬂame kernel may also considerably slow the ﬂame
speed. Markstein [30] suggested a linear dependence of laminar ﬂame speed on ﬂame

stretch given by:

S1,: : Shadi - MaaK’
(35)

where:

Ma is the Markstein number

S 1,, is the adiabatic stretched laminar ﬂame speed

K is the stretch factor and
K = (2/r) (dry /dt), for a spherical ﬂame propagation, (rf) being the ﬂame kernel radius.
This equation indicates that larger initial ﬂame kernels will have smaller stretch factors,

and therefore, a lower reduction in laminar ﬂame speeds.

2.2.2 Turbulent Flame Speed.

Whereas the laminar ﬂame speed has a propagation velocity that depends
uniquely on the thermal and chemical properties of the mixture, a turbulent ﬂame has a
propagation velocity that depends on the characteristics of the ﬂow, as well as the
mixture properties. A simple theoretical formula proposed by Damkohler for the

turbulent ﬂame speed is given by

s=s+w,
(36)

22

where u’is the scale of turbulent velocity ﬂuctuations (m/s).

The ﬂame is assumed to spread by the process of turbulent entertainment, with
burning occurring in the entrained region at a rate controlled by the tm'bulent parameters.
The ﬂame front is assumed to entrain mixture at a rate that is governed by the turbulent

intensity and the local laminar ﬂame speed. Thus, the burning rate can be given by

dm ,
—dt_ = AAA“ + SI)
where: (37)

m = mass entrained in to the ﬂame front
)0, = density of the unburned charge
A f: ﬂame front area (excluding area in contact with the electrode surface)

a ’=turbulent velocity factor
Turbulence is typically inhomogeneous and always time varying. Chen and
Veshagh [17] assumed the following to express turbulent intensity:

0 The turbulent ﬂow is isotropic and homogeneous over the combustion period near the

compression.

0 The turbulence velocity factor is proportional to the mean piston speed and does not

change substantially over the combustion period:

'-
u—ch

(33)

23

where Vp is the mean piston speed and c is the constant, which depends on the design

of the inlet port, the inlet valve shape, and many other factors.
0 The integral length scale (10) retains a constant value over the combustion period. This
assumption is roughly correct, as some studies suggested that the integral scale almost

maintains a constant value over the compression TDC position [16].

Semenov [36] suggested the following expression for turbulence intensity, as

governed by the conservation of angular momentum:

1/3
#4:")
ng

Here, the subscript ig is for the values at the time of ignition and subscript u is for

 

(39)

the values of unburned mass. Although many different theoretical formulae have been
developed to predict the turbulent ﬂame speed, there appears to be general agreement
that turbulent ﬂame speed depends only on turbulent intensity and does not involve any

other turbulence property, such as the such as length scales, etc.

2.3 Burn Duration

The rate at which fuel-air mixtures burn increases from a low value, immediately
following the spark discharge, to a maximum, about half way through the burning
process. This burning rate then decreases almost to zero as the combustion process ends.
It is convenient to use mass-fraction burned to characterize different stages of the spark
ignition engine combustion process by their duration in crank-angle degrees, thereby

deﬁning the fraction of the engine cycle that they occupy. The ﬂame development

24

process from spark discharge, which initiates the combustion process to the point where a
small but a measurable traction of the charge has burned, is one stage. It is inﬂuenced
primarily by the mixture state, composition, and the motion in the vicinity of the spark
gap. The major portion of the charge burns as the ﬂame propagates to the chamber wall,
in the second stage. This stage is obviously inﬂuenced by thermodynamic conditions
throughout the combustion chamber. The ﬁnal stage, when the remainder of the charge
burns to completion, can not be quantiﬁed easily because energy release rates are

comparable to other energy transfer processes that are occurring.

A functional form often used to represent the mass fraction burned versus the

crank angle curve is the Weibe function.

n+1
xb = 1— exp[—a( 6;? ) ]

 

(40)

Here, 9 is the crank angle, 00 is the start of combustion, A6 is the total combustion
duration (X5 = 0 to X), = 1), and a and m are adjustable parameters. Varying a and m
changes the shape of the curve signiﬁcantly, actual mass-fraction-burned curves have
been ﬁtted with 0 =5 and m=2. The crank angles can be converted to time (in seconds) by

dividing by 6N (with N in rev per min).

It has been observed experimentally [24] that dming cycles in which the ﬂame
kernel moved towards the center of the combustion chamber, the burn duration was
shorter than during cycles in which it moved away from the center of the cylinder. The

factors affecting the burn duration are given at Appendix A.

25

2.3 Effect of Initial Flame Kernel Growth

The initial ﬂame kernel grth has a strong inﬂuence on the overall combustion
process in the IC engines. The parameters governing the initial ﬂame kernel growth and
subsequent ﬂame propagation are:

0 The ignition system, which includes the spark duration and ignition energy.
0 Effect of electrical energy on ﬂame growth.

0 Heat losses to the electrodes.

0 Flame kernel advection from the spark gap.

0 Characteristics of charge motion.

0 Effect of convection velocity.

The details of these parameters are discussed in Appendix B.

26

CHAPTER 3

EXPERIMENTAL APPARATUS AND PROCEDURES

3.1 Experimental Set Up

In order to evaluate the comparative effectiveness of the Pyrotek spark plug and a
base-line Champion spark plug, emission measurements were carried out for three
different types of test engines. The details of these test engines are listed in Appendix C.
The emission measurements of these engines were carried out in the sequence described

below.

The two stroke Homelite Super 2 chainsaw engine was mounted on a Homelite
eddy current dynamometer in a test cell. The ambient temperature in the test cell was held
within a few degrees of constant value and was always between a 20°C to 30°C limit.
The four stroke Tecumseh VLV60 side valve engine and Tecumseh OI-IHSO overhead
valve engines were mounted on a Micro—Dyn 15 hydraulic dynamometer, in the same test
cell as the chain saw engine. The four stroke engines were mounted on a hydraulic
dynamometer and could be run at a constant programmed speed, while the torque output

produced by the engines was measured.

Each engine was run according to the appropriate EPA CFR test procedure, using

Indolene (or Indolene mixed 32:1 with two-stroke oil for two-stroke engine). This fuel

27

was supplied to the engine by a fuel cart, which measured the mass ﬂow rate of fuel
consumed by the engine. On this cart, 3 Micro Motion Coriolis meter D800068100
carried out the fuel ﬂow measurement. Fuel was supplied to each engine at a constant
pressure with the level of fuel at the height of the half-ﬁrll level of the original fuel tank
of each engine. The engine exhaust gases were ducted though a dilution ttmnel positioned
in a manner such that all the exhaust products and sufﬁcient quantities of ambient

dilution air were drawn through it.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- Fuel art _’ Engine >
D t Dilution
ynamomeer
Dynamomctcr _, tunnel
controls PC
1 r
, DatapocessingPC HC NOX C0 C02 Exhaust
11
Exhaust to
atmosphere

 

 

Figure 3: Layout of Test Equipment

28

The intake of the dilution tunnel was positioned slightly away from the mufﬂer to
avoid any chances of creating unnecessary vacuum or additional suction on the muffler.
This diluted mixture of exhaust gases was then drawn through a critical ﬂow nozzle by a
ring compressor. The critical ﬂow nozzle regulates the total ﬂow rate of ambient air and
exhaust gases. The size of the critical ﬂow nozzle was changed between the tests in order
to make best use of the measurement range of the emission analyzer. A partial sample of
this diluted mixture was fed to the emission analyzers, while remaining gases were
expelled from the test cell. In addition, during periods of exhaust-gas sampling, the
background level of ambient air were also recorded and subtracted from the ﬁnal data

output.

Diluted exhaust-gas samples were analyzed continuously throughout each test for
their C0, C02, HC (propane equivalent), and N0, content by a Horiba Model 11136-1
emission bench. The details of the layout are given in Figure 3. A brief overview of the
working principles of the Horiba emissions bench is given at Appendix D and more

complete details are given in [23].
3.1.1 Test Equipment Calibration

The emission analyzers were calibrated before the start of tests using a gas divider
at 10, 20, 30, 40, 50, 60, 70, 80, and 90% settings between the span and zero
concentrations. While the C02, HC and N0, responses were linear to better than the
required 2% accuracy throughout their range, the response of the C0 analyzer was ﬁtted

with a fourth- order polynomial to achieve better than 2% accuracy.

29

The torque calibration on the dynamometer was performed by adding different
weights at a distance of 45.720m (18 inch) from the torque-sensor load cell. All indicated

torques were within 2% of the applied torque.

