DEVELOPMENT OF A COMPRESSED NATURAL GAS (CNG) MIXER FOR A

TWO STROKE INTERNAL COMBUSTION ENGINE

DEVARAJAN A/L RAMASAMY

UNIVERSITI TEKNOLOGI MALAYSIA

PSZ 19:16 (Pind.1/97)

Universiti Teknologi Malaysia

BORANG PENGESAHAN STATUS TESIS

JUDUL: DEVELOPMENT OF A COMPRESSED NATURAL GAS (CNG) MIXER

FOR A TWO STROKE INTERNAL COMBUSTION ENGINE

SESI PENGAJIAN: 2004/2005-2

Saya DEVARAJAN A/L RAMASAMY

(HURUF BESAR)

mengaku membenarkan tesis (PSM/Sarjana/Doktor Falsafah)* ini disimpan di PerpustakaanUniversiti Teknologi Malaysia dengan syarat-syarat kegunaan seperti berikut:

1. Tesis adalah hakmilik Universiti Teknologi Malaysia.2. Perpustakaan Universiti Teknologi Malaysia dibenarkan membuat salinan untuk tujuan

pengajian sahaja.3. Perpustakaan dibenarkan membuat salinan tesis ini sebagai bahan pertukaran antara

institusi pengajian tinggi.4. **Sila tandakan ( )

SULIT(Mengandungi maklumat yang berdarjah keselamatan ataukepentingan Malaysia seperti yang termaktub di dalamAKTA RAHSIA RASMI 1975)

TERHAD (Mengandungi maklumat TERHAD yang telah ditentukanoleh organisasi/badan di mana penyelidikan dijalankan)

TIDAK TERHAD

Disahkan oleh

(TANDATANGAN PENULIS) (TANDATANGAN PENYELIA)

Alamat Tetap: No. 70, Jalan Batu Nilam 9 PM DR. ROSLI ABU BAKARKaw. Perumahan Bukit Tinggi Nama Penyelia42000, Klang, Selangor

Tarikh: Tarikh:

CATATAN: * Potong yang tidak berkenaan.** Jika tesis ini SULIT atau TERHAD, sila lampirkan surat daripada pihak berkuasa

/organisasi berkenaan dengan menyatakan sekali sebab dan tempoh tesis ini perludikelaskan sebagai SULIT atau TERHAD.Tesis dimaksudkan sebagai tesis bagi Ijazah Doktor Falsafah dan Sarjana secara penyelidikan, atau disertasi bagi pengajian secara kerja kursus dan penyelidikan,atau Laporan Projek Sarjana Muda (PSM).

I hereby declare that I have read this thesis and in my

opinion this thesis is sufficient in terms of scope and quality for the

award of the degree of Master of Engineering (Mechanical)

Signature :

Name of Supervisor : PM. DR. ROSLI ABU BAKAR

Date :

PART A – Confirmation Of Cooperation*

It was confirmed that this thesis research project was accomplished with the cooperation

between __________________ and _________________.

Confirmed by:

Signature : ____________________________ Date: _______________

Name : ____________________________

Position : ____________________________

(Official Stamp)

* If the thesis/project research involves cooperati on.

PART B – For The School Of Graduate Studies Office Usage

This thesis had been checked and approved by:

Name and Address of External Examiner : __________________________________

__________________________________

__________________________________

__________________________________

Name and Address of Internal Examiner : __________________________________

__________________________________

__________________________________

__________________________________

Name of Other Supervisor (If Available) : __________________________________

__________________________________

__________________________________

__________________________________

Confirmed by Assistant Registrar of SPS:

Signature : ____________________________ Date: _______________

Name : ____________________________

UTM(PS)-1/02

Sekolah Pengajian Siswazah

Universiti Teknologi Malaysia

PENGESAHAN PENYEDIAAN SALINAN E-THESIS

Judul tesis : DEVELOPMENT OF A COMPRESSED NATURAL GAS (CNG) MIXER FOR A TWO STROKE INTERNAL COMBUSTION ENGINE

Ijazah : SARJANA KEJURUTERAAN MEKANIKAL (AUTOMOTIF)

Fakulti : FAKULTI KEJURUTERAAN MEKANIKAL

Sesi Pengajian : 2004/2005-2

Saya DEVARAJAN A/L RAMASAMY mengaku telah menyediakan salinan e-thesis sama seperti

tesis asal yang telah diluluskan oleh panel pemeriksa dan mengikut panduan penyedian Tesis dan

Disertasi Elektronik (TDE), Sekolah Pengajian Siswazah, Universiti Teknologi Malaysia, Januari

2005.

___________________________

(Tandatangan pelajar)

_____________________________

(Tandatangan penyelia sebagai saksi)

Alamat tetap:

N0 70, Jalan Batu Nilam 9,

Bukit Tinggi,

41200 Klang,

Selangor

Nama penyelia: PM DR. ROSLI ABU BAKAR

Fakulti: FAKULTI KEJURUTERAAN MEKANIKAL

Nota: Borang ini yang telah dilengkapi hendaklah dikemukakan kepada SPS bersama penyerahan CD.

DEVELOPMENT OF A COMPRESSED NATURAL GAS (CNG) MIXER FOR A

TWO STROKE INTERNAL COMBUSTION ENGINE

DEVARAJAN A/L RAMASAMY

A thesis submitted in fulfilment of the

requirements for the award of the degree of

Masters of Engineering (Mechanical)

Fakulti Kejuruteraan Mekanikal

Universiti Teknologi Malaysia

OCTOBER 2005

ii

I declare that this thesis entitled “Development of a Compressed Natural Gas (CNG)

Mixer for a Two Stroke Internal Combustion Engine” is the result of my own research

except as cited in the references. The thesis has not been accepted for any degree and

is not concurrently submitted in candidature of any other degree.

Signature: ....................................................

Name : ....................................................Devarajan A/L Ramasamy

Date : ....................................................

iii

This work is dedicated to my beloved ones,

My Father Mr Ramasamy Rengasamy

My Mother Mrs. Ganam Govindan

My Brother Mr.Saravanan Ramasamy

My Sisters Ms Kalaivani Ramasamy and Ms Karthikayeneee Ramasamy

And

My sweetheart Ms Gayadri Krishanan

iv

ACKNOWLEDGEMENT

I would like to express my greatest appreciation and gratitude, to my

supervisor, Dr. Rosli Abu Bakar for his gu idance, patience for giving advises and

supports throughout the progress of this study. Throughout this work, I have learned

much from him. Special thanks are also given to my co-supervisor, Dr Normah

Ghazali for the guidance to comment my thesis.

Not forgotten, special thanks to En Mardani and all the technicians at the

Automotive Laboratory of Universiti Teknologi Malaysia, especially Wak Sairaji, En.

Mazlin, En. Subki, En. Hishamudin and Tuan Haji Wahab. They were not hesitant to

answer all my doubts and spending their time to guide me during my experimental

work.

A great appreciation is acknowledged to the Ministry of Science, Technology

and Innovation for the funding under the Intensified Research in Prioritized Area

(IRPA), vot 74516. The project is hoped to succeed under the guidance and

knowledge of Prof. Ir. Dr Azhar Abdul Aziz.

Last but not least, I would like to thank all of my friends and teammates

especially Mr. Wong Hong Mun, Mr. Gan Leong Ming, Mr. Chong Chin Lee and

Mr. Fadzil Rahim, for their support and encouragement given to me, especially

during the hard times.

v

ABSTRACT

Compressed Natural Gas (CNG) has been accepted widely as an alternative to

gasoline. More importantly the use of CNG in two stroke engines will drastically

reduce the high emission output from these engines as these engines are widely used

around the world. A conversion kit is used to apply the fuel in engines. A bi-fuel

conversion system converts engines without much modification to other systems.

They are normally produced for four stroke application. This kit has to be studied to

be modified for two stroke application. The part that connects the engine to the kit is

called a gaseous fuel mixer. This part mixes the air and fuel due to its venturi shape.

A mixer provides fuel suction at different engine speeds due to pressure difference at

the throat. The optimisation of the throat is important as a small throat will cause

poor performance at high speeds while a large throat will reduce fuel suction. The

smaller throat size creates higher velocity and lower pressure. This low pressure

creates fuel suction into the mixer. The mixer was designed for a two stroke engine

air flow. Computer aided design (CAD) and computational fluid dynamic (CFD)

software were used as a tool for the design. The design is optimised for inlet and

outlet angles, number and size of the hole at the throat circumference and also the

throat size. The prototype design was manufactured based on optimised dimensions

of the mixer that were obtained from CFD analysis. The mixer was validated to show

that the CFD analysis was correct. Testing apparatus were used to do the validation.

The apparatus consists of a laminar flow element (LFE), a smoke generator, a digital

manometer and a gaseous flow meter. It was used to validate the flow pattern,

pressure drop from the mixer and the air fuel ratio given by the mixer.

vi

ABSTRAK

Gas Asli Termampat (CNG) telah diperaku i sebagai satu alternatif kepada

petrol. Penggunaan gas in dalam enjin dua lejang mampu mengurangkan pengeluaran

pencemaran tinggi dari enjin ini. Ini kerana penggunaan enjin dua lejang adalah

banyak di dunia. Bahan api ini digunakan pada engine melalui kit penukaran.

Penukaran enjin petrol ke CNG perlu dilakukan dengan modifikasi kecil pada enjin

asal. Oleh itu, kit penukar CNG dwi-bahanapi digunakan. Unit ini dibuat lazimnya

untuk enjin empat lejang, oleh itu, ia perl u dikaji bagi penggunaan dalam enjin dua

lejang. Bahagian pada alat ini yang bersambung kepada enjin dinamakan sebagai

pencampur bahanapi bergas. Ia menyebabkan gas bercampur pada bahagian yang

berbentuk venturi. Pencampur ini memberikan sedutan gas kepada enjin pada halaju

enjin yang berbeza disebabkan perbezaan tekanan pada bahagian yang dipanggil

leher. Ubahsuai leher adalah penting bagi operasi alat ini. Ubahsuai leher adalah

perlu kerana leher yang kecil akan menyebabkan prestasi enjin yang rendah pada

kelajuan tinggi manakala leher yang besar tidak dapat memberi sedutan gas yang

diperlukan. Tekanan rendah menyebabkan sedutan pada pencampur ini. Pencampur

direkabentuk untuk aliran udara pada enjin dua lejang. Rekabentuk berbantukan

computer (CAD) dan Dinamik Bendalir berbantukan computer (CFD) digunakan

sebagai alat rekabentuk. Rekabentuk pencampur diubahsuai dengan menggunakan

CFD pada sudut masukan dan keluaran, bila ngan lubang dan saiz lubang pada leher

serta saiz leher itu sendiri. Prototaip dibuat berdasarkan dimensi pencampur yang

diperolehi daripada analisis CFD. Untuk membuktikan analisis CFD pengesahan

telah dilakukan. Peralatan ujikaji telah digunakan untuk melakukan pengesahan ini.

Ia terdiri daripada elemen aliran laminar (LFE), penghasil asap, manometer digital

dan meter aliran gas. Peralatan ini digunakan bagi tujuan pengesahan bentuk aliran,

kejatuhan tekanan dan nisbah udara kepada bahan api yang diberi oleh pencampur

ini.

vii

CONTENTS

CHAPTER TITLE PAGE

TITLE i

DECLARATION ii

DEDICATION iii

ACKNOWLEDGEMENT iv

ABSTRACT v

ABSTRAK vi

CONTENTS vii

LIST OF TABLES xi

LIST OF FIGURES xii

LIST OF APPENDICES xiv

LIST OF SYMBOLS xv

1 INTRODUCTION 1

1.1 Problem Statement 2

1.2 Objectives 3

1.3 Scope 3

1.4 Methodology 3

viii

2 LITERATURE REVIEW 5

2.1 Two Stroke Engine 5

2.2 CNG as Fuel for Two Stroke Engines 6

2.2.1 CNG as an Alternative Fuel 7

2.2.2 Combustion Characteristics of CNG 10

2.2.3 Emission Reduction from CNG Usage in

Two Stroke Engines 11

2.2.4 Other Issues Regarding CNG Usage 13

2.3 CNG Mixer 14

2.3.1 Current Trends in CNG Mixer Design 15

2.3.2 Sizing of the Mixer Throat 18

2.3.3 Pressure Drop in the Mixer 19

2.3.4 CNG Mixer and Engine Conversion Kits 23

2.4 Summary of Literature Review 24

3 DESIGN OF A VENTURI BURNER MIXER 25

3.1 Conceptual Design 26

3.2 Procedure of Mixer Design 28

3.2.1 Initial Throat Size 29

3.2.2 CFD Simulations of the Mixer 30

3.2.3 Inlet and Outlet Angles of the Mixer 34

3.2.4 Number of Holes at Throat Circumference 36

3.2.5 Size of Hole at Throat Circumference 37

3.2.6 Throat Size Optimisation 37

3.3 Prototyping the Mixer 38

3.4 Validating the Mixer Design 39

3.4.1 Testing Apparatus 39

3.4.2 Testing Procedure 42

3.4.2.1 Smoke Mixing in Mixer 43

3.4.2.2 AF Ratio Test 43

ix

3.4.2.3 Pressure Drop Test 46

4 RESULT AND DISCUSSION 47

4.1 Designing of the Mixer 47

4.1.1 Initial Throat Size 47

4.1.2 CFD Simulation of the Mixer 48

4.1.3 Inlet and Outlet Angles of the Mixer 48

4.1.4 Number of Holes at Throat Circumference 52

4.1.5 Size of Hole at Throat Circumference 56

4.1.6 Throat Size Optimisation 58

4.2 Prototyping the Mixer 63

4.3 Validating the Mixer Design 65

4.3.1 Smoke Mixing in Perspex Prototype 65

4.3.2 AF ratio Testing of Mixer 67

4.3.3 Pressure Drop Testing of Mixer 69

5 CONCLUSION AND RECOMMENDATION 72

5.1 Conclusion 72

5.3 Recommendation 73

REFERENCES 74

APPENDICES 77

Appendix A 77

Appendix B 79

x

Appendix C 109

Appendix D 117

Appendix E 125

Appendix F 128

xi

LIST OF TABLES

TABLE NO. TITLE PAGE

2.1 Energy content of alternative fuels relative to petrol

and diesel 8

2.2 Proven natural gas reserves 8

2.3 Average natural gas composition in Malaysia 9

2.4 Methane gas properties 10

2.5 Typical 2-stroke emissions 12

2.6 Current regulation that is available for two-stroke

engines 12

2.7 Fuel price 13

3.1 Specification of the analysed engine 29

3.2 Properties of air 33

5.1 Specification of the mixer designed 73

xii

LIST OF FIGURES

FIGURE NO. TITLE PAGES

1.1 Methodology 4

2.1 Operation of a two stroke engine 6

2.2 Type of CNG mixers currently being used in the market 15

2.3 Power test results for different mixer designs 16

2.4 Venturi upstream of the carburettor 18

2.5 Mixer after throttle in intake system of injection engine. 18

2.6 Schematic plot of velocity and pressure across a venturi 20

2.7 Pressure profile during intake stroke of an engine 21

2.8 Pressure drop in air cleaner and intake manifold 22

3.1 Methodology for designing the CNG mixer 25

3.2 The concept models 27

3.3 Proposed shape of the mixer 28

3.4 Location of throat diameter 30

3.5 Simulation steps for each simulation 32

3.6 Overall simulation stages done on the mixer 34

3.7 Simulation model for inlet and outlet angles 35

3.8 Schematic diagram of flow test rig to measure air flow 40

3.9 Schematic of smoke generator connected to test rig 41

3.10 Schematic diagram of pressure measurement 42

4.1 Pressure plot along the centre line of the mixer at different

inlet and outlet angles 49

4.2 Lowest pressure at the throat diffuser wall 50

4.3 Pressure ratios of each model inlet and outlet angle changes 51

4.4 Eight holes mixer model 53

4.5 Ten holes mixer model 54

xiii

4.6 Twelve holes mixer model 55

4.7 Effect of AF ratio on hole sizes at throat circumference

at all speed range 57

4.8 Effect of throat diameter size on air fuel ratio 60

4.9 Simulation pressure drop due to different throat size

at all engine speed 62

4.10 Perspex model for flow testing 63

4.11 Assembled view of Aluminium mixer 64

4.12 Components of Aluminium mixer 64

4.13 Simulation of smoke at 1000 rpm, 2000 rpm and 3000 rpm

air speed 66

4.14 Experiment and simulation results of AF ratio 68

4.15 Simulations and experiment pressure drop 71

xiv

LIST OF APPENDICES

APPENDIX TITLE PAGES

A Thesis Gantt Chart 77

B CFD Analysis 79

C Apparatus and Experiments 109

D Technical Drawings 117

E Material Selection 125

F Mesh Independant Analysis 128

xv

LIST OF SYMBOLS

AF Air fuel ratio -

1A Area in inlet m2

2A Area at throat m2

C Viscosity constant -

Cv Specific Heat J/kgK

Dr Delivery ratio -

HL Losses in pipe Pa

k Turbulent kinetic energy J/kg

1m Inlet mass flow rate kg/s

N Engine speed rpm

aQ Volumetric air flow rate m3/s

1Q Measured flow rate m3/s

2Q Actual flow rate m3/s

atmp Atmospheric pressure Pa

QH Heat source per unit volum e J/m 3

qi Diffusive heat flux J/s

Si Mass-d istributed external force per unit mass N/kg

U Fluid velocity m/s

1v Velocity at inlet m/s

2v Velocity at throat m/s

p Pressure drop Pa

airP Pressure drop in the air cleaner Pa

uP Intake pressure drop upstream Pa

thrP Pressure drop across throttle Pa

xvi

valveP Pressure drop across intake valve Pa

1 Air density at inlet kg/m 3

f Turbulent viscosity factor. -

ij Kronecker delta function -

Turbulent dissipation J/s

Angle º

ik Viscous shear stress tensor Pa

Dynamic viscosity kg/m s

l Dynamic viscosity kg/m s

t Turbulent eddy viscosity kg/m s

CHAPTER 1

INTRODUCTION

Current trends in the automotive industry are ever changing especially

regarding the usage of alternative fuels. The search for the best alternative fuel that

produces the least amount of emission has sparked concerns to many researchers.

Maxwell (1995) stated that many studies on alternative fuel have been carried out

and researchers are looking at natural gas, liquefied petroleum gas (LPG), methanol,

ethanol, and hydrogen. All of these fuels have their advantages and disadvantages

which are cost, availability, environmental impact, usage in vehicle, safety and the

acceptance by consumers.

Current fuel price inflation and also current oil crisis, drastic moves were

taken by many countries to reduce petroleum usage and finding other alternatives to

its usage. In developing countries, the concern of finding alternative fuels has

started and already had become an issue. With gas reserves three times more than

petroleum oil, Malaysia is increasingly turning its attention towards natural gas. The

national petroleum company of Malaysia, PETRONAS has embarked on the

Natural Gas for Vehicles (NGV) program where NGV dispensing facilities are

available at some selected PETRONAS service stations, located in high traffic

density areas of Kuala Lumpur and Johor Bahru. The government support for the

NGV program was seen in 25% reduction on car road tax for using NGV as well as

requiring new taxis in the Klang Valley to use CNG by engine conversion systems.

2

In automotive applications, natural gas can be used in three forms based on

how the natural gas is stored. One of the most popular forms of natural gas is the

compressed natural gas (CNG), which is natural gas in pressurised form. The other

least popular methods of obtaining natural are liquefied natural gas and the

absorption natural gas.

CNG is a good alternative to petrol and diesel. Consumers would easily

accept this form of alternative as it has low operational cost due to subsidised price

and its usage could provide cleaner engine emissions. The main reason behind CNG

fuel being cleaner is that natural gas is principally comprise of 90% methane, which

is the simplest form of hydrocarbon. Even so, the CNG fuel available today still

lack in some qualities compared to petroleum fuel. For example, CNG fuelled

engines normally possess lower engine performance compared to petrol.

The main reason is that CNG fuelling systems creates a lot of losses in

terms of volumetric efficiency. This happens as CNG must be supplied to the

engine through a mixing device before the mixture of CNG and air is drawn into the

engine. This causes less fuel in the combustion chamber and reduces volumetric

efficiency. Currently petrol fuelled engine are converted into a CNG fuelled engine

by means of a fuel mixing device.

1.1 Problem Statement

Currently, there are no specific CNG mixers specifically designed for two

stroke engines in the market. All of the conversion kits that are available for four

stroke engines only. A proper CNG mixer should be designed for two stroke engine

application. A supercharged 150 cc two stroke engine has been chosen for CNG

conversion. Direct usage of a conventional four stroke engine CNG mixer for two

stroke engines is not possible as they are too large a size for a small two stroke

engine air requirements. The design of the mixer has to consider the whole range of

engine operating condition in order to provide a complete view of its performance.

3

The existing four stroke engine CNG mixers are usually not properly refined

and optimised to enable good air fuel mixing. In addition, the efficiency of the

current mixer design is also an issue as it is designed for simplicity which only

offers practicality but lack in efficient air flow performance throughout the engine

speed. Therefore, a straight forward conversion is not possible.

1.2 Objectives

The objectives of the study are as follows:

1) To design a venturi burner type CNG mixer for a two stroke engine

according to the engine’s air requirement using CFD.

2) To fabricate the optimised prototype of the CNG mixer and test it on a

flow bench machine.

1.3 Scope

The scopes of the research are as follows:

1) Preliminary design of the CNG mixer.

2) Optimising the CNG mixer design using CFD as a design tool .

3) Fabrication of the prototype CNG mixer.

4) Testing and validation of the CNG mixer design.

1.4 Methodology

A general methodology was followed in the research as indicated in the flow

chart as shown in Figure 1.1:

4

Start

Literature review

Concept design

Designing of mixer

Figure 1.1 Methodology

Prototyping the mixer

Meet design criteria

No

Yes

Validating the mixer design

End

CHAPTER 2

LITERATURE REVIEW

2.1 Two Stroke Engine

Air pollution in many Asian cities is increasing due to the proliferation of

vehicles powered by simple two-stroke cycle engines.In Asia, there are an estimated

70-100 million two stroke engines operating which are motorcycle, tricycle and auto

rickshaws to name a few.