By checking the fuel ﬂow through the Micro Motion meter for speciﬁed time, and
collecting and weighing the fuel in the beaker, the calibration of the fuel ﬂow meter was
veriﬁed to better than 1% accuracy. The fuel ﬁlters of the fuel cart were regularly
renewed. Vibrations in the fuel cart attributed to the fuel pumps were minimized by
mounting the pumps externally. Additional ﬁlters were also installed to further damp

vibrations which would otherwise cause inaccurate meter readings.
3.1.2 Recorded Data

In addition to the emission data, a number of other quantities were also measured
along with the exhaust emissions. These included engine speed (RPM), torque output,
power output, the atmospheric and differential pressure, cylinder head temperature, room

and exhaust temperatures, relative humidity, and fuel ﬂow rates.
3.2 Test Procedure

The test procedure detailed in the CFR40 part 90 was followed where applicable.
All the equipment was calibrated prior the commencement of tests. For the gas sampling
test procedure, analyzers were turned on for a period of half an hour to give it a sufﬁcient
time to warm up. Next the analyzers were zeroed, spanned, and tested for the HC hang
up, before all the tests. Background concentrations of the ambient air (dilution air) during
the each test were also taken in order to include the background effect. Every effort was

made to complete each particular test in the same day in order to reduce the variations of

30

the test parameters due to the changes in the ambient conditions. The test engines were
warmed up until the exhaust temperature stabilized at a particular throttle setting or load.
This was done to ensure the same conditions for each test. Before the commencement of
each test, the spark gap of the spark plugs was adjusted to the engine manufacturer
recommendations. A sample data sheet indicating measured and calculated quantities is

given in Figure 4 below.

 

 

Emissions Operating Conditions Temperatures

Fuel ﬂow (gr/hr) 423.7 Speed (rpm) 2797.
HC (C_1) ( ) .1802 Power (brake hp) 1.3883 T_nozzle (‘C) 46.9

 

NOX ( ) .3232 Torque (ft-1b) 2.606 T_cyl (C) 103.0
C02 ( ).5228 Rel. Humidity (%) 15.00 T_exhaust (C) 253.8
CO ( ).3594 T_room (C) 30.5 T_oil (C) 27.6
CH4 ( ).0000 P_diff (kPa) .110 T_fuel (C) 27.3

P_abs (kPa) 97.5

CO(g/kW/hr); NOx(g/kW/hr); HC(g/kW/hr); C02(g/kW/hr)
CO(g/hp/hr); NOX(g/hp/hr); HC(g/hp/hr); C02(g/hp/hr)

54.7 8.0 4.88 1147.10
40.8 5.9 3.64 855.74

CO(grm/hr); NOx(grm/hr); HC(grm/hr); COZ(grm/hr)

56.7 8.2 5.05 1188.02
BSFC (gr/kW/hr): 409.08 Fuel burned (%): 98.81
Dil.Fac 30.66 Dil.Fac (EPA) 29.73 AFR: 13.94
[CO]/[COZ] .0477

 

Figure 4: Output Data File Indicating Measured and Calculated Quantities

31

 

3.2.1 Air-Fuel Ratio

The air fuel ratio was calculated according to Spindt’s method- an equilibrium
combustion model equation involving fuel composition and exhaust gas composition
only. For this method, exhaust gas concentration was determined using the emission
analyzer. Following ratios were calculated using the concentration of the C0, C02, 02

and hydrocarbons (per carbon atom) and known fraction of carbon in the fuel (F c):

R: Pea/Pea, 1
(41)
Q: Poz/PC 1
(42)
F6 =(Pco+Pco,)/(PC0+PC0, +PCH)‘
(43)
Then,
A/F= 1.7,[1 1.492FC(ﬂ/—2+—Q)+1—211—+F—C)]-
1+R 3.5+R
(44)

The accuracy of this method when compared with measured air-fuel ratios was found to
be above 95% even for poor combustion. The comparative results along with detailed

calculation are given in [37].
3.2.2 Two Stroke Chainsaw Engine

The two stroke Homelite Super 2 chainsaw engine was controlled manually by its
throttle cable. The emission measurements were made at two settings: idle and wide open

throttle. The engine was locked at wide-open throttle to test it at rated mode, and locked

32

at the closed (idle) position for the idle mode. In this case, the idle mode was quite
variable from test to test, possibly because of the length of the time required to reach a
steady state. The CFR requirement of steady state, in which the cylinder head

temperature remains within a 10°C band for three minutes, was met in all tests.

A series of idle tests was carried out in which the engine was run with one spark
plug, until a nominally steady state was reached and data was then taken. Then another
idle test was performed with a different spark plug, and so on. It was found that the
sequential tests at one mode were more repeatable. Therefore, all three spark plugs were

ﬁrst tested at idle.

Some minor adjustments had to be made to the engine carburetor in order to ﬁnd
an idle mixture/speed at which the engine would run evenly for all three spark plugs

tested. At wide-open throttle (WOT), no such difficulty was encountered.
3.2.2 Four Stroke Utility Engines

Two different types of four stroke engines were used as described in sect 2.1.
These four stroke utility engines were mounted on the hydraulic dynamometer cart. In
order to control the engine speed and throttle settings more precisely, the engine governor
was disconnected and an external throttle cable was connected instead. The use of the
external throttle cable provided much more stable engine performance than did the engine
governor. Before taking the emission data, each engine was run at its maximum rated
load (maximum power output) at 2800 rpm, and its torque was recorded. Once the WOT
power was known, these engines were run at various ﬁactional loads (100%, 75%, 50%,
25%, 10% of the WOT power achieved) as required by CFR40 part 90. Two different

approaches were adopted for each engine when run at various fractional loads. In the ﬁrst

33

approach, the engine was run at all ﬁve load settings in sequence with the same spark
plug, prior to changing to a second plug. In this approach, the load settings at which the
engines operated could not be repeated precisely, and, consequently, they operated at
slightly different air to fuel ratios. In the second approach, at each setting, the engine was
operated at a given load with each of the spark plugs installed, with no changes made to
the throttle, speed and load setting. The data collected in this approach was found to be
more consistent since, at a given load, the engine operating at essentially the same air to

fuel ratio.

The engines were also run at the idle settings by disconnecting the engine ﬁ'om
the dynamometer. The carburetors of the four stroke utility engine were of ﬁxed jet
design, with no adjustment available. In the idle mode (about 1500 rpm), all three spark
plugs chronically misfnred occasionally, although no chronic misﬁres were observed in
any of the other load settings. The repeatability of tests of the four stroke engines was
higher than the two-stroke engine, with the highest repeatability being for the Tecumseh
overhead valve engine. The entire tests were repeatable within ﬁve percent of any brake

speciﬁc emissions measurement.

Apart from the effect of each spark plug’s different gound electrode geometry on
engine emissions, other aspects of spark plugs on engine emissions were also analyzed.

These included the spark plug orientation, spark gap, center electrode geometry, and

engine speed.

34

CHAPTER 4

TEST RESULTS

4.1 Overview

A comparison between the emissions/performance of three different engines ﬁtted
with: i) a brand-new spark plug of the manufacturer’s recommended Champion model
(the baseline plug); ii) a brand-new Pyrotek (crown-shaped electrode) version of the
baseline plug; and iii) a well-used Champion spark plug was carried out. Effects of
different gound electrode geometries would then be revealed by differences between the
Pyrotek (new) and the baseline (new) spark plugs. Changes in performance through usage
could also be assessed by carrying out the same tests with a well—used Champion
sparkplug, which would be indicative of emissions levels during normal usage of these
engines. The two stroke Homelite Super 2 chainsaw engine (1 hp) engine was tested with
all three kinds of spark plug (baseline Champion model DJ 7Y). The four stroke Sidevalve
(L-head) (Tecumseh VLV60, Champion J 19LM) and overhead valve (Tecumseh
OHHSO, Champion N4C) engines were tested principally with the Pyrotek (new) and
baseline (new) spark plugs, to clearly differentiate the effect of gound electrode
geometry. The results are discussed in the order in which these tests were performed.

Details of these results in tabulated data form are presented at Appendix E.

35

4.2 Two Stroke Homelite Super 2 Chainsaw Engine

Two stroke Homelite Super 2 chainsaw engine used in this study was a well used engine
with over 50 hours of operation. For this engine, the tests performed were basic CFR40
comparisons of the emissions levels of this engine with three different types of spark
plugs, i.e. Pyrotek, a brand new Champion spark plug and a well used Champion spark
plug. In these tests, the engine was operated at ﬁrst at wide-open throttle and then at idle.
Each test was then repeated so that three-test averages could be constructed. Figure 5
shows the comparative level of hydrocarbon emissions for the three spark plugs at these

two throttle positions.

At 6% load (the nominal idle), the difference between emission levels was not
very distinct. However, at WOT, the difference in the emission level was quite
signiﬁcant. There was approximately a 13% reduction in the hydrocarbon emissions
when using the Pyrotek spark plug compared to the brand new spark plug. In addition, a
17% reduction in hydrocarbon emissions was observed when compared with the well
used spark plug. The results in Figure 7 show that NO, emissions were higher for the
pyrotek spark plug as compared to the new and well used Champion spark plugs.
Although the N0Jr emissions were quite low in total, relative to regulatory standards for
small engines, the trend indicates that the Pyrotek plug produces more N0,r with a
subsequent reduction in the hydrocarbon emissions. Moreover, at the same throttle setting
the Pyrotek produces 4.5% and 12% more torque than the new spark plug and the well

used spark plugs respectively as shown in Figure 6.