In a two stroke engine, there is a power stroke in every revolution whereas

there is only one power stroke for every two revolutions in four stroke engines. With

this, it can be said the two stroke engine can produce power better than the four

stroke engine (Bryan, 2002). The higher power production allows the two stroke

engine to have a higher power to weight ratio and are simpler in design. Evolutions

of the two stroke engines have seen many changes in their designs. There are designs

of two stroke engines with single or multi cylinders, with turbochargers or even with

superchargers.

The basic operation of a two stroke engine has not differed much from how

the first two stroke engine was designed. Figure 2.1 describes an operation of a two

stroke engine with a supercharger system.

6

Compress

Air drawn intothe engine driven supercharger to be

compressedIgnition

Piston descends due to Compressed air sent

to the aircompartment

combustion

Figure 2.1 Operation of a two stroke engine

As seen in Figure 2.1, the intake of a two stroke engine is based on the

scavenging process. Scavenging occurs when the intake ports are uncovered, the

compressed air and fuel pushes in and displaces the remaining exhaust gases. There

are many types of scavenging process that is used in a two stroke engine among them

are cross, loop or uniflow scavenging (Ferguson, 2001).

2.2 CNG as Fuel for Two Stroke Engines

CNG is a type of alternative fuel to petrol and is vastly accepted around the

world. With current fuel price inflation and also current oil crisis that is going on,

drastic moves have been taken by many countries to reduce petrol usage and find

other alternatives to its usage.

Compressed air fuel mixtureforced into

cylinder through intake port

Exhaust escapes through exhaust

valves

Fuel is mixed with compressed air

in air compartment

Scavenging

7

2.2.1 CNG as an Alternative Fuel

By definition, alternative fuels are fuels that can be derived from non-crude

oil resources. Crude oils are petroleum based fuels. Some of the types of alternative

fuels that are available today which are not from crude oil are Natural Gas (NG),

Liquefied Petroleum Gas (LPG), Methanol, Ethanol, Hydrogen and others (Maxwell,

1995).

Amount energy released by burning of fuel in air is dependant on the type of

fuel. This was seen in Table 2.1, which shows in lower heating values for some

known alternative fuels. Judging by their energy density, ethanol and methanol must

be burned more to produce energy equal to that of petrol. Other fuels such as NG,

LPG and hydrogen also have a lower density compared to petrol. With this they can

provide more energy per kilogram equivalent of petrol. Vehicles running in any of

these fuels are more efficient as they produce more energy for the same given mass

of petrol. Here engine design plays an important role in providing the efficiency to

the engine to make full use of the higher energy content derived from these fuels.

Due to its high energy content and large availability, NG has been chosen to

be studied further worldwide. Table 2.2 shows the distribution and availability of NG

throughout the world. NG has already been used in more than 1 million vehicles in

the world since 1993 and the number is increasing as more consumers are exposed to

the numerous benefits of NG. It has been used for domestic and industrial sectors

(heating, thermal energy production, chemical industries). NG has also been labelled

as a clean fuel in ecological considerations (Poulton, 1994).

In Malaysia, the national oil and gas company PETRONAS had introduced

the NGV program in 1986. Now there are almost 4,000 taxis called the Enviro 2000

vehicles in Kuala Lumpur since 1998. It was expected that the Malaysian car

manufacturer, PROTON would be producing around 40,000 NGV’s by 2004 for the

Malaysian market (Taib Iskandar Mohamad, 2003).

8

Table 2.1 Energy content of alternative fuels relative to petrol and diesel (Maxwell,

1995)

Fuel Density

(kg/m3)

Energy Content

(MJ/m3)

Energy Relative

to equivalent

mass of Petrol

Energy Relative

to equivalent

mass of Diesel

Petrol 621.8 4257 100% 91%

Diesel 622.2 4694 110% 100%

LPG 422.1 3113 115% 109%

Methanol 658.5 2100 49% 45%

Ethanol 652.5 2813 66% 60%

NG 351.2 2814 120% 113%

Table 2.2 Proven natural gas reserves, 1991, (Poulton, 1994)

Area

Trillion

Cubic

Meters

Billion

Tonnes Oil

Equivalent

Share of Total

(%)

North America 7.5 6.7 6.1

Latin America 6.8 6.1 5.4

Western Europe 5.1 4.6 4.1

CIS/E Europe 50.0 45.0 40.4

Middle East 37.4 33.7 30.1

Africa 8.8 7.9 7.1

Asia/Australasia 8.4 7.6 6.8

Total 124.0 111.6 100.0

The growth and development of NGV industry in Malaysia was slow due to

the lack of refuelling stations and unavailability of original equipment manufacturer

(OEM) of gas conversion kits. On the contrary, the government is supportive as they

had given import duty and sales tax exemption on the conversion kits. Tax reductions

of 25% for bi-fuel vehicles and 50% on monogas vehicles were given to boom the

industry by the government.

9

Malaysia is currently using EURO II emission regulation and will continue

the standard until 2010 (Jitendra Shah, 2001). The exhaust emission by NGV

vehicles are well below EURO II limits on carbon monoxide, hydrocarbon and

nitrogen oxide. In addition, these vehicles can travel up to 170 km per filling of NGV

rendering its performance equivalent to a petrol-powered vehicle. More importantly,

these NGV vehicles can generate a significant saving on fuel expense as natural gas

is cheaper compared to petrol (Yusoff Ali, 2003).

The composition of CNG also varies between countries. The principle

ingredient of CNG is methane. Methane makes up to 90 percent of CNG. Aside of

methane, the composition contains small portions of other gases such as ethane,

propane, butane, pentane and hexane. It can also contain nitrogen, helium, carbon

dioxide and hydrogen sulphide. Malaysia owns the 12th largest NG reserve in the

world according to Gas Malaysia (2003). The average NG composition in Malaysia

is shown in Table 2.3, while the properties of methane are shown in Table 2.4.

Table 2.3 Average natural gas composition in Malaysia (Gas Malaysia, 2003)

Natural Gas Composition Percentage

Methane 92.73

Ethane 4.07

Propane 0.77

n-Butane 0.06

i-Butane 0.08

Other Hydrocarbon 0.01

Nitrogen 0.45

Carbon Dioxide 1.83

10

Table 2.4 Methane gas properties (Gas Malaysia, 2003)

Methane characteristics

Density (kg/m3) 0.715

Gross Calorific Value (kCal/cm3) 9530

Burning Velocity (m/s) 0.3

Upper Flammability Limit 15.4

Lower Flammability Limit 4.5

Auto Ignition Temperature (oc) 640

Theoretical Air Requirement (m3) 9.74

2.2.2 Combustion Characteristics of CNG

Fuel energy can be harvested only through combustion. Any fuel that

combusts completely will produce simple by-products. A complete burning of fuel in

air to produce this condition is called a stoichiometric condition. When a

stoichiometric air fuel mixture combust, it will produce energy from combustion,

water vapour, carbon dioxide while other composition such as nitrogen and inert

gases will remain constant. Theoretically, combustion burns the fuel completely

without much emission such as carbon monoxide, hydrocarbon from excess fuel and

nitrogen oxides from reaction with surrounding air.

Stoichiometric air fuel (AF) ratio was calculated by first knowing the

combustion chemical reaction of the fuel in air. AF ratio is defined as the mass of air

to the mass of fuel in a mixture. Knowing CNG consists of 90% methane (CH4), the

chemical reaction of CNG fuel is therefore given by:

OHCOOCH Energy2224 22 (2.1)

The equation states that 1 mole of methane would completely combust with 2 moles

of oxygen to produce energy as well as 1 mole carbon dioxide and 2 moles of water.

11

Air is a composition of many gases, where 20.95% of the whole composition

is oxygen and the rest are nitrogen and other gases. Since the mass of 1 mole of air is

28.96 g, therefore 1 mole of oxygen is contained in a mass of 138.23 g of air (Lenz,

1992). On the other hand, the mass of 1 mole of CNG or methane from equation 2.1,

is given as (1 X 12.01 + 4 X 1.008) or 16.042 g. By knowing the mass of each mole

of air and CNG, the stoichiometric AF ratio was found as stated below.

(2.2)

23.17042.16

46.276

042.16

23.1382

ratioAFfmam

According to Yeap (2002), for idling at normal engine operating temperature,

the engine still demands a fairly rich mixture, for which the AF ratio usually in the

range of 11:1 to 13:1. A good mixer must be able to meet these requirements. The AF

ratio becomes too rich and incombustible for the engine at 9.77:1. At too rich

conditions the engine will have less oxygen to burn with the fuel, causing a rise in

hydrocarbon emission or in the worst case it will stall the engine.

2.2.3 Emission Reduction from CNG Usage in Two Stroke Engines

The main problem that occurs in a two stroke engine is they are characterised

by very high levels of hydrocarbon (HC), carbon monoxide (CO), and particulate

matter (PM) emissions. Table 2.5 shows the current emissions of two-stroke engines.

Clearly from Table 2.5, it can be seen that the usage of CNG in two stroke engines

can drastically reduce the emissions produced by normal petrol fuelled engines.

12

Table 2.5 Typical 2-stroke emissions (Bryan, 2002)

Type Of Engine CO

g/kW-h

HC

g/kW-h

NOx

g/kW-h

55 kW 2-stroke

outboard engine 185 134 2.74

7.4 kW 2-stroke

Johnson OMC-J10RCSE

Engine

519 236 0.73

Cooper GMVC-10C 2-stroke

natural gas engine1.88 9.1 0.67

Comparisons of emission are generally based on standards followed by

regulating bodies of a country. All vehicles should follow these regulations in term

of emission. The guideline for two stroke engine emission in India and Unites Stated

of America is shown in Table 2.6. The overall emission of the two stroke engine

using CNG is lower than standard emission regulation of two stroke engines. It was

only slightly higher in term of HC emission of a small engine regulation standard in

California. The California emission standard is considered one of the highest

standards in emission regulations in the world.

Table 2.6 Current emission standard that is available for small engines by

experiments (Bryan, 2002)

CO

g/kW-h

HC

g/kW-h

NOx

g/kW-h

Indian 2-stroke

Genset regulation 603 603 -

US EPA 2-stroke

emission factor, 1999 522 206 0.67

California 4-stroke emission

factor < 19 kW (25 hp) 322 5.4 2.4

Euro II Standard * 4 7 0.15

* Euro II standard is the emission standard followed by Malaysia

13

2.2.4 Other Issues Regarding CNG Usage

Apart from having lower emissions than petrol, the CNG fuel also has its

advantages and disadvantages in other areas. Advantages of using CNG fuel in

vehicles include (Rosli, 2002):

a. Higher Octane number in the range of 120 to 130, which is

considerably higher than 93 to 99 Octane for petrol. A high Octane

number ensures that CNG fuel can run at high compression ratio

without any knocking phenomena to the piston that will cause damage

to the engine.

b. Higher flammability compared to petrol that makes it appropriate to run

on lean burn technology.

c. Safer; as it is lighter and dissipates quickly. Due to this it ignites

quickly, but only when the fuel to air ratio was between 5 – 15% by

volume.

d. Because it is a clean burning fuel, it reduces the required maintenance

cost of vehicle; it can be half of petrol—oil changes can be done for

more than 15,000-30,000 km, spark plug points can be changed at

intervals up to 120,000 km.

e. Plenty of reserve; there is an estimated 65-70 year supply of natural gas.

Besides made from fossil fuel, natural gas can also be made from

agricultural waste, human waste and garbage.

f. Cheaper per litre equivalent than petrol, in Europe 14-17% less than

petrol and 12-74% less expensive than diesel. In Malaysia, the CNG

price is half less compared to petrol as shown in Table 2.7

Table 2.7 Fuel price

FuelPump Price/Litre 1)

CNG RM 0.56Petrol (Unleaded) RON 97 RM 1.41Petrol (Unleaded) RON 92 RM 1.36Diesel RM 0.83

Note: 1) Source: Petronas pump price (2005)

14

However, CNG fuel has some disadvantages that limit its potential to achieve

the optimum engine performance, are as stated below:

a. Since CNG is available in gaseous form, it has a low density. CNG

from the mixer drawn into the engine cylinder displaces approximately

8% to 10% of Oxygen by volume. This reduces the amount of Oxygen

due to larger space occupied by the CNG in the combustion chamber.

b. CNG has a low flame speed. Its burns slower than conventional fuels,

such as petrol and diesel. As much as 60% decrease in burning velocity

has been measured. This prolongs the total combustion duration

compared with diesel and petrol. This can cause a further reduction in

the engine output of 5 to 10%.

Even though CNG has disadvantages, the advantages outweigh the

disadvantages.

2.3 CNG Mixer

The principle characteristic of a CNG mixer was known to analyse the

operation of the mixer. The operation of a mixer is that the change in velocity

causes a change in pressure in the contraction passage which in turn effects a

change in flow of the fuel to join and mix with the main airflow in the required

proportion (Heywood, 1988).

The sizing of the mixer was basically based on the air flow that is drawn by

any engine (Maxwell, 1995). A mixer would also cause pressure drop to the air flow

as it was a device that restricts air flow to create fuel suction. A general

understanding of pressurised flow was also studied to understand pressure drop in

the CNG mixer.

15

2.3.1 Current Trends in CNG Mixer Design

The venturi mixer acts as a carburettor to meter the amount of fuel to the

engine. One research that show the characteristic consideration of the CNG mixer

was done while studying the exhaust gas recirculation (EGR) system. Baert (1999),

found that requirements for a good mixer are as follows:

1) A compact design of the mixer for the EGR system

2) Minimal flow restriction during the intake process. This means less

pressure difference.

3) A good suction pressure mainly in the throat due to venturi effect

from the pressure difference. This will enable more fuel to be sucked

into the system.

CNG mixers have also evolved in design. There are many types of mixer

available. The main three types of mixer design were identified as the venturi, fan

and venturi-burner mixer as shown in Figure 2.2 (Roslia, 2002).

Venturi Fan

Venturi-BurnerFigure 2.2 Type of CNG mixers currently being used in the market

According to (Mardani, 2003), the design with combination of venturi and

burner mixer had better performance than all the other mixer type. By measuring

16

the power output a deduction was found that the venturi-burner type of mixer had

the 5% value closer to petrol power output. This is as shown in Figure 2.3.

15

20

25

30

35

40

45

50

55

60

1500 2000 2500 3000 3500Engine Speed (rpm)

Pow

er (

kW)

Petrol

CNG venturi-burner

CNG venturi

CNG fan

Figure 2.3 Power test results for different mixer designs (Mardani, 2003)

In Mardani’s research, the inlet and outlet angles for the design are obtained

by looking at the pressure drop caused by the angle. The angle that produced lower

pressure drop was chosen as it does not restrict the engine. Apart from this, the

angles are also sized by the amount of turbulence and velocity at the throat. The

angle that has higher turbulence with lower velocity would cause a homogenous

mixture of AF. The design which consists of the combination of 60° of inlet angle

and 30° of outlet angle with 8 holes at the throat was proved to increase the engine

performance using CNG to standard of petrol engines (Mardani, 2003).

Apart from the angle, the holes of a CNG mixer are normally designed at the

throat of the mixer. This was found by Luiz (1996), who developed the Mercedes-

Benz Natural Gas Engine M 366 LAG with a lean burn system by incorporating a

venturi type mixer. The gas was added into the air stream from the holes of the mixer

at the narrowest section of the venturi (throat). The mixer was placed after the

turbocharger, near the intake valve. The air fuel ratio was controlled by the

17

characteristic of the pressure regulator, mixer and the gas flow valve which is

commanded by the engine management system.

Another important parameter is the size of the throat. The throat size affects

the engine performance. Maxwell (1995) observed this when testing on the GMC

(5735cc) engine. The engine was installed on a SuperFlow Model SF-800

dynamometer to obtain data such as engine speed, brake torque, brake horse power,

water temperature, inlet air temperature, exhaust temperature and barometric

pressure. The CNG mixer also gives lower flow rate compared to the General Motors

throttle body. This means the mixer causes a large restriction to the airflow. Peak

power was reduced by 16.2% and 19.7% for small and large venturi respectively

when running on CNG. They concluded that the large venturi produced less suction

than the small venturi hence induced less fuel to the engine. The power loss of about

10 percent was due to natural gas occupying significant portions in the intake system.

Other parameters that are also considered important in a mixer design are

nozzle distance from inlet, venturi tube diameter, throttle opening and mixture ratio.

This was observed by Mikio (1998), after looking at the mixing of natural gas and air

in a two-dimensional CNG mixer. The Schliren method was used to see the mixing

effect. The parameters were found to affect the mixing of natural gas and air in a

CNG vehicle. The inadequate mixing gave an adverse effect on engine combustion

and emission characteristics.

The location of the mixer in the intake system of the engine is another

parameter to note. The positioning can be done by various methods. According to

Lenz, (1992), the location of the mixer for a carburettor engine is as shown in Figure

2.4. It is applied before the throttle. In a fuel injection engine the mixer was usually

put after the throttle as shown in Figure 2.5.

18Air Inlet

GasSupply

Figure 2.4 Venturi upstream of the carburettor

Figure 2.5 Mixer after throttle in intake system of injection engine

AirInlet

GasSupply

2.3.2 Sizing of the Mixer Throat

As mentioned earlier, the sizing of the mixer throat was done based on the

air requirements of two stroke engine. The required air flow (Qa) depends on the

engine speed (N), the displacement of the engine (Vd) and the delivery ratio (Dr).

Knowing this, the air flow rate in a two stroke engine was calculated as,

Da VNDrQ (m3/s) (2.3)

From the equation 2.3, the delivery ratio is important as it depicts the

amount of air that will be drawn into the engine (Ferguson, 2001). The delivery

ratio was defined as,

Delivery ratio, Dr =densityambientvolumedisplaced

cycleperairdeliveredofmass (2.4)

19

Different engines provides different delivery ratio. The delivery ratio was

found to be as low as 85% of the airflow for naturally aspirated engine or it was as

high as 150% for supercharged engines (Willard, 1997).

Mohamed (1998), cited the highest suction that was created by the air stream

occurs only when the air velocity is traveling in near boundary of compressible and

incompressible flow, which is at 150 m/s. The contraction of the throat causes a rise

in velocity linearly and the highest velocity at the throat could not be increased above

150 m/s without neglecting the change in density of the flow. Knowing throat

velocity as 150m/s, the area of the throat was found from the continuity equation.

Thus, by solving the circular cross sectional area of the mixer’s throat, the diameter

of the mixer throat is found.

2.3.3 Pressure Drop in the Mixer

As mentioned previously in Bernoulli principle, the faster the fluid travels at

the throat of the venturi, the lower the surrounding pressure at that region as shown

in Figure 2.6.

Pressure at the throat is lowest when the velocity is the highest. When the

flow exits the venturi the pressure does not return to its original value. This is the

overall pressure drop due to the restriction of the throat as shown in the Figure 2.6.

This also occurs in a CNG mixer as the shape of the mixer is similar to the venturi.

20

321

Diffuser angleVenturi

Y

X

Pressure

Velocity

FlowDirection

Figure 2.6 Schematic plot of velocity and pressure across a venturi

Applying an ideal full sized venturi in the engine is almost impossible as the

venturi needs a long length to reduce pressure loss in the flow. To compromise this

carburettors use air funnels at the throat to provide minimal pressure loss and

creating the largest vacuum at the narrowest point. The overall pressure loss was

found very little during low flow separation occurring at funnel diffuser angles of 7°

to 12°. The usage of funnel can reduce diffuser length and the overall length of a

mixer (Lenz, 1996).

The pressure drop in a venturi can be calculated by using the Bernoulli’s

equation which is,

LHgzvP 2

21

= constant (2.5)

Since the axis mixer was along the horizontal axis, the effect of gravity was

neglected and a more simplified Bernoulli was arrived for two points as in Figure 2.6

as,

LHvPvP 222

211 2

1

2

1 (2.6)

21

The pressure drop is obtained by subtracting the pressure of the flow entering

the mixer to the flow exiting the mixer. This can be given by,

LHvvPPP 21

221221 2

1 (2.7)

Losses in pipes occur as the term HL and computations of the losses needs

also to be included. A simulation and experimentation was seen a good way to

compute the losses and also the pressure drop (Lenz, 1996)

A clearer picture of how the pressure drop occurs in an engine is shown in the

profile of pressure during the intake in Figure 2.7.

P0

airP Po, To

IntakeManifold

uP

thrP TDC BDC

valveP

P P

Engine

Figure 2.7 Pressure profile during intake stroke of an engine (Heywood, 1988) Po, To

is pressure and temperature at atmospheric condition; airP , is the pressure loss in

the air cleaner; is the intake losses upstream of throttle; , is the losses

across throttle; losses across intake valve.

uP thrP

valveP

Heywood (1988), also found that a large pressure drop would cause

performance problems for the overall engine operation. This is because a venturi

arrangement can only meter fuel over a certain range of flow rates and pressures.T

22

The selection of small throat size of mixer will cause a higher restriction of air

entering to the intake manifold at high engine speeds. As flow rates increase, the

venturi will begin to "choke". Consequently, the engine will not achieve desired

operation with respect to increased throttle opening.

At the other extreme, when the velocity of the air in the venturi was very low

for example during idle or start-up, the pressure drop across the venturi becomes

small. A too small pressure drop due to a larger throat also causes less suction to at

the mixer throat. This extremity concerns with engine starting, idle and low-speed

throttle response to inadequate suction. The advantage of slightly oversized mixer

appears on over-the road applications. This was due to the fact that it achieves

optimum performance at higher engine speeds.

In a specific throat size of the mixer the overall pressure drop of the device

can be seen. The pressure drop across the venturi will increase as the engine speed

increases. This condition is similar to any component that restricts air flow to the

engine. Figure 2.8 shows how the pressure drop graph is plotted for components in

the intake manifold (Heywood, 1988).

0

10

20

30

40

50

60

70

0 1000 2000 3000 4000 5000 6000 7000 8000

Engine Speed,rev/min

DP,

mm

Hg

Air Cleaner

Patm

Pthrottle

Throttle

Patm-Pthrottle

Figure 2.8 Pressure drop in air cleaner and intake manifold, Patm is the atmospheric

pressure and Pthrottle is the pressure after the throttle (Heywood, 1988)

23

Since the pressure drop will increase as the air flow in the mixer increases the

fuel flow will also increase due to suction created by the lower pressure at the throat.