36

hydrocarbons(Grams/hour)

Torque(Ft-Lb)

Comparative effect of spark plug ground electrode geometry
on hydrocarbon emissions(C FR 90), two stroke Super 2
chainsaw engine

 

 

 

 

 

 

 

 

6% 100%
°/o throttle setting

Figure 5: HC vs % Throttle Setting for Three Spark Plugs

Comparative effect of spark plug ground electrode
geometry on torque levels (CFR90), two stroke
chainsaw engine

0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05

.100%

 

Torque(ft-lb) Torque(ft-lb) Torque(ft-lb)
Pyro New Used
Torque

Figure 6: Torque vs Three Spark Plugs

37

Comparative effect of spark plug ground electrode geometry
on NOx emissions(CFR 90),for two stroke Super 2
chainsaw engine

-6% .1007.

 

Nox (grins/hour) Nox (grins/hour) Nox (grins/hour)

spark plugs

Figure 7: NOx vs Three Spark Plugs

After three successive tests with a given spark plug at a given mode, the only
change made was to replace one spark plug with another, restart the engine, and run it
until it had fully warmed up. Comparisons of results from successive tests with the same
spark plug indicated the repeatability of the emissions measured was about 5% for this
two-stroke engine. Since the reduction in emission achieved with the Pyrotek plug was
considerably more than this uncertainty, it was considered a signiﬁcant effect attributed

to geometric differences in the spark plugs.

4.3 Four stroke Tecumseh side valve (VLV60) utility engine

The side-valve engine used in these studies was a brand new engine which had
been broken in for six hours in accordance with the manufacturer’s recommendations.
The engine was run at WOT, the maximum torque at its rated speed of 2800 rpm was

recorded, and the engine emissions were then measured according to CFR90 procedures.

38

In studies of the Tecumseh side valve engine operating with different spark plugs,
emissions were measured at each of 6 loads required for a CFR40 certiﬁcation test, as
well as comparing emissions as functions of angular position of the spark-plug gound

electrode and as functions of different spark-plug gaps.

4.3.1 Effect Of Ground Electrode Geometry

In this test, the effect of gound electrode geometry on emissions was evaluated.
The procedure adopted was to record the maximum torque produced at WOT and then to
measure emissions according to CFR90 requirements. These requirements specify that
the engine be operated at full (100%) load and subsequently at 75%, 50%, 25%, and 10%
of its full load, as well as at idle. These tests were repeated three times for each spark
plug. The procedure of adjusting the throttle to achieve target loads, for each of three
different spark plugs, resulted in the engine operating at slightly different air-to—fuel
ratios for each plug, at nominally the same load. This added variability in air-fuel ratio
was believed to obscure the effect of gound electrode geometry on emissions

measurements.

Figure 8 shows the hydrocarbon emission levels for the Pyrotek and baseline
spark plugs for this side valve engine. It can be seen that there was a signiﬁcant reduction
in hydrocarbon emissions in the case of Pyrotek spark plug, at 50% and 75% load.
However, at these loads, the engine ﬁtted with the Pyrotek plug operated at an air-to-fuel
ratio slightly higher than it did when ﬁtted with the baseline spark plug as can be seen

from Figure 9. A similar trend is observed for the NOx emissions shown in Figure 10.

39

Hydmwbon
emissions(grams/hour)

AFR

    

Comparative effect of spark plug ground electrode geometry
on hydrocarbon emissions
(CFR90) TecumsthLV60 side valve engine

.NEW DPYROTEK

 

 

 

 

Figure 8: HC Ennissions vs % Load

A ir-fuel ratio at different loads

14.00

12.00

 

.New
.Pyronek

    

 

0% 10% 25% 5096 75% 10096
%Lo|d

Figure 9: APR vs % Load

Comparative effect of ground electrode geometry on NOx
emissions (CFR90) Tecumseh VLV60 sidevalve engine

251

 

.NOx (grms/hour) NEW

 

 

 

 

 

    

 

 

 

 

 

 

 

5
5 20 .NOx(grms/hour)PYROTEK
E 15
9i:
.2 10
.3
E
g 5
Z

0-

0% 10% 25% 50% 75% 100%
%Load

Figure 10: N03: vs % Load

4.3.2 Effect of Ground Electrode Orientation

In this series of tests, emissions of the side valve engine were measured when the
gound electrode was oriented in each of several different angular positions. The gound
electrode oriented with its open end facing the exhaust valve was taken as the datum, and
regarded as 0° orientation, as shown in Figures 11 and 12. Successive emissions tests
were then carried out with the Pyrotek spark plug rotated counter clockwise at 45°
increments from 0° to 180°, and with the baseline Champion spark plug rotated at 90°
increments. Spark plug indexing washers of thickness 0.030-inch, 0.004-inch, 0.005-inch
and 0.060-inch were used to permit the required orientation to the spark plugs witlnout
damaging the cylinder threads. The change in spark plug position had negligible effect on

compression ratio as illustrated by calculations in Appendix F. These effects of angular

41

position on emissions were monitored at two load settings: ﬁrst, at the idle condition with
the engine disconnected from the dynamometer; and second at 38% of full load. The
emissions measured with the Pyrotek spark plug revealed there was a signiﬁcant

inﬂuence of gound electrode orientation on the combustion characteristics.

 

Figure 12: Layout of Combustion Chamber

42

In comparison with the baseline spark plug, changes in emissions (particularly
hydrocarbon emissions) through rotation of the Pyrotek spark plug were large. Five
angular positions at part load and two at no-load settings were considered. In the case of
the baseline spark plug, only three angular positions were studied at part load. These
changes in the hydrocarbon errnissions for different orientation angles are shown in Figure
13.

Effect of ground electrode orientation on hydrocarbon
emissions atpart load(Tecumseh VLV60)

10.5 L ,‘. -..- ‘_ n- L
10.25
g 10
G
.8 i 9.75
E
g 9.5
"3.5 9.25
:1: .3
E 9
8.75
8.5

 

0 45 90 135 180
Orientation Angle(degrees)

Figure 13: Orientation Angle vs HC Emissions at Part Load

The results of Fig.13 show that the Pyrotek spark plug was more sensitive to the
orientation compared to the baseline spark plug, with improved emission performance at
orientation angles of 0°, 45°, 90°, and 135°. Of these positions, the 0° orientation gave
the lowest hydrocarbon emissions. The performance of the Pyrotek spark plug when
facing the inlet valve (180°) was inferior to the 0° orientation. Changes in hydrocarbon
emissions for base line spark plug are quite low when compared to the estimated test-to-
test uncertainty level of 3%. Thus, it is possible that there may not be any real change in

hydrocarbon emissions for different orientations of the baseline spark plug.

43

At idle, with the Pyrotek gound electrode facing the center of the cylinder
(90°orientation), the unburned hydrocarbons were signiﬁcantly (50%) lower than when

facing the nearest wall (270°orientation).

Effect of ground electrode orientation at idle

Hydrocarbon emissions(glmI/hour

 

"C(Pyrotek) HC(Bnaeline)
0 rientation angle

Figure 14: Orientation Angle vs Hydrocarbon Emissions at Idle

No signiﬁcant effect on hydrocarbon emissions was observed when the baseline
spark-plug electrode was rotated, as is evident from Figures 13 and 14. The results were
almost identical for all three orientations at part load. In case of the idle tests, the
hydrocarbon emissions with the gound electrode facing the center and the nearest part of

the cylinder wall were scarcely different.
4.3.3 Inﬂuence Of Spark-Plug Gap

In this series of experiments, the gaps of the Pyrotek and baseline spark plugs
were varied from .0025 inch to 0.0045 inch, and the corresponding effects on emissions

measured. These tests were performed at a ﬁxed throttle setting corresponding to no load,

with using spark plugs with the different gaps in successive tests. The changes observed

in the hydrocarbon emissions when changing spark gaps are shown in Figure 15. It was
observed that HC emissions measured with the Pyrotek spark plug were more sensitive
to the spark gap used than when the baseline spark plug was ﬁtted. There was a 13%
increase in the HC emission level when the spark gap was reduced from the
recommended 0.0030 inch to 0.0025 inch. In addition, the hydrocarbon emission level
increased whenever the gap was changed from 0.003 inch to a larger gap. In the case of
the baseline spark plug, the hydrocarbon emissions did not vary much as the gap was
changed from 0.0025 inch to 0.0035 inch. The lowest hydrocarbon emissions were found
at a gap of 0.003 for the Pyrotek plug and at a 0.0035-inch gap for the baseline spark

plug.