This reduces AF to a richer mixture. Generally engines require a high amount of fuel

at high speeds and the mixture of air and fuel tends to get richer as the engine speed

increases.

The mixer that can create more suction of fuel can have smooth engine

operation as fuel flow will not bottleneck the performance of the engine at high

speeds. Nevertheless, a too rich mixture would also cause the engine to stall as the

fuel is more than oxygen to burn. As discussed earlier on CNG fuel mixture, a

stoichiometric mixture is important to enable proper fuel combustion. With this the

mixer should allow the air fuel mixture to be near the stoichiometric range.

2.3.4 CNG Mixer and Engine Conversion Kits

The CNG mixer is the part that connects the engine with a CNG conversion

system. The system itself is made from many components that functions to bring fuel

to the mixer when the engine is operated. There are many types of conversion that

can be used to convert an engine to run on CNG. The conversion systems can be

grouped into three groups which are:

1. Bi-fuel engines, this is a spark ignition petrol engine converted to

natural gas by fitting various components such as a gas

mixer/carburettor, regulator, shut-off valves, control systems and fuel

storage tanks. This arrangement retains petrol fuel system which can be

used when CNG refuelling facilities is unavailable.

2. Dedicated natural gas engines, which are engines optimised for natural

gas. They can be made from petrol engines or be designed mainly for

CNG usage.

3. Dual-fuel engines, these are diesel engines that operate with mixture of

natural gas and diesel fuel. Here, a pilot fuel (diesel) is injected within

the gas mixture so that ignition can occur. This injection is required

because natural gas has a low Cetane rating which is suitable for

24

compression ignition engines. The pilot fuel will ignite first to start the

ignition process.

In the market bi-fuel conversion was a popular option as the user can change

between the fuels without making drastic changes to the existing engine. In the

research the bi-fuel conversion kit was chosen to be studied as it can be obtained

readily from the market as in Appendix C.

2.4 Summary of Literature Review

The literature obtained for two stroke engine and CNG fuel was a preliminary

knowledge for the development of the mixer. The different designs of CNG mixer

showed that the venturi-burner mixer was better and has better performance in actual

testing.

The working principle of the mixer was based on the venturi principle from

Bernoulli equation. It creates a low pressure in the throat area so that natural gas can

be sucked into the engine. The calculation of throat size was seen possible by

knowing the air speed and usage of continuity equation.

Pressure drop causes fuel suction. Due to this, fuel and air mixing

characteristic must reach a certain characteristic that will enable proper operation of

the mixer in the engine. A near to stoichiometric operating condition will allow the

mixer to work efficiently by mixing the fuel and air to the required burning condition

of the engine so that complete combustion of the fuel occurs.

CHAPTER 3

DESIGN OF A VENTURI BURNER MIXER

The research is based on the CNG mixer development for the two-stroke

engine. The flow of the research was carried out as shown in Figure 3.1. The findings

of the study have aided the researcher to begin the designing process of the mixer.

Start

Conceptual Design

Procedure of Mixer Design

Computational Fluid Dynamics Simulation

Design and Fabrication

Validating the Mixer Design

End

Figure 3.1 Methodology for designing the CNG mixer

26

3.1 Conceptual Design

Firstly the design process starts with a concept design. The initial idea for the

shape

ed for 150 cc supercharged two stroke engines.

ent.

at

l angles were used for the inlet and outlet angles to reduce

made along the circumference of the throat to dispense the

nents rather than a single part as the

terial was considered to be used to

s the venturi burner type of mixer shape was for design, many ideas were

though

he first concept in Figure 3.2(a) shows a two inlet mixer, the two inlets

require

design was not developed and needs some improvements.

of the mixer was done from some sketches. The sketches were developed

based on the design criteria. This had been predetermined for the mixer design by

considering machining process involved in the prototyping of the mixer. The design

criteria were set as follows:

The mixer was siz

The length of the mixer must be 60 mm to fit the engine compartm

It has to be a venturi burner type CNG mixer with a diffuser angle of 7°

the throat.

Symmetrica

mixer size.

Holes were

CNG fuel to the air stream evenly.

The model was divided into compo

process of machining would be easier.

A light weight and easy to fabricate ma

make the mixer.

A

t for the shape of the mixer. These shapes are as shown in Figure 3.2. The

shapes were in a development stage, where a constant improvement of the previous

design was made. Each shape was discarded until a final acceptable design was

reached.

T

modification to the piping of the CNG kit. Having two pipes for CNG also

creates a burden for installation of the mixer. The complexity of the shape also

demands high machining cost as the process of machining identified are casting and

rapid prototyping. This increases the production cost of the design. Due to this, the

27

Figure 3.2 The concept models

Figure 3.2(b) show ngential porting with one

CNG inlet. This idea was to give a swirl to the CNG flow but was quiet hard to be

fabrica

bricated. Figures 3.2(c) is a

oncept model with a section view. The model was split in components rather than a

whole

(b)

(c)

(a)

s the concept of the mixer with ta

ted with conventional machining. A casting process was still needed. The cost

of fabricating this shape is also estimated to be too high.

The design had to be cost effective and easily fa

c

part for machining. This also reduces the machining cost as complex

machining is not needed. A turning process was chosen to create this concept. This

concept model also uses circular holes made along the circumference of the throat.

28

The holes are easily made by drilling process. Casting process was not needed for

this concept.

The last concept model was accepted as the initial shape of the mixer. This

design was found to be easier to be fabricated and cheaper due to its simple

machining process. The concept was further refined until the shape in Figure 3.2 as

the proposed design. Further simulation in CFD will continue to optimise the

proposed design.

Proposed shape of the mixer

.2 Procedure of Mixer Design

ed to start the design of the mixer. The design

riteria that were outlined were fulfilled in the process of designing. The design starts

Inlet

Outlet

Throat

CNGInlet

Hole at the throat

Figure 3.3

3

The conceptual design was us

c

by finding the initial throat size of the mixer by using equations found from

literature. The other dimensions of the mixer was then finalised from CFD simulation

and analysis. CFD was a tool to simplify the design process.

29

3.2.1 Initial Throat Size

based on the specification given by a two stroke engine.

ns of the engine are as shown in Table 3.1. Hoses that fit the two

stroke

an, 2002)

The mixer was sized

The specificatio

engine were found to be of 36 mm diameter. The size was used to target the

outer diameter of the mixer so that the mixer can fit the hoses available in the market.

A thickness of 2 mm was given to the mixer walls for connecting purposes, thus

making the inner diameter of the mixer to be sized at 32 mm.

Table 3.1 Specification of the two stroke engine (G

Volumetric displacement 150 cc

Bore x Stroke 57.6 mm

Maximum engine speed 8000 rpm

Type of scavenging system Uniflow scavenging

Effective compression ratio 8.2866

Maximum combustion pressure 49 bar

Maximum temperature 2688.88 K

Type of fuel Petrol

Supercharger Blower

Since the inner diameter was predetermined by hose size, the initial size of

e mixer throat was calculated from air flow characteristics of the two stroke engine.

In this

th

engine the air flow was increased by the use of a blower mechanism. Due to

this the engine air flow is calculated by taking the deliver ratio of the two stroke

engine as 150% as suggested by Willard (1997). The equation for air flow of the two

stroke engine at any given rpm was given by,

60

5.1 Da

VNQ (m3/s) (3.1)

By knowing the air flow of the engine, the initial approxim tion of the throat

ize was computed by using continuity equation. An assumption of inviscid and

a

s

30

steady flow was used. As known earlier in Chapter 2, the throat size was the

minimum cross section area of the mixer that makes the air flow reaches speed of

near compressible flow. Figure 3.4 shows where the throat is located in the mixer.

Maximum air flow rate of the engine was used to find the throat area.

Figure 3.4 Location of throat diameter

V1

ThroatInlet Outlet

V2D2D1

The maximum equation 3.1, using

e maximum engine speed of 8000 rpm. The air speed at the throat was taken as 150

m/s as

theoretical air flow rate was found from

th

discussed in literature. Thus, the throat area was obtained by solving the

continuity equation for incompressible flow as,

22

112 V

Q

V

VAA a (3.2)

The initial throat diameter, D2 was then obtained by solving the circular area

btained from equation 3.2. The throat diameter was used as a reference for the

.2.2 CFD Simulations of the Mixer

G mixer continues to achieve the criteria set

arlier in the concept stage. Simulations need to be done to save time and cost of

designi

o

beginning step of designing the concept model of the mixer.

3

The process of designing the CN

e

ng the mixer. As the air flow in the mixer is governed by fluid mechanics, the

simulation of fluid motion would be needed to analyse the mixer. A computational

31

method is chosen to calculate the fluid motion, by using a commercial CFD package,

CosmosFloworks 2001 (Rosli, 2004).

In CFD, the fluid motions are determined by solving the the Navier-Stokes

equation of mass, momentum and energy equation. The three equations can be

written in the conservation form as follows:

0ukkxt

(3.3)

ii

ikkik

i Sx

Puu

xt

u (3.4)

Hkkiikkkk

QuSuquPExt

E (3.5)

here u is the fluid velocity, W is the fluid density, Si is a mass-distributed external

tforce per unit mass, E is the to al energy per unit mass, QH is a heat source per unit

volume, ik is the viscous shear stress tensor and qi is the diffusive heat flux.

Turbulence normally occurs in any fluid flow. Prediction of turbulent flow

was done by the Reynolds averaged Navier-Stokes equations. Through the averaging

procedure, extra terms known as the Reynolds stresses would appear in the

equations. Following Boussinesq's assumption, the Reynolds stresses was defined in

this model as:

ijijl

l

i

j

j

it

Rij k

xu

x

u

xu

32

32

(3.6)

t is defined using two basic turbulence properties, namely the turbulent kinetic

energy, k and the turbulent dissipation rate, . A constant is also derived empirically

as, C in the equation with value of 0.09.

2kCft (3.7)

hus, theT k model is used to solve turbulence by the software package.

Generally, the CFD analysis follows steps shown in Figure 3.5.

32

Figure 3.5 Simulation steps for each simulation

After the initial mixer modelling w CAD, the CFD analysis continues with

mesh generation. It is done by splitting th volume of fluid in the model into small

volume

boundary conditions inputs were given to the CFD software to begin the calculations.

The ac

t

ith

e

s which is called mesh and iterates the calculation for each volume The CFD

package uses rectangular mesh to the volume of the fluid in the design (Rosli, 2003).

When a specific a mesh was reached as in appendix F, the mixer model’s

curacy of the boundary is important as it would determine the simulation

outcomes. Three boundaries are needed to analyse the CNG mixer they are,

1. Air inlet

2. Flow rate at outlet

3. CNG inle

Modelli erng of mix

Meshing of model

Sett ioning boundary condit

Iterations

Satisfactory

YesNo

Start

Ana ltslysing resu

End

33

1. ir inlet

oundary is the air inlet. A static air inlet was given assuming the air

at the starting of the simulation was not moving. The properties of air were taken as

shown

Pressure 101325 Pa

A

The first b

in Table 3.2. The values of pressure and temperature were taken from an

aneroid barometer from the lab. The other values are interpolated from fluid dynamic

air properties found in any fluid properties table at the given pressure and

temperature. The air property was used as the first input to the inlet of the mixer.

Table 3.2: Properties of air

Temperature 299.15 KSpecific Heat, gKCv 1006 J/kDensity, 1.225 kg/m3

Dynamic viscosity, 1.80x10-5 kg/ms

2. Flow

e engine operates the air will get sucked into the intake

anifold. This causes the flow rate at the outlet of the mixer. The suction flow rate

was cal

undary was the CNG fuel inlet. The fuel is in ambient pressure as

in actual usage a bi-fuel system delivers the fuel at ambient conditions. This

bounda

ulation strategy of the mixer in CFD was done in stages to determine the

required dimensions of the mixer. The dimensions are to meet the design criteria set

based on fabrication needs. The simulation strategy would follow Figure 3.6.

rate at outlet

When the two strok

m

culated from equation 3.1. This flow rate is the second boundary to the mixer.

The boundary was calculated and simulated for engine intervals of 1000 rpm.

3. CNG inlet

The third bo

ry allows the software to compute the amount of fuel sucked into the mixer

by means of the venturi shape and the air flow of the two stroke engine by the mixer.

Methane properties that are available in the software are similar to that found in

literature.

Sim

34

Start

Stage 1 Inlet and outlet angle analysis

Stage 2 Number of hole umferences at throat circ

Stage 3 Size of hole a rcumferencet throat ci

Stage 4 Throat size isationoptim

End

Figure 3.6 Overall simulation stages done on the mixer

.2.3 Inlet and Outlet Angles of the Mixer

Firstly, simulations done on Stage 1 are to obtain the inlet and outlet angles

e needs to be at maximum when the

ngine speed is at 8000 rpm using equation 3.1. This airflow rate was applied at the

mixer o

ith a 7°

diffuser angle to reduce flow separation. A 5 mm gap was at least needed to enable

pipe fit

3

for the mixer. For the analysis, the airflow rat

e

utlet while ambient pressure opening was given to the inlet. The air in the

mixer is travelling at the highest velocity according to the top engine speed.

During the simulation for the angle analysis the throat size was kept at the

initial size obtained from equation 3.2. The throat was given a design w

tings to the mixer at both ends. After giving the minimum gap at the ends, the

angles were computed using simple trigonometry. The throat length was

predetermined to be 20 mm from comparison of existing four stroke mixers

available. The angle was then varied to make the mixer fit in 60 mm. Figure 3.7

show where the inlet and outlet angles was placed in the mixer for CFD analysis.

35

Figure 3.7 Simulation model for inlet and outlet angles

The pressure across the mixer was found for the whole length of the mixer.

The pressure p cs of the air

ow in the mixer. A graph would show the effect of air flow pressure at each angle

inlet and outlet are not same in the

ixer. The most effective angle is the angle that produces the highest suction per

pressur

he hole drilled at this location would allow the mixer to operate

most efficiently as the suction of fuel is the highest.

result was subtracted from the

inlet pr ssure, P1 to get the highest suction reading. The overall pressure drop also

Inlet angle Outlet angle7°Diffuser

Gap5 mm

60 mm

P1 P2P3

lot was needed to see the overall pressure characteristi

fl

and to lead to the choosing of the best angle.

The criteria for choosing the right angle was by looking at a pressure ratio.

The reason for the criteria is that pressure at the

m

e drop. The ratio was found by finding the pressure difference from the inlet

to the highest suction and dividing it with overall pressure drop in the system. The

highest suction occurs at the location of the fastest air movement in the mixer. This

location happens at the throat because it restricts the air. As the mixer needs holes at

the throat based on the design criteria, the location of the highest suction was seen on

the throat wall, P2.

A pressure plot along the wall was needed to show exactly the point of

maximum suction. T

Knowing this location, a surface integral by the software was then used to

find the averaged pressure at this cross section. The

e

36

can be found by subtracting the pressure of the inlet, P1 from the outlet, P3. The ratio

was obtained from the equation:

Pressure ratio =31

21

PP

PP (3.8)

e efficiency of each angle that could

produce the higher amount of suction per pressure drop needs to be obtained from

FD analysis. The highest efficiency was chosen to be the angles used for the

sign.

of Holes at Throat Circumference

Obtaining the angle and location of the hole leads the design process to Stage

i at circumference of the mixer

esign. A CNG inlet was added to the design and the boundary of fuel is applied to

this inl

t speed, the mixer generates the lowest suction. The

low suction must be sufficient at the fuel inlet to enable the fuel to flow due to

suction

two holes at once. Then, the hole size were

determ itting of the number of holes at the throat. With even numbers,

the num

A pressure ratio graph that shows th

C

de

3.2.4 Number

2. Wh ch is to find the number of holes at the thro

d

et as in Figure 3.3. The size of the inlet needs to match the hose that connects

to the available conversion kits.

For this stage, the boundary of air should be for lowest engine speed of 1000

rpm. This is because at the lowes

generated in the throat. The profile of fuel mixing with air must be qualitative

and needs visualisation using CFD.

Firstly, the numbers of holes were taken as even numbers to ease machining

as the drill can be used to make

ined by the f

ber of holes that covers the most area at the throat needs to be chosen. This

was to not waste too much material between the angles at the circumference. After

this consideration, the number of holes was increased and simulations are done for

each number. The number of hole that gives a good mixing of fuel qualitatively at

the throat is finalised for the number of hole. A cross-sectional plot of the mixer at

37

throat was used to visualise the mixing. The colour contour plot from CFD

simulation decides which number hole gives better mixing.

3.2.5 ize of Hole at Throat Circumference

tage 3 of the process in designing the mixer was to enhance the size of hole

the number of holes was determined

Stage 2, the size of the holes are based on commonly available drill sizes. The size

uel flow at the fuel inlet should be in the range of 9.77 at the

chest limit and 17.23 at the leanest limit. A lower than 9.77 AF would cause the

mixer t

oat Size Optimisation

The last simulations in Stage 4 were conducted to find the correct throat size

determined. A variation of the throat size was

one by giving a small increment to slowly increase the throat size and looking at

S

S

to get good fuel and air mixing at the throat. As

in

of the hole would affect the AF ratio of the model. AF ratio given by the different

hole sizes was seen for the entire operation of the mixer from 1000 rpm to 8000 rpm.

To choose a suitable hole size the AF ratio is compared between the hole sizes

simulated at the throat.

The AF ratio obtained from dividing the surface integral of the mass airflow

at the air inlet and mass f

ri

o have too much fuel suction and this would cause the engine to have less

oxygen to complete the combustion. Meanwhile, a too lean AF ratio above 17.23

means less fuel sucked by the mixer and the hole size restricts the fuel suction. This

could cause incapability of the engine to combust the mixture. The AF ratio obtained

must fall in this range of rich at 9.77 and lean limit of 17.23 to choose the hole size to

be drilled.

3.2.6 Thr

after all the other parameters were

d

CFD results that were obtained by the enlargement process. Two parameters are

obtained by this enlargement to find the best throat size. The first being the AF ratio

and the second was the pressure drop.

38

The choosing criteria for the throat size was the first parameter, this is the AF

ratio. The throat size was increased from the initial throat size to the appropriate

throat size that gives AF ratio near to the stoichiometric limit of 17.23. The limit

nables

at gives the restriction to the air flow a pressure drop result is

expected in this simulation. The pressure drop value can be used for the validation of

the sim

.3 Prototyping the Mixer

After the CFD predictions were made, the mixer was built as a prototype. The

a xer drawings from the CAD drawings was sent for

brication of the prototype.

or better known as Perspex. With this material, a

transparent mixer was fabricated. The prototype needs to be created transparent to

visualis

use it is easy to machine, light weight and easily available

in the market. In addition, Aluminium is also corrosion resistant. These are reasons

why Al

e the mixer to mix air and fuel to the correct combustion value. A venturi is

predicted to give range of AF ratio and is constant. Due to this, a range of AF ratio

near stoichiometric value at a certain range of engine speed needs to be considered.

The mixer throat that can give AF ratio range near stoichiometric condition is chosen

to be fabricated.

The second parameter obtained from throat enlargement was the pressure

drop. As the thro

ulation results. Two validations were done to prove that the mixer simulation

is accurate in real life.

3

inform tion of the optimised mi

fa

Two prototypes were aimed to be generated. The first was a mixer made from

polymethylcrylate (PMMA)

e the flow of gas.

The second mixer was targeted to be made entirely from Aluminium. This

material was chosen beca

uminium was chosen as the material for the second prototype.

39

Fabrication process for machining the prototype are turning and drilling as

described earlier in the conceptual design of the mixer. The process was applied for

both m er prototypes.

3.4 alidating the Mixer Design

he two prototypes were made to validate the simulation results. Tests were

o tions to get a firm confirmation of CFD and

ctual testing. The first test was using the Perspex prototype of the mixer. This

.4.1 Testing Apparatus

The first apparatus was a flow test rig that is readily available at the

ntre (ADC) lab in UTM. The flow rate of air was

easured by this test rig. A flow test rig was used to test the mixer with air flow

similar

ix

V

T

done n the prototypes in three varia

a

prototype was used to visualise the CNG mixing in the mixer as the prototype is

transparent. Smoke was generated and flowed into the CNG inlet to replicate the gas

movement. The second test was the AF ratio that the mixer provides during its

operation. The AF ratio shows the amount of air and fuel that the mixer induced due

to suction from the engine. The third test was the pressure test, where the overall

pressure drop was found. The results were compared with simulation results. A set of

equipment were targeted to do all three tests.

3

Automotive Development Ce

m

to the engine conditions. The schematic diagram of the test rig is shown in

Figure 3.8. A centrifugal blower was used to generate suction and blowing to the

mixer prototype. The electronically controlled blower can produce the equivalent air

flow of the two stroke engine. The unit of measurement in the flow test rig is in litres

per minute (LPM). Air flow given by the blower is measured by means of a

calibrated laminar flow element.

40

Laminar Flow ElementMixer

DigitalDisplayUnit

PVC Piping

Centrifugal BlowerInverter

Figure 3.8 Schematic diagram of flow test rig to measure air flow

As the flow element of the flow test rig was used for a long time, a

recalibration needs to be done. Calibration was done by measuring the air flow at the

flow outl he flow

rate, Q ound from the test rig was converted into velocity. A velocity meter was

used to

) Measurement of AF ratio

i)

he first test is to visualise the mixing of two fluid flows in the mixer. As

CNG is colourless, the fuel was replaced with smoke to see the mixing clearly. The

et area. The blower was then set to the predetermined flow rate. T

f

measure the velocity again at this outlet. Finally, the two velocities are

compared to get a correction of the air flow rate given by the test rig.

The calibrated flow test rig was used with other measuring instruments to do

all three tests. The following tests were carried out for the mixer prototype:

i) Flow visualisation

ii

iii) Pressure drop across the mixer

Flow visualisation

T

41

vis ation was done by cualis ombining the use of the test rig with a smoke generator to

visually analyse the mixing phenomenon. The generator was available at the

Thermodynamic Lab in UTM was chosen as it provides continuous smoke at

atmospheric pressure. Figure 3.9 shows the schematic location of the smoke

generator when it is assembled to the test rig.