Influence of spark gap on hydrocarbon emissions

.Pyrotek ‘

 

Hydrocarbon emiadamGnm/hom
0
in

 

 

0.0025 0.003 0.0035 0.004 0.0045
Spark gap(inch)

Figure 15: Spark Gap vs HC Emissions

4.4 Four stroke Tecumseh overhead valve (OHHSO) utility engine

The overhead-valve engine used in these studies was a brand new engine which

had been broken in for four hours according to the manufacturer’s recommendations. The

45

engine was run at WOT, the maximum torque at its rated speed of 2800 rpm was
recorded, and the engine emissions were then measured according to CFR90 procedures.
Emissions tests were carried out at 100%, 75%, 50%, 25%, and 10% of the maximum
load and at idle. In these experiments, at each throttle setting, the three spark plugs were
tested in succession. In this way, all comparative tests at the same load were performed at
identical throttle settings and at about the same air ﬁrel ratios, making it easier to identify
the effect of using a particular spark plug. In addition to the 6-mode CFR90 emissions
tests on conventional and crown-shaped (Pyrotek) plugs, a series of experiments was
performed to examine some effects on engine emissions of center electrode geometry,
and engine speed at part load. The details of the tests carried out are discussed separately

in the following sections.

4.4.1 Effect Of Ground Electrode Geometry

The hydrocarbon emissions for the baseline and Pyrotek spark plugs are shown in
Figure 16 and the corresponding N0, emissions are displayed in Figure 17. The data
show a clear trend with the Pyrotek spark plug tending to produce lower hydrocarbon
emissions with a slight increase in N0Jr emissions. There was, on average, a 7% reduction
in hydrocarbon emissions throughout the test range. This reduction in hydrocarbon was
most pronounced at 100% load and least sigiiﬁcant at 75% load. There was no
signiﬁcant change in hydrocarbon emissions on account of using different spark plugs at
idle. In the case of N0, emissions, this trend is almost reversed. Increases in the N0,
emissions were signiﬁcantly lower than decreases in hydrocarbon emissions when

substituting a Pyrotek for a baseline spark plug.

46

Comparative effect of spark plug ground electrode geometry
on HC emissions (CFR 90)Tecumseh-OHH50

.New

.Pyrouk

 

 

 

 

HC emissions(grama/hour)

 

Figure 16:HC vs % Load

Comparative effect of spark plug ground electrode geometry
on NOx emissions (CFR 90)Tecumseh-OHH50

16
14
12
10

NOx emissions(gams/hour
an

 

ON.O

0% 10% 25% 50% 75% 100%
% Load

Figure 17: N0, Emissions vs % Load

4.4.2 Effect Of Center Electrode Geometry

The spark discharge from the Pyrotek spark plug was studied in its original cylindrical

shape and when modiﬁed by inserting a V-notch in the center electrode. It was observed

47

that, in the case of the unmodiﬁed Pyrotek plug, the spark travels from center electrode
edges to the ground electrode along the shortest path. Multiple exposure photographs of
the spark discharge in quiescent ambient air were taken and the paths followed by sparks

for botln spark plugs are shown in Figure 18.

 

Figure 18: Spark Patterns For Two Spark Plugs

Figure 18 indicates that, in the case of the Pyrotek plug, sparks were only observed at the
outer edges of the center electrode. This observation suggested that the middle of the
center electrode’s upper surface does not participate in the discharge process, and so it
might serve only to conduct energy away from the initial ﬂame kernel. Hence, it was
decided to make a V-notch in the pyrotek spark plug in order to reduce the contact area

fraction of the electrodes and possibly reduce heat losses.

A V-notch was made in the center electrode and another series of tests comparing
the engine emissions using these three types of spark plugs was carried out. Figure 19
shows the spark pattern observed with the Pyrotek plug with its center electrode
geometry modiﬁed in this way. The spark discharge patterns in Figures 18 and 19 appear
to be qualitatively the same so this modiﬁcation may have remedied some disadvantage

of an unnecessarily large contact area ratio. Figures 20 and 21 shows the emission levels

48

for the three types of spark plugs installed in this overhead-valve engine. When using the
V-notched center electrode, a 6.5% reduction in the hydrocarbon emissions was observed
relative to the unmodiﬁed Pyrotek spark plug and an 11% reduction was achieved
compared to the baseline spark plug. The V-notch geometry was found to be particularly
effective in reducing hydrocarbon emissions at low loads. This observation is consistent
with the ﬁndings of Hood [4]. Since this engine was operating at an equivalence ratio of
roughly 1.3, where Heywood [24] has noted that these heat transfer effects are not as
signiﬁcant as at leaner mixtures, the potential beneﬁts of the V-notch geometry might be
even geater at equivalence ratios closer to l. The reduction of electrode contact area by
making a V-notch indicates that the heat loss from the ﬂame kernel to the electrode plays

an important role in the ﬂame propagation characteristics of spark igiited ﬂames.

 

Figure 19: Pyrotek Spark Plug Center Electrode Geometry (V-notch) and Spark Pattern

49

Effect of ground and center electrode georn etry on hydrocarbon
emissions

ll W'_—‘
16 .Pyionek .
14 gpyrorekuqroove)

     

 

 

 

Hydrocarbon
emissions(grama/hour)

 

Figure 20: HC Emissions vs % Load

Effect of ground and center electrode geometry
on NOx emissions

.Pyrouk

 

Nox emissions(Grams/hour)

0% 10% 25% 50% 75% 100%
% Load

Figure 21: N0Jr Emissions vs % Load

4.4.3 Effect of Engine Speed On Engine Emissions

A series of experiments were carried out at ﬁxed throttle settings and different
engine speeds to explore how gound-electrode geometry might affect engine emissions
at different engine speeds. In these tests, all other controllable parameters were kept

constant while the engine speed was changed from 2400 rpm to 3000 rpm. The changes

50

in speciﬁc and brake-speciﬁc hydrocarbon emissions as functions of engine rpm are
shown in Figure 22 when using baseline and unmodiﬁed Pyrotek spark plugs. When
ﬁtted with the Pyrotek spark plug, hydrocarbon emissions decreased almost linearly with
increasing engine speed. In contrast, emissions of this engine equipped with the baseline
spark plug rise to a maximum at 2600 rpm and then decrease at higher engine speeds.
The differences in these responses can be attributed to the differences in spark-plug
geometry. A possible interpretation is that each electrode may inﬂuence bulk gas motions

in the vicinity of the spark-plug gap differently, which is consistent with the ﬁndings of

 

 

 

 

 

 

 

 

 

 

 

Halldin [38].
Effect of engine rpm on hydrocarbon emission
6 _
A
m 5 / -
g X
a 4
e t:
o w
G 3
8 W
3
8 2
'3 +pyrotek(grm/hr)
2.; + pyrotek(grrn/hp/hr)
1 -.—baseline(grm/hour) ,
+baseline(gms/hp/hr)
0

 

 

2400 2600 2800 3000
rpm

Figure 22: Hydrocarbon Emissions vs Engine RPM

51

CHAPTER 5

DISCUSSION OF RESULTS

5.1 Effect of Ground Electrode Geometry

The use of the Pyrotek spark plug has been shown to consistently reduce
hydrocarbon emissions at the expense of increasing NOx emissions, for a two stroke
chainsaw engine (nominally operating at 05 51. 5) and for two four stroke utility engines (11>
5 1.3, side valve & t1) 5 1.35 overhead valve). This reduction in hydrocarbon emissions,
and increase in N0, and torque is consistent with a faster binning rate, which would
reduce the burn duration and would cause lower hydrocarbon emissions. A possible
explanation for this reduction in hydrocarbon emissions and increase in N0, emissions is
that the Pyrotek gound electrode geometry favors a larger irnitial ﬂame kernel. As was
shown in Figure 18, in section 4.4.2, the electrical discharge in the pyrotek plug is over a
longer path than in the baseline spark plug, which therefore leads to a larger initial ﬂame
kernel. It is known that larger irnitial ﬂame kernels increase the burning rates, torque, and
NOx emissions and lower the hydrocarbon emissions [20]. Maly [20] experimentally
compared the inﬂuence of the breakdown, are, and glow phases of the discharge on
engine performances and noted these effects. In these experiments, he noticed that ﬂame
kernels with the largest radii, in the breakdown phase, caused the highest N0Jr levels

whereas ﬂame kernels of the smallest radii (achieved with Capacitor Discharge Ignition

52

(CDI) resulted in the lowest N0Jr concentrations, with corresponding reductions in

hydrocarbon emissions at near stoichiometric mixture compositions.

Larger initial ﬂame kernels increase the burning velocity and reduce the burn
duration, when other parameters, including the air fuel ratios, are unchanged. They are

discussed in detail below. The burn rate is given by

dm
EL = puAfSI

(45)

and for a ﬁxed laminar ﬂame speed, geater ﬂame area correspond to faster burn rates. If
the Mallard-Le-Chatalier model is applied to the ﬂame ﬁ'ont spread in an IC engine, and

the ﬂame kernel is initially at its adiabatic ﬂame temperature:

S, = Jar" RR -
(46)
substituting (46) into (45), we get
dm
7% = puAfJa*M = pu(4zzrf2 )Ja'l'RR.
(47)

Then a larger initial ﬂame kernel radius equates to a faster burn rate. Higher burn rates
reduce the burn duration, reduce the total heat losses through time dependent heat transfer
and so result in maintaining higher burned gas temperatures, and according to the Mallard

Le-Chatalier ﬂame theory, higher reaction rates/ ﬂame speeds.

53

Another possible explaination is that convection velocities in the vicinity of the
spark gap, which would convect the ﬂame kernel away from the electrodes, are ernhanced
by the crown shape of the gound electrode. They would then increase the ﬂame kernel
gowth rate by reducing heat losses to the electrode [24]. Heywood [24] using Schlieren
photog'aphy, observed that the ﬂame kernels which move away from the electrodes gow
faster. If the gound electrode of the Pyrotek spark plug (Figure 23) also changed the
tnnbulence intensity in the vicinity of the electrodes, it might also promote a faster initial
ﬂame kernel development than with the baseline spark plug. According to the Brown,

Stone and Beckwith turbulent combustion model [41]:

dm ,
'27 = P0421" + S!)
(48)

Higher values of turbulence velocity factor (u ’) would promote turbulent combustion at

increased burn rates.

Another important design aspect of the gound electrode geometry is the shape of
the joint between the gound electrode and the plug body. As is evident from Figme 24,
the Pyrotek plug has a larger surface area compared with the baseline spark plug. It is
possible that this increased area may also inﬂuence the bulk charge motion in its
immediate vicinity in a manner favoring faster ﬂame propagation. If the beneﬁts of this
effect outweigh the increased heat losses which might be incurred, this kind of joint
might be advantageous. However, this aspect of spark plug shape was not studied in

detail.

54

 

Figure 23: Ground Electrode Geometry

 

Figure 24: Ground Electrode Base

5.2 Effect of Ground Electrode Orientation

Since neither of the spark plugs are axisymmetric, their position with respect to
the momentary ﬂow direction inﬂuences the local ﬂow ﬁeld around the electrodes and
thereby affects the ﬂuid ﬂow in the early stages of ﬂame kernel gowth. As the gound
electrode orientation inﬂuences the velocity ﬁeld within the spark-plug gap, it might
facilitate early ﬂame kernel convection either towards the gound electrode or away from
the gound electrode, or neither, depending on the gound electrode orientation. The in-
cylinder ﬂow, the residual and inﬂow conditions, as well as intake system dynamics and

cyclic variations in the ﬂow ﬁeld are strongly dependent on the inlet port, inlet valve, and

55

cylinder head geometry. These ﬂows appear to become unstable, either during the intake
or compression process, and breakdown into three-dimensional turbulent motion. The jet-
like character of the intake ﬂow, interacting with the cylinder walls and piston motion
creates large-scale rotating ﬂow patterns within the cylinder [38] in some engines.

Therefore, it is difﬁcult to predict the path a ﬂow might follow in the vicinity of a spark.

It is thought that the emissions produced by engines with spark plugs with larger
and longer gound electrodes are more sensitive to the orientation of the spark plug [2].
This may be because larger gound electrodes generate broader wakes behind the
electrode if the gound electrode faces the momentary ﬂow direction. For spark plugs
with smaller gound electrodes and shorter projections, engine emissions are not very
sensitive to the ﬂow direction [2]. The emissions measured with the Pyrotek spark plug
were more sensitive to the gound electrode orientation than with the baseline Champion
spark plug. A possible explanation for this observation is that the larger electrode area at
the gound electrode base of the Pyrotek spark plug, shown at Figure 24, has a beneﬁcial
inﬂuence on the ﬂuid motion in the vicinity of the spark gap which outweighs any

disadvantages of increased heat loss.

The angular positions of the spark plug were denoted as 0° facing the exhaust
valve, 90° facing the center of cylinder and 180° facing the intake valve, as shown in
Figure 22. As noted by the Anderson and Asik [25], for a spark plug of baseline shape in
a 0.4L Bowl- in-piston single cylinder engine, the 180° and 90° orientation have
improved combustion performance over the 0° orientation. These results are consistent

with our results using a baseline spark plug in a side valve engine. Anderson and Asik’s

56

explanation for the reduced performance as given in [25] is that: in the 0° orientation, the
spark plug may shroud the gap from the charge motion. The electrode may create a
recirculation region near the gap, which allows the charge a longer time for the formation
of a stable ﬂame kernel. In the 180° position, in contrast, it may allow the ﬂow to move
the ﬂame kernel to regions where they experience increased contact with the electrode,

geater heat losses reduced ﬂame speed

2 if?

1"

14 -

 

(b)

57

180° orientation

 

((0

Figure 25: Cartoons of the Inﬂuence of Plug Orientation on Flame Front Shape

Another possible explanation for the improved performance of the Pyrotek spark
plug may be the different gound electrode geometry. The gound electrode, when facing
the exhaust valve (0° orientation), may cause the deﬂection of the any swirl ﬂow in a
manner inwhich the ﬂame kernel is moved away from the electrode region toward the

center of the combustion chamber, Figure 25(a). This may enhance its initial gowth rate

58

through reduced heat transfer to the electrode. The same effect may cause faster burning
rates at 45°, 90° and 135° orientations. In the case of the gound electrode facing the inlet
valve 180°, the larger shrouding effect cited by Anderson and Asik[25] may cause a
longer residence time for ﬂame kernel formation. This may cause excessive heat losses to
the electrodes thus reducing the burning rates. The results obtained using the Pyrotek
spark plug contradict the ﬁndings of Asik and Anderson [25] which were better for 180°
orientation. However, they are in ageement are in ageement with the ﬁndings of
Zeigler, Schaudt, and Herweg [26] for ignition in a constant ﬂow chamber, which were

better for the 0° orientation.

In the case of the idle runs, with the gound electrode facing the center of the
cylinder (90°orientation), the unbnn'ned hydrocarbons were signiﬁcantly lower than when
the gound electrode was facing the nearest part of the cylinder wall (270°orientation).
This might be explained if the local mean ﬂow in the vicinity of the spark plug were
toward the wall, thus blowing the ﬂame kernel, on the average, towards the wall as
shown in Figure (25d). Heat losses from the ensuing ﬂame front to the wall may result in
delayed combustion through this quenching effect. In additiorn, shrouding of the spark
gap may also play a vital role in this conﬁguration, causing more heat losses to the
electrode surface. The hydrocarbon emissions of the baseline plug when facing the center
and nearest part of the cylinder wall are scarcely different. This may be due to a reduction
of Anderson and Asik’s shrouding effect of the gound electrode of the baseline spark

plug on ﬂame kernel gowth and charge motion.

In comparison with the ﬁndings of the other studies, the results obtained are in

ageement with those of Zeigler, Schaudt, and Herweg [26] and Mantel [21] for the

59

Pyrotek spark plug, and with Anderson and Asik [25], in the case of the baseline spark
plug. It appears, for a given bulk ﬂuid motion, there exists an optimal spark plug
orientation. However, the bulk ﬂuid motions are often different at different engine
speeds, loads, and tlnrottle settings so a single optimization for a given engine may not

always be possible.

5.3 Effect of Center Electrode Geometry

The center electrode of the Pyrotek spark plug was modiﬁed as shown in the
Figure 26. In this case, a reduction in hydrocarbon emissions was achieved relative to the
unmodiﬁed Pyrotek and base line spark plugs. The disadvantage of an unnecessarily
large electrode area, compared with the baseline spark plug, was partially overcome by
making a V-notch in the center electrode. Based on experimental observations that spark
discharge occnu'red mainly at the periphery of the center electrode, the ﬂame kernel
contact area can be reduced by reducing the effective surface area of the center electrode.

In the ﬂame gowth phase, convective heat loss plays an important role, as modeled as

glass 2 “C(71) — If)’

(49)

where A, is the contact area with the electrodes. The contact area ﬁaction can be used to
estimate the effect of heat loss to the electrode during ﬂame gowth. As noted in [24], this
contact area fraction is lower for the tlnin electrodes when using spark plugs of baseline

design.

The HC emissions results of Chapter 4 show that plugs with smaller contact area
yield greater beneﬁts at idle and low loads. This observation is also consistent with the
ﬁndings of Hood [4]. Since this engine was operating at an equivalence ratio of roughly
1.3, where Heywood [24] has noted that these heat transfer effects are not as signiﬁcant
as at leaner mixtures, the potential beneﬁts of the V-notch geometry might be even

geater at equivalence ratios closer to 1.

 

Figure 26: Center Electrode Geometry

5.4 Effect Of Spark Gap

The inﬂuence of the spark gap manifests itself through two effects. The ﬁrst effect
is to change the breakdown voltage and energy discharged: the larger is the gap, the
higher is the voltage required to break down the gap, and the larger the energy released in
this phase of discharge. The second effect is to change the proportion of energy lost to the

electrodes through heat transfer.

61

It was observed that emissions measured using Pyrotek spark plugs were more
sensitive to the spark gap than those using a baseline plug. A plausible explanation is: for
small gaps, the electrodes remove excessive quantities of heat from the incipient ﬂame,
thus causing a reduction in the ﬂame propagation speed and therefore increasing the burn
duration. This effect seems to be more pronounced when there is a large contact area, as

with the Pyrotek gound electrode.