LaminarFlowElement

Figure 3.9 Schematic of smoke generator connected to tes

The gaseous

ow meter was used to read the flow rate at the fuel inlet of the mixer. The meter

. The gaseous flow meter measures the flow rate in

LPM also. The gaseous flow meter is as shown in Appendix C. With this instrument

both th

across the mixer

he third test was the pressure drop test. As the mixer causes pressure drop to

the air flow, a pressure test was needed to be done by combining the test rig to a

used by connecting it directly to the hoses,

t rig

Smokegenerator

ConeAdapter

Black screen

Mixer

Adapter

Camera

ii) Measurement of AF ratio

second test was the measuring of the AF ratio in the mixer. A

fl

was also available at the ADC lab

e measurements of air flow rate and fuel flow rate were obtained to do the AF

ratio measurement.

iii) Pressure drop

T

digital manometer. A manometer was

42

giving reading of pressure difference. The manometer was able to measure the range

of the p

Figure 3.10 ent

he chosen apparatus was used to do expe set of testing

procedures. The procedures are followed as a guide to conduct the tests.

he goals of the experiment were to validate the CFD simulation. Each

experim nt was done based on the air input calculated for each engine speed. The

t rig. Air speed was taken when the system comes to a

teady condition. The minor flow fluctuation also needs to be minimised. The errors

were re

ressure drop given by the mixer. A schematic of the manometer connected to

measure pressure drop is shown in Figure 3.10.

Schematic diagram of pressure measurem

2

1 3

Mixer

Digital manometer

T riments based on a

3.4.2 Testing Procedure

T

e

values were set at the flow tes

s

duced by doing many tests and averaging all the results obtained.

43

3.4.2.1 Smoke Mixing in Mixer

In the first experiment of smoke mixing of mixer, the air flow was set in the

ow test rig and pictures of smoke were taken at different engine speeds. The smoke

hows the path taken by the CNG if it was used in the system.

he smoke was generated by a smoke generator connected to the test rig. A

blowin

smoke flow from the

surroundings.

the second experiment, the combination of laminar flow element and

gaseous flow meter readings gives flow rate of air induced by the blower and the

flow rate of CNG. In actual test, the fuel that needs to be used to find AF ratio was

rig. Due to some limitations, the CNG fuel could not be used

the test rig. The limitations are as follow:

water to the regulator. As the mixer

2)

fl

s

T

g condition was used in the flow test rig to eliminate smoke from entering the

blower. A flow pattern was seen to determine whether the design can induce mixing.

A picture was taken with a black screen to differentiate the

3.4.2.2 AF Ratio Test

In

CNG from a bi-fuel test

in

1) The bi-fuel conversion kit needs heated water to heat the decompressed

gas from the storage tank. As the gas is regulated from 200 MPa to

atmospheric pressure the regulator becomes extremely cold and reduces

the flow of gas due to condensation in the regulator. Normally the engine

cooling system provides heated

needs to be tested separately a heating unit needs to be designed. The

limitation in resources did not allow the design of a heating unit.

The operation of the bi-fuel kit also needs input of spark timing from the

engine. This timing enables the regulator to energise the solenoids to

open and flow the CNG fuel to the mixer. This was seen as a limitation

as a timing device needs to be designed to operate the bi-fuel system if

the mixer was tested separately.

44

3) Apart from this, the usage of CNG in open testing imposes some safety

issues. The gas could not be contained after the testing and the highly

combustible gas would ignite if strict safety procedures are not applied.

main task here is to measure thThe e AF ratio even though the fuel is not

availab

readily ava

properties s e

to change the air properties to methane properties. The flow similarity follows

Reynol

or the first step, since the flow meter is calibrated to methane, the reading

shows

r manual (Sierra, 1994).

From the manual a K, factor was used to correct the flow rate measured as Methane

to the a

le. As CNG or methane cannot be used, air is used to replace the fuel as it is

ilable. Nevertheless the composition of air does not share the same

uch as density and viscosity with methane. A similarity analysis was don

ds number similarity (Andreas, 2001).

Two steps are taken to analyse the fuel flow. The first was to convert the fuel

flow rate measured by the gaseous flow meter to actual air flow rate. The second step

was to do the similitude analysis to find the fuel flow rate.

F

the flow rate of methane instead of air. This produces errors to the readings if

the correction of the flow is not done. The actual air flow rate was found by

calculations procedures given by the gaseous flow mete

ctual flow rate of air from the meter.

air

Methane

K

K

Q

Q

2

1 (3.9)

Subscripts 1 and 2 denote the measured and actual flow rates respectively. MethaneK is

0.72 and airK is 1.00 from the manual. After the air flow rate is obtained the value

was converted to air velocity by looking at

The air velocity was used in step two to change it back to fuel velocity.

pro

the cross sectional area at the fuel inlet.

For the second step, similarity of the flow rate at the CNG inlet need to be

used. The totype tested with air is geometrically similar to the model tested with

methane. For making both prototypes similar, a dimensionless property must be

made similar. The dimensional property is Reynolds number given as,

45

vd(3.10)Re

Where, is the fluid density, d is the diameter of the pipe, v is the fluid

velocity and is the viscosity of the fluid.

Using flow sim

ethane and the model with air. This can be written as,

ilarity the Reynolds number is same for the model with

m

AirMethane

vdvd (3.11)

Thus, the corrected air velocity can be converted to methane velocity by,

airair

MethaneairMethane vv

Methane

(3.12)

he velocity of methane is then converted to mass flow rate of fuel to the mixer,T

inletCNGMethaneMethanef Avm (3.13)

After getting the mass flow rate of fuel the mass flow rate of air has to be

und in order to find the AF ratio. The suction of the blower combines the two air

ing

the air flow rate, the air flow rate from inlet, te

flow test rig, . The subtracted flow rate is, . The

e AF ratio. The results

a gr

experiment and simulations. The air induced by the blower at the inlet is obtained by

fo

flow from the CNG inlet and the air flow from ambient condition, Q . For findTotal

the CNG , was subtrac d with the 2Q

readings shown on by Total air

airflow need to be converted into air mass flow rate, am by including air density.

The results am are divided by fm to obtain th are

calculated for each engine speed and aph of AF ratio was compared between

subtracting the flow rate obtained by the gaseous flow meter.

Q Q

46

A repetition of 10 times was done. Again the average readings were is taken.

The AF ratio was calculated and graph of AF ratio versus engine speed was plotted.

A comp

The third experiment was done to measure pressure drop at different engine

peeds. The pressure drop was measured in Pascal (Pa), as shown in Figure 3.9

3. Points 1 and 3 are tapped at 20 mm away from the

mixer so as not to be too far away from the mixer inlet and outlet. The CNG inlet,

point 2

f 1000 rpm. A plot of pressure drop

versus engine speed was produced.

ined and discussed.

arison was done on the two graphs.

3.4.2.3 Pressure Drop Test

s

previously for points 1 and

, was left open to atmospheric condition as the given input in the simulation

model is methane at atmospheric conditions.

Since there is fluctuation in the reading, the experiment was done 10 times.

The average of the 10 readings was taken for each pressure reading obtained.

Pressure readings were taken for intervals o

Validation was done on both the simulation of pressure drop and AF ratio.

Graphs of CFD analysis were compared with the corresponding experimental graphs.

The error of the two graphs was obta

CHAPTER 4

RESULTS AND DISCUSSIONS

The prototype of the mixer was based on the methodology planned in Chapter

3. Each step of the method was followed and results at each step were used to

continue on the other step. The steps had led the research to fabricate the prototype

of the mixer.

4.1 Designing of the Mixer

The process of design continues with the final concept of the mixer as seen in

Chapter 3. Calculation and CFD simulation was done to do the design for reduction

of cost in producing the mixer.

4.1.1 Initial Throat Size

Mixer throat size was determined by the maximum air flow characteristics of

the two stroke engine. The purpose was to find the correct inner diameter of the

throat which would be used to design the mixer prior to CFD analysis. The size of

the throat from maximum engine speed of 8000 RPM was given by,

s

mV

NQ Da

3

03.060

5.1

48

With throat air speed at compressible limit, was assumed as 150 m/s. The

throat area was given by:

2v

2A =2v

Q= 2.0 x 10-4 m2,

The diameter was found as: mA

d tt 016.0

4

The throat diameter was found to be 16 mm. The size is a guide to further

develop the design to meet the design criteria set earlier in Chapter 3.

4.1.2 CFD Simulation of the Mixer

A number of simulation were done to find the design dimensions. The results

were obtained for the following dimensions:

1) Inlet and outlet angle of the mixer

2) Number of holes at throat circumference

3) Size of holes at throat circumference

4) Throat size optimisation

4.1.3 Inlet and Outlet Angles of the Mixer

The results of CFD were plotted in Figure 4.1 as a pressure graph. The graph

showed the effect of the changes for each angle that was analysed. The range of the

angles for the inlet and outlet started by including a predetermined 5 mm gap to fit

the hoses at the inlet and outlet of the mixer. This was shown in Figure 3.7 earlier.

With this the simulations were done with the angles were varied from 60°, 70°, 80°

and 90°.At angles less than 60° the configuration does not meet the 5 mm gap

decided from design criteria and these angles would not be chosen to analysed. At

more than 90° the angle produced large opening from the inlet to the throat which

was seen as a waste of material to fit the 60 mm length.

49

84000

86000

88000

90000

92000

94000

96000

98000

100000

102000

0 0.01 0.02 0.03 0.04 0.05 0.06

Mixer Length (m)

Pres

sure

(Pa

)60 deg70 deg80 deg90 deg

Outlet

Inlet

2

1 3

Inlet and outlet angles

P3

P2Diffuser

P1

Figure 4.1 Pressure plot along the centre line of the mixer at different inlet and outlet

angles; 1 indicates inlet, 2 indicates throat diffuser and 3 indicates outlet

The graphs showed clearly the effect of the constriction at the throat on the

pressure profile. The graph also proves Bernoulli effect that the contraction at the

throat size cause higher fluid flow velocity at the throat which causes pressure to be

lowered. The converged air stream lines exert pressure to the wall. This causes more

pressure to be built up at the inlet of the mixer. When the air passes the throat, the

high pressure difference forces the air velocity to increase and reduces the pressure at

the throat. The pressure curve was found similar to the theoretical curve found in

literature as shown in Figure 2.6.

The stream lines will eventually diverge by the angle given at the throat

diffuser and outlet. The diverged flow looses its velocity and eventually increases the

pressure at the outlet to almost back to initial pressure at the inlet. The pressure does

not return back to its original condition because energy was used to overcome the

restriction. Judging by Figure 4.1, the lowest pressure drop occurred at 90° angle, but

the amount of pressure drop from the intake also was considered high comparatively

to the other angles. A choosing criteria needs to be followed at this stage.

50

The criteria to be met was that the angle must give high suction but with

lower overall pressure drop by looking at the pressure ratio. To find the pressure ratio

the point of highest suction must be found first. An accurate value of this lowest

suction was found by looking at the pressure plot along the throat wall in Figure 4.2.