As the spark gap is increased above the recommended value, the ﬂame has more
room to gow without the inﬂuence of the electrodes. However, when the spark gap
becomes too wide, a higher voltage is required across the gap, the energy release rate
become faster and the spark duration become smaller. The spark may lose excessive
energy to the unburned mixture through heat transfer. Increasing the spark gap also
requires higher breakdown voltages with associated reductions in the spark duration. This

reduced spark duration slows ﬂame kernel gowth [12] resulting in higher HC emissions.

5.5 Effect of engine rpm on comparative emissions of two spark plugs

When ﬁtted with the Pyrotek spark plug, hydrocarbon emissions decreased almost
linearly with increasing engine speed. In contrast, emissions of tlnis engine equipped with
the baseline spark plug rise to a maximum at 2600 rpm and then decrease at higher
engine speeds. The differences in these responses can be attributed to the differences in
spark-plug geometry. A possible interpretation is that each electrode may inﬂuence bulk
gas motions in the vicinity of the spark-plug gap differently, which is consistent with the
ﬁndings of Halldin [38]. It is also related to the observation of Witze, Hall and Bennett

[34], that the swirl velocity increases linearly with engine speed, which might cause a

62

proportional increase in the ﬂame kernel gowth rate. Shen and Jiang [35] also studied
this effect by comparing the early ﬂame kernel development at different engine speeds,

and noticed that turbulence intensity varies almost linearly with the mean piston velocity.

63

Chapter 6

Conclusions

. The Pyrotek spark-plug gound-electrode geometry changed the emission

characteristics of several small engines compared to baseline Champion spark plugs.

. Use of the Pyrotek spark plug caused reductions in HC emissions from these engines
which were attributed primarily to faster ﬂame kernel gowtln rates. The faster ﬂame
kernel gowth rates were attributed to the larger irnitial ﬂame kernels produced by the

Pyrotek spark plug over a larger effective gap.

. A secondary effect of the Pyrotek geometry was thought to favor the faster initial
ﬂame-kemel gowth by enhancing convection of the initial ﬂame kernel away ﬁ'om

the electrodes, thereby reducing the heat losses.

. At low loads, the beneﬁts of the Pyrotek electrode may be outweighed by higher heat

losses to the electrodes. This can be partially overcome by reducing the center

electrode contact area.

. The Pyrotek spark plug is more sensitive to its gound electrode orientation when

compared to the baseline spark plug. This is attributed to the effect of the larger

surface area in promoting heat transfer and the shape of Pyrotek spark plug electrode

on the local ﬂow ﬁeld.

. There exists an optimum spark gap and gound electrode orientation for each

combustion chamber geometry and engine operating conditions.

65

 

Appendix

 

 

Appendix A

Factors Effecting Burn Duration

A.l Engine speed

Mixture burning rate is strongly inﬂuenced by the engine speed. Burning rate
throughout the combustion process increases almost as rapidly as engine speed,
increasing in-cylinder velocities. This increase in burn duration due to the increased in-
cylinder velocities causes an increase in the turbulent intensity, resulting in more

turbulent combustion.

A.2 Equivalence ratio

The inﬂuence of the equivalence ratio is small for near stoichoimetric mixtures in
the range as 0.8< 0 <1.2. The burn duration increases with larger deviations from
stoichiometry due to the decrease in the laminar speed or equivalently the reaction rate.
The burn duration varies signiﬁcantly less over the same range of air fuel ratios if the
spark timings are varied, such that combustion takes place closer to the TDC. This
indicates that the increase is mainly caused by the initial combustion process where the
laminar speed is important. The increase in the burn duration is determined from the

inﬂuence of chemical kinetics and is more pronounced at low levels of turbulence.

66

A.3 Flow field effects

Flow ﬁeld effects can be divided into small-scale effects and the large-scale
effects. The small-scale effects are local in nature, such as the local turbulence levels, air
fuel ratio ﬂuctuations, and heat transfer rates. These small-scale effects are highly
inﬂuenced by the large-scale effects, such as mean ﬂow structures, geometry, and overall
fuel ratio. The turbulence, for instance, enhances the local ﬂame propagation, but the
mean ﬂow structures effect the total front area and the ﬂame position witlnin the
combustion chamber. The total combustion duration is determined by the combined
effect of propagation speed and necessary ﬂame travel distances to burn up the charge.
The best performance is obtained when both the small-scale structure and the large-scale
structure are more favorable for the combustion process. The concept of using both large-
scale and small-scale ﬂow to promote the combustion has already been implemented.
Fast burn engines have shown improvement in the fuel consumption and a control of

emissions [28].

A.4 Effect of swirl and tumble

Swirl and tumble also inﬂuence burn duration. Experiments have revealed that
engine with high swirl and tumble had a faster burn rate as compared to those with
quiescent ﬂow in every stage of the combustion [18]. The geatest improvement on
account of tumble occurs at the early part of the cycle whereas in the case of swirl the
burn rate improvement is in the middle of the burning process. As reported by [18], when

tumble is present, the organized motion is broken up into small-scale turbulence during

67

the compression process so that the turbulence intensity is highest at the end of
compression. In the case of swirl, although the swirl velocity decays somewhat during the
compression and expansion, the magnitude of the mean velocity and turbulence remains

reasonably high, and tlne burn rate is sustained until later in the cycle.

68

Appendix B

Factors Effecting the Flame Kernel Growth

8.] Ignition energy and spark duration

The three phases of ignition process [20] are: the breakdown phase, the arc phase,
and the glow phase. The break down phase was found to have high thermal efﬁciency. In
addition, ignition systems with longer spark durations (1.5 - 2 ms) are also very efﬁcient,
because they permit the maintenance of a hot central core. From this hot kernel core, heat

may diffuse to the ﬂame surface and mask cycle to cycle variations near the spark gap.

The lengtlnening of the spark discharge duration is particularly effective in
achieving stabilized combustion [12]. Longer spark dnn'ation provides a continued supply
of electrical energy as kinetic energy to the mixture around the spark gap. Analytical
results and constant volume combustion chamber tests verify that longer spark duration
promotes ﬂame initiation and makes more reliable ﬂame propagation possible [12]. The
length of the spark duration is regarded as the period from ignition to the onset of

combustion pressure rise.

69

8.2 Effect of electrical energy on flame growth

Based on experimental observations [6], the cycles in which the ﬂame kernel is
convected away from the electrodes are more likely to be the cycles with more electrical
energy transferred to the kernel. The discharge channel follows the ﬂow ﬁeld and is
stretched in length leading to a larger voltage drop in the positive column, which implies
that the higher electrical energy leads to a fast initial ﬂame gowth. The cycles with larger
ﬂame radii are the cycles, which result in earlier 10% burned location, indicating faster

ﬂame kernel gowth rate.

8.3 Heat losses to spark plug electrodes

Generally, the presence of the electrodes can be described as a heat or radical
sink, which can eventually lead to the “ﬂame quenching” in the case of unsuccessful
ignition. The ﬂame kernel and electrode geometry play a signiﬁcant role in the heat
transfer ﬁom the ﬂame kernel to the electrodes. The character and relevance of the
geometrical interactions between ﬂame and electrode are of particular interest in an
engine, where the ﬂame development process is known to vary from cycle to cycle. The
inﬂuence of this heat transfer from the ﬂame kernel to the electrodes geatly depends on
the contact area ratio. The electrodes with smaller area are found to have smaller heat

losses thus resulting in better combustion [24].

B.4 Flame kernel advection

It has been observed experimentally that smaller ﬂame kernel are essentially

centered in the spark gap whereas larger ﬂame kernels are convected away from the spark

70

gap center. As observed by the Heywood [24], there exists a correlation between the
ﬂame radius and the ﬂame kernel displacement. Flame kernels, which are convected
away from the gap center, gows faster. The cycles in which ﬂame kernels do not move
away ﬁ'om the electrodes, have slower gowth. The basic cause of this phenomenon is

heat losses from the ﬂame kernel to the electrodes.
B.S Characteristics of charge motion

The ﬂame initiation is accelerated when the ﬂow ﬁeld convects the initial ﬂame
kernel. Cycle to cycle analyses [24] of ﬂame irnitiation have shown that a ﬂow velocity at
the ignition point is favorable to succesful initiation. When the ﬂame kernel is convected
away from the electrode gap, the heat losses to the electrode wall are limited. The
secondary effect of the ﬂow is to stretch of initial ﬂame. The characteristics of the charge
motion have a signiﬁcant effect on the expansion velocity. The expansion velocity is
enhanced by an increase in the entrainment area due to the wrinkling of the ﬂame front
by the turbulence. In addition, the ﬂame front may be stretched considerably if the ﬂame

kernel is ‘anchored’ at the spark gap when there is substantial bulk motion.
B.6 Effect of convection velocity

The overall convection of the ﬂame kernel in the early stages of ﬂame
development is due to the bulk charge motion at the spark plug. This bulk ﬂow could
move the ﬂame kernel away or toward the electrodes and change the heat losses from the
ﬂame. The high bulk ﬂow would also lengthen the discharge channel and would increase
the electrical energy to the kernel. In addition, the direction of the ﬂame kernel toward or
away ﬁ'om the wall also inﬂuences the ﬂame geometry relative to the combustion

chamber at the later stages of ﬂame development. There is a strong correlation which

71

exists between the expansion speed and the ﬂame kernel convection velocity. This

correlation can be explained by three factors

0 Turbulent velocity ﬂuctuations are higher, if the bulk charge motion velocity is

higher, hence there is a faster ﬂame due to the turbulent enhancement of the ﬂame

speed.

0 For ‘anchored’ ﬂames, the higher convection velocity gives a larger ﬂame stretch.

0 Higher convection velocities tends to advect the ﬂame kernel out of the spark plug

region, hence reducing the heat losses to the spark plug electrodes. [24].

72

Appendix C

Engine Data:

Table 3: Two Stroke Chainsaw Engine

3.12
32.4

1W0

 

Table 4: Four Stroke Tecumseh Side Valve (VLV60) Engine

 

 

 

 

 

Bore (cm) 7.1018
Stroke(cm) 5. 1993
Displacement(cm3) 207
Rated kW 4.4742
Conﬁguration vector light weight vertical
(side valve)

 

 

 

 

73

 

Table 5: Four Stroke Tecumseh Overhead Valve (OHI-ISO) Engine

 

 

 

 

 

 

Bore (cm) 6.6675
Stroke(cm) 4.92252
Displacement(cm’) 171.89963
Rated kW 3.7285
Conﬁguration Overhead valve Horizontal

 

 

 

74

Appendix D

Overview of Emission Bench and Dynamometer

D.1 Dynamometer

The Hydraulic dynamometer dissipates the energy produced by the engine
through an oil to air heat exchanger. It uses a hydraulic pump driven by an electric motor.
The ﬂow rate is controlled by the computer via a stepper motor, which governs the
engine RPM. The RPM are set by the ﬂow rate with the use of Stepper motor and the
engine can be loaded to desired setting while rpm is kept constant. A torque sensor is

mounted between the engine and the hydraulic pump to monitor the torque loading of the

engine.

D.2 Emission Measurements

D.2.1 C0 and C02 measurements

The C0 and C02 meters employ a non-destructive measurement technique. This
involves passing a speciﬁc wavelength of infrared light (IR) through the sampled gas. An
IR detector on the opposite side detects how much IR light has passed through the

sample. Since C0 and C02 absorb the IR light (over different frequency ranges), the

75

geater the concentration of C0 or C02 in the sample, the less IR light that will pass

tlnrough.

D.2.2 HC measurements

The HC meter uses a different approach and is tuned to speciﬁc carbon compositions
witlnin hydrocarbon mixtures. Composition of the two stroke exhaust will mirror the
composition of the two stroke fuel whereas, the composition of the four stroke exhaust is
mostly C2 and C2 molecules [33]. Therefore, a span gas is used which it closely
resembles the exhaust gas in question. The analyzer detects the amount of the HC ’5
present in the sample using the ﬂame ionization technique. The ﬂow of a sample with
carbon compounds through the hot ﬂame leads to an ionization current. This current is

proportional to the volume fraction of the HC ’s present within the sample.

D.2.3 N0, measurements

The NO, meter uses a chemical illumenescent process. Nitrogen compounds in
the exhaust gas are actually a mixture of N0 and N02, which are usually written N0... In
the analyzer, the N02 is ﬁrst converted to NO. The resulting sample then react with 03,
which is generated by an electric discharge through the oxygern, at a low pressure in a
heated vacuum chamber. The light emitted by the reaction is measured by a photo-

multiplier and indicates the N02 concentration within the converted sample.

76

E.1

Appendix E

Experimental results in tabulated data form

Ground electrode geometry

E.1.1 Two Stroke Chain Saw Engine

TABLE 6: Effect of Three Different Spark Plug on Two Stroke Engine Characteristics

 

 

 

 

 

 

 

 

 

 

Spark Plug Pyrotek New(base) Used(base) Pyrotek New(base) Used(base)
Parameters 6% 6% 6% 100% 100% 100%
Power(Brake Hp) 0.05 0.05 0.04 0.84 0.81 0.74
"Thrque(Ft_Lb) 0.06 0.06 0.05 0.46 0.44 0.41
Exht Temp(C°) 214.10 224.80 212.60 393.03 398.93 385.33
CO (gms/hour) 8.00 56.50 60.60 680.30 716.83 687.67
C02 (grins/hour) 229.98 224.46 212.27 870.52 806.01 745.74
N0. (grms/hour) 0.00 0.00 0.00 0.63 0.47 0.47
HC (gms/hour) 31.03 32.70 34.12 203.08 229.99 237.72
AFR 14.26 8.87 8.82 11.77 12.86 12.92
RPM 4340.00 4433.00 4573.00 9571.00 9537.33 9521.67—

 

 

 

 

 

 

 

 

 

 

77

E.l.2 Four Stroke Tecumseh Sidevalve (V LV60) Engine

TABLE 7: Effect of Ground Electrode Geometry on Four Stroke Engine Emissions

0.00 0.36 0.72 1.37 2.19 2.91

0.00 0.81 1.36 2.58 4.1 5.47
T Exht(C°) ' 153.97 274.17 317. 10.23 376.50 370.47

233.67 1.37 17 5.80 375.87 .67
Baselnne 150.03 5.40 68.30 .3 586.43 1 5.87

.70 16.06 864. 964.67 1559.02 1438.42

0.67 0.60 1. 10.5 19.67 2.83

17.70 6.67 6.83 4.5 l 74 5
l4. 1 l. 11. 0

 

 

78

E.1.3 Four Stroke Tecumseh Overhead Valve (OHH50) Engine

TABLE 8: Effect of Ground Electrode Geometry on Overhead Valve Engine

0.00

Temp Exht(C°) 112.73 188.03 1.5 .67
l.
. . l .
169.83 243.30 162.07 76.90 367.57 583.00
8 .
. . l. .
0.03 0.60 1.40 13.90

28.00 6.69 4.78 4.17
4

8.81 11.82 12.86 13.98 12.61 12.23

 

79

 

E.2 Effect Of Ground Electrode Orientation

TABLE 9: Effect of Ground Electrode Geometry (Pyrotek) Spark Plug at Part Load

Torque(Ft- )
Exhaust Temp (C°)

HC )

)
AFR

TABLE 10: Effect of Ground Electrode Geometry (Baseline) Spark Plug at Part Load

9.36

9.88

9.3

 

11.68 11.65 11.66

 

 

 

 

 

 

 

 

 

 

 

 

Orientation 0° 90° 180°

Power(Bhp) 1.2420 1.2500 1.2430

Torque(Ft-Lb) 2.3000 2.3140 2.2940
Exhaust Temp (°C) 288.7300 295.0000 889.3000
CO(gns/hr) 353.7600 350.5000 350.1300

HC’s(gns/hr) 9.2900 8.9600 8.8200
C02(gms/hr) 876.5000 886.1160 888.1400
AFR 11.6800 11.7200 11.7300

 

80

 

TABLE 1 1: Effect of Ground Electrode Geometry (Pyrotek) at Idle

270°

Exhaust Temp (°C) 118 l

.6
) . 1

) 7 . 84
) 243.6 275.7

I

 

TABLE 12: Effect of Ground Elecuode Geometry (Baseline) at Idle

 

 

 

 

 

 

 

 

Orientation 270° 90°
RPM 1709 1574
Exhaust Temp (°C) 107.8 121.7
CO(gns/hr) 82.5 126.13
HC’5(gms/hr) 63.67 5277—“
C02(gns/hr) 223.14 242.936

 

 

 

81

 

E.3. Effect Of Center Electrode Geometry

TABLE 13: Effect of Center Electrode Geometry (V-Notch) on Emissions

0.91 1.78
l.
1.71 3.35

121.03 184 215. 275.17
1.
. . l .
17 377. 170. 117.17
-Groove 16. 59. 1 .4 1

435. 616.23 883.91 1313.

0.1 . . 13.53
. . . l .
16. 1 5. 6
15. 11. 5.3 6.07

10.83 10.81 12.

82

W
2.47
.l .
4.65 .46
328.57 385.97
267. 573
7 .1

1569.87 1773 1

 

E.4.

E.5.