83000

84000

85000

86000

87000

88000

89000

90000

91000

92000

0 0.005 0.01 0.015 0.02

Diffuser Length (m)

Pres

sure

(Pa

)

60 deg70 deg80 deg90 deg

MIN

Inlet and outlet angles

Figure 4.2 Lowest pressure at the throat diffuser wall.

From Figure 4.2 it can be seen that the lowest pressure occurs at the same

location for each angle. The location was at a point 4 mm from the diffuser. The

location also measured 24 mm from the mixer inlet if seen from Figure 4.1. This

point was taken as the point to make the holes along the circumference of the throat.

The location of the holes was then used to find the integral at the cross-

section. A pressure ratio was found by including the average surface integral at the

inlet throat and outlet of the mixer. The pressure ratio was obtained from the

equation 3.10. The angle that gives best pressure ratio was then found from a graph

plotted as shown in Figure 4.3.

51

1

1.1

1.

1.

1.

1.

1.

Pres

sure

Rat

io, (

P

2

3

4

5

6

1.7

1.8

50 60 70 80 90

1- P

2)/(

P2-

P3)

Inlet and Outlet Angles, º

Figure 4.3 Pressure ratios of each model inlet and outlet angle changes

From Figure 4.3, it was obvious that the 60° gave more relative suction

pressure to the overall pressure drop. The high suction was seen important as the

mixer was used to suck in fuel. On the other hand, the prediction of low pressure

drop ensures less energy was used by the mixer to overcome the restriction given.

The saving of energy here translates into a more efficient engine operation. The ratio

also depicts the efficiency of the angles. The efficiency reduces as the angle

increased due to more flow separation that occurs during the expansion at the outlet.

This shows that sharp angles are bad for smooth airflow as more turbulence were

produced.

As a result the inlet and outlet angle of the main body of the mixer was

optimised at 60 . The simulation also showed that the lowest suction was

approximately at 24 mm from the inlet of the mixer for all angles simulated. The

reason being the flow entering the throat restriction produces a vena contracta. This

causes the highest velocity in the mixer happens at a slight distance from the throat

and not at the throat. This is as shown in Appendix B. The high velocity in turn

Two views of mixing air and fuel were seen. The views are a 3 dimensional

view for overall comparison of the mixing and a cross-sectional side view at the

throat holes. The results are shown in Figure 4.4, 4.5 and 4.6 for the three numbers of

holes simulated. The quality of the flow from the simulation plot shows how well the

mixing occurs.

As the largest number was 12, a 3 mm hole size was used to start the

simulation. A larger than 3 mm size does not fit the mixer throat with 12 holes. It

causes jagged edges at the throat due to the holes being too close to each other.

Simulations were done to find the number of holes at the throat

circumference. Since the number of holes was made by considering even number of

holes, 8 holes is the most likely number to be simulated first. When increasing the

number of holes it was found that they caused decimals when dividing with 360°.

For example, if 8 hole numbers was chosen it can be easily divided with the 360° in

the throat giving 45° per hole. If the number was 14 holes the division is 25.714° per

hole. It is difficult to fabricate the design if there are decimals in the design.

Simplification without decimal was done. Twelve holes were found to be the

maximum increment without causing decimals in the angle calculated. Due to this

the numbers of holes were simulated for three conditions of 8, 10 and 12 holes.

4.1.4 Number of Holes at Throat Circumference

reduces the local pressure in the area. The location of the low pressure was used to

drill the location of the holes along the circumference of the throat.

52

1

82

Thr

oat

= 1

6 m

m

Hol

e =

3 m

m

73

64

5F

igur

e 4.

4 E

ight

hol

es m

ixer

mod

el

53

12

10

39

Thr

oat

= 1

6 m

m

Hol

e =

3 m

m

84

57

6

Fig

ure

4.5

Ten

hol

es m

ixer

mod

el

54

55

4

3 5

2 6

1

78

12

11 9

10

Thr

oat

= 1

6 m

m

Hol

e =

3 m

m

Fig

ure

4.6

Tw

elve

hol

es m

ixer

mod

el

With the number of holes determined, the CFD analysis continued to find the

better hole size based on available drill bits in the market. The largest size of the drill

was considered by the size govern by the 12 holes chosen earlier in the design stage,

while the smallest is the smallest easily available drill bit in the market. The drill bits

were identified as 1mm, 1.5 mm, 2.0 mm, 2.5 mm and 3.0 mm. The results are

shown in Figure 4.7. The results to the 1 mm hole size AF ratio was not plotted in the

graph as the values are too lean.

4.1.5 Size of Hole at Throat Circumference

From the results, twelve hole gives most coverage of fuel at the throat. This is

true from the colours obtained from the simulation. The blue colours around the

edges are less comparing to the 8 and 10 holes. The blue colour is the location that

the fuel does not mix with the air. The occurrence of this location shows the less

coverage of the fuel at the throat cross section. With this 12 holes will be done to on

the mixer throat to enable good mixing at the throat.

All the models were capable of mixing the fuel around the mixer throat. The

induction of fuel was due to the lower pressure region at the throat. As CNG moves

from the higher pressure at the CNG inlet to lower pressure at the throat, the fuel was

virtually sucked into the mixer. It was found that the model predicted for the location

of the hole was correct and does provide adequate suction to the mixer. This was

because the mixer simulation shows fuel entering the mixer without the pressure

from the fuel inlet. It is suction that pulls the fuel in. The fuel was then seen mixing

with the air which is shown by the gradual change in colour of the simulation from

red (CNG fuel) and blue (pure air) to colour of light blue (mixed CNG and air).

56

57

510152025

010

0020

0030

0040

0050

0060

0070

0080

0090

00

3.0

mm

AF=

9.7

7

2.5

mm

2 m

m

1.5

mm

2.5

mm

2.0

mm

1.5

mm

AF=

17.

23

3 m

mD

rill

Size

Fig

ure

4.7

Eff

ect o

f ho

le s

izes

on

AF

ratio

at t

hroa

t cir

cum

fere

nce

at a

ll sp

eed

rang

e

Eng

ine

spee

d (r

pm)

AF Ratio

58

The overall trend of results show that the AF ratio becomes richer as the

speed was increased. The AF ratio also becomes rich when the area of the hole was

increased using the larger drill size. This shows that more suction of fuel occurs with

higher engine speeds and larger flow area. The limits for choosing the better hole are

drawn in the graph at AF ratio of 9.77 and 17.23.

From the graph, the area covered by hole of 1.5 mm produces a lean mixture

with most of the points above the stoichiometric limit. The reason would be the small

opening restricts the flow of fuel into the air stream. The reverse was seen happened

for the 2.5 mm and 3 mm hole as the opening area is larger. The amount of fuel was

too much and shows the overall rich condition.

From the result, the hole size of 2 mm was the best available option as the

range was found in between the very rich and stoichiometric limit. With the hole size

determined, the last stage simulation was done which is the throat size optimisation.

4.1.6 Throat Size Optimisation

Since the initial diameter was 16 mm the throat size was increased to find the

size which is in the stoichiometric range. The throat was increased slightly and each

increment was simulated to do the optimaisation. An increment of 1 mm was seen

the best for this small increment. This was because it was small enough to be

machined easily but enough produce the required resolution of results until the

stoichiometric limit.

Considering the increment, the throat size range was found to be 16 mm, 17

mm 18 mm, 19 mm, 20 mm and 21 mm. At size larger than 21 mm, the simulated

result was found too lean as it was too much above the stoichiometric limit. Due to

this the largest throat size to be simulated was set at 21 mm.

The results of simulation and actual running have errors themselves as

simulation only takes in ideal condition to calculate. As it was known that there will

be difference in simulation and experiment, the 19 mm throat diameter is the best

model to predict at this stage as the model for validation is designed next. Should

there an error occurred as predicted, the experiment is expected to put the graph with

more points at the stoichiometric range. If the 20 mm diameter was chosen it will put

the results in more lean limit and not stoichiometric.

It was shown that the 20 mm throat has the most points in the AF ratio range

specified. All the other mixer throat diameters are considered as too lean or rich for

the mixer design. Nevertheless, CFD can only predict the results and the actual

results was expected to be leaner than the simulated results.

The usage of different throat diameter showed variation in AF as seen in

Figure 4.8. The size of the mixer throat has to provide a good range of AF ratio,

which has to be near stoichimetric condition. In the graph a limit of AF of 5 % of

stoichiometric condition was drawn. For all the results a similar trend was shown.

The shape of the graph was exponential and the graph gradually straightened out.

This was due to the venturi shape that gives AF ratio at certain range. A venturi is

known not to give constant AF ratio in engines.

The throat optimisation shows that the larger sized mixer compared to the

initial throat size is a better working model at stoichiometric. The larger throat size

will give a small change in pressure drop.

Two results were found for the simulation done on the throat sizes. The

results were obtained as planned in Chapter three from the use of the CFD software

package. The first is the AF ratio and the second is the pressure drop of the mixer.

59

60

10121416182022242628

010

0020

0030

0040

00

Fig

ure

4.8

Eff

ect o

f th

roat

dia

met

er s

ize

on a

ir f

uel r

atio

5000

6000

7000

8000

9000

AF=

17.2

3A

F= +

5%

AF=

-5%

21

mm

20

mm

19

mm

18

mm

17

mm

16

mm

Eng

ine

spee

d (r

pm)

Thr

oat

Size

AF Ratio

Fuel suction was the most important aspect of the mixer and the main

operation of the mixer was based on the lower suction at the throat. Due to this, the

19 mm throat predicted is having a balance of low pressure drop to suction compared

to other throat diameters simulated. The pressure drop is around 4230 Pa. The small

pressure drop value can be neglected of giving effect to engine operation. The

predicted the last model of 19 mm will be made into the prototype.

A mixer with large pressure drop was not a good design as the possibility of

engine stalling was there due to loss of energy. This was seen in the initial model of

the mixer with 16 mm throat size. Whereas, a low pressure drop will cause reduction

of the suction pressure needed to suck in fuel. Low pressure drop was seen at large

diameters of 21 mm.

In the pressure drop results, the results are as shown in Figure 4.9. All the

graphs produced the same trend, which was the exponential increase in pressure drop

as the engine speed was increased. The pressure drop curve proves the theory from

Bernoulli, which says that as more airspeed was given to a restriction the pressure

drop would increase to the square of the velocity. Apart from this, the pressure drop

closely relates with the efficiency of the mixer. The smaller throat uses more energy

of the flow to overcome the restriction while the larger throat uses lesser energy. The

energy comes from the engine or in other words, the existence of pressure drop

reduces some efficiency from the engine. The mixer will always produce pressure

drop as the restriction is needed to create the suction to induce fuel to the throat.

61

62

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

1000

0

010

0020

0030

0040

0050

0060

0070

0080

0090

00

21

mm

20

mm

19

mm

18

mm

17

mm

16

mm

Thr

oat S

ize

Fig

ure

4.9

Sim

ulat

ion

pres

sure

dro

p du

e to

dif

fere

nt th

roat

siz

e at

all

engi

ne s

peed

Eng

ine

spee

d (r

pm)

Pressure (Pa)

63

4.2 Prototyping the Mixer

Most of the dimension of the mixer was obtained from simulation. These are

can be summarised as below:

1. Symmetrical inlet and outlet angle of 60°

2. The location of the fuel