 

Inﬂuence of Spark Gap on Emissions

TABLE 14: Irnﬂuence of Different Spark Gaps on Engine Emissions

Gap
1

) 407. 430.57 420.43 445.37 442.90

Exht Temp (°C) . . . 67 .
-HC ) 9 10.14 10. 10.64

Exht Temp (°C) 234.80 302.47 248.83 286.63 306.03

Effect of Engine Speed on Emissions

TABLE 15: Effect Of Engine Rpm On Exhaust Emissions With Pyrotek Spark Plug

 

 

 

 

 

 

 

 

 

 

 

 

 

Rated RPM 2400 2600 2800 3000
Power(Bl-IP) 1.385 1.452 1.460 1.408
Torque(ﬁ-lb) 3.046 2.942 2.745 2.470
CO(gns/hr) 11.267 19.100 31.100 33.567
C02(gns/hr) 1156.507 1221.033 1273.327 1311.527
NOx(gns/hr) 14.000 13.367 1 1.233 8.700
HCﬁgns/hr) 4.760 4.587 4.120 4.020

Exhaust Temp (°C) 246.367 256.400 258.900 263.667—
AFR 14.28 14.24 14.18 14.18
NOx(gns/hp/hr) 10.067 9.233 7.733 6.200
HC's(gns/hp/hr) 3.437 3.160 2.823 2.853

 

 

 

 

 

83

TABLE 16: Effect of Engine RPM On Exhaust Emissions With Baseline Spark Plug

N

Exhaust Temp (°C)
AF R

 

 

APPENDIX F

Effect of Use of Indexing Washers on Compression Ratio

Engine Data:

Displacement = 10.49 in3

Stroke = 1.938 in3

Bore = 2.625 m3

V, = (Ir/4)D21= 10.48823 n13

Given 7= 7.5 = (V, + V2)/ V2 :5 V2: 1.613573 in3

One turn of spark plug is equal to the 0.050-inch washer if turned in or turned out.
By measurements:

OD of spark plug = 0.5435 inch

II) of spark plug = 0.336 inch

Then AV = 11/4 {(0.5435)2 — (0.336)2 }* 0.050 = 7.1666*10 ‘3 inch3

If we turn in the spark plug by one complete turn then the new compression ratio would
be:

New compression ratio = (V, + V2 —A V)/ ( V2 — AV) = 7.5289

Percentage change in compression ratio 0.386 %

For 1/2 or 1A turn, the effect would be much smaller.

85

LIST OF REFERENCES

 

LIST OF REFERENCES

. Milton, B. E., “Thermodynamics, Combustion and Engines,” Chapman and Hall,
First edition, 1995.

. Danials, C, F., Scilzo, B, M., “ The Effect of Electrode Design on Mixture
Igniteability,” SAE paper 960606, 1996.

. Nishio, K., Oshima, T., Sher, E., “Cyclic Variability In Spark Ignition Engines, A
Literature Survey,” SAE Paper 940987, 1994.

. Hood, S., “ The V-Groove Electrode Spark Plug,” SAE paper 901535,l990.

. Hadjiconstatinou, N,. Mirn, K., Heywood, J. B., “Relation Between Flame
Propagation Characteristics and Hydrocarbon Emissions Under Lean Operating
Conditions in Spark Ignition Engines,” Twenty Sixth Symposium (International) on
Combustion! The Combustion Institute, 1996(PP 2637-2644).

. Bianco, Y., Cheng, W, K., Heywood, J. B., “ The Effect Of Irnitial Flame Kernel
Conditions on Flame Development In SI Engines,” SAE paper 912402, 1991.

. Witze, P, O., “ The Effect Of Spark Location On Combustion In A Variable Swirl
Engine” SAE paper 820044,1982.

. Herweg, R., Begleris, P., Zettlitz, A., and Zeigler, G, F, W., “ Flow Field Effect On
Flame Kernel Formation In A Spark Ignition Engines,” SAE paper 881639, 1988.

86

 

 

9.

10.

11.

Hall, M.J., “ The Inﬂuence of Fluid Motion on Flame Kernel Development And
Cyclic Variation In A Spark Ignited Engine,” SAE paper 890991, 1989.

Anbarasu, A. M., and Abata, D. L., “ Modeling of Heat Loss to the Spark Plug
During Early F lame Growth in A Spark Ignition Engine,” ICE-Vol 23, Engine
modeling, ASME 1995.

Maly, R., Saggau, B., Wagner, E., and Zeigler, G., “ Prospects of Ignition
Enhancement,” SAE paper 830478, 1983.

12. Nakai, M., Nakagawa, Y., Hamai, K., Sone, M., “ Stabilized Combustion in A Spark

13.

14.

15.

16.

17.

18.

19.

20.

Ignited Engine Through A Long Spark Duration,” SAE paper 850075, 1985.

Brereton, G. J., Bertrand, E., and Macklem, L., “ Effect Of Changing Ambient
Humidity And Temperature On The Emissions Of Carbureted Two And Four Stroke
Hand Held Engines,” SAE paper 972707, 1997.

Turns, S. R., “ An Introduction To Combustion: Concepts And Applications,”
Mcgaw Hill , First Edition , 1996.

Gaydon, A. D., and Wolﬂnardt, H. G., “ Flames, Their Structure , Radiation and
Temperature,” Chapman and Hall, Ltd., London, 1970.

Fraser, R. A., and Bracco, F. V., “Cycle Resolved LDV Integal Length Scale
Measurements In An ICE Engine,” SAE paper 880381,1988.

Chen, C., and Veshagh, A., “A Reﬁnement of Flame Propagation Combustion Model
for Spark-Ignition Engines,” SAE Paper 920679.1992.

Lancaster, D. R., Krieger, R. B., Sorenson, S. C., and Hull, W. L., “Effect of
Turbulence on Spark Ignition Engine Combustion,” SAE paper 760160, 1976.

William, F. A., “ Combustion Theory,” Second Edition, Addison- Wesley, Redwood
City, CA, 1985.

Maly, R. R., “ Spark Ignition, its Physics and Effects on the Internal Combustion
Engine,” Fuel Economy In Road Vehicles Powered By Spark Ignition Engines,
Hilliard and Springer, G. S., Plenum Press, 1983.

87

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Mantel, T., “ Three Dimensional Study Of Flame Kernel Formation Around A Spark
Plug,” SAE paper 920587, 1992.

Kanury, A. M., “ Introduction to Combustion Phenomena,” Combustion Science and
Technology, Volume 2, Gordon and Breach Science Publishers, 1984.

Bertrand, E. C., “Effect Of Ambient Conditions of A Small Carbureted Overhead
Valve Four Stroke Utith Engine,” Master Thesis, 1997.

Heywood, J. B., and Pischinger, S., “ How Heat Losses to the Spark Plug Electrodes
Affect Flame Kernel Development In An SI Engine,” SAE paper 900021, 1990.

Anderson, R. W., and Asik, J. R., “ Ignitability Experiments in a Fast Burn, Lean
Burn Engines,” SAE paper 830477,1983.

Ziegler, G. F. W., Schaudt, R. A., and Herweg, R., “ The Inﬂuence of Spark Plug
Orientation On Flame Kernel Formation in Flowing Mixtures,” SAE paper
865031,1986.

De Soete, G. C., “ Propagation Behavior of Spark Ignited Flames in Early Stages,” C-
59/83.

Mattavi, J. N., and Amann, C. A., “ Combustion Modeling in Reciprocating
Engines,” Plenum, Newyork, 1980.

Hacohen, J., Yossaﬁ, D., Thomas, J., Thurly, R., Belmont, M.R., “ Flame Stretch and
Heat Transfer Effects Dnning Flame Initiation in SI Engines,” American Chemical
Society Paper 93359, 1993.

Markstein, G H., “ Non Steady Flame Propagation,” Pergamon, New York, 1964.

Metaghalchi, M., Keck, J.C., “ Burning Velocities of mixture of air with methanol,
Isooctane, and Indolene at high Pressure and Temperatures,” Combustion and Flame,
48; 1982.

Lefebvre, A. H., “ Gas Turbine Combustion,” McGraw-Hill, 1983.

Eccleston, B. H., Hurn, R.. W., “ Exhaust Emissions from Small Utility Internal
Combustion Engines,” SAE Paper 720197, 1972.

Witze, P.O., Hall, M. J., Bennet, M. J., “Cycle Resolved Measurements of Flame
Kernel Growth and Motion Correlated with Combustion Duration,” SAE Paper
900023, 1990.

88

 

 

35. Shen, H., Jiang, D., “ Investigation on the Flame Irnitiation and Early Development in
a Spark Ignition Engine,” SAE Paper 922239, 1992.

36. Semenev, E. S., “Studies of the Turbulent Gas Flow in Piston Engines,” NASA
Technical translation F-97, 1963.

37. Spindt, R. S., “Air Fuel Ratios from Exhaust Gas Analysis,” SAE Paper
650507,]965.

38. Halldirn, C., “ Velocity and Turbulence Measurements Close to a Spark Plug,” SAE
Paper 920154, 1992.

39. Heywood, J, 8., “Internal Combustion Engine Fundamentals,” McGraw-Hill, 1988.
40. Wichman, I. 8., “Introduction to Combustion” Course Notes, 1997.

41. Brown, A. G., Stone, C. R., Beckwitln, P., “Cycle-by-Cycle Variations in Spark
Ignition Engine Combustion- Part 1: Flame Speed and Combustion Measurements
and a Simpliﬁed Turbulent Combustion model,” SAE Paper 960612, 1996.

89

HICHIGQN STnTE UNIV. LIBRQRIES
1|HI111111111111111111111111111111
31293017746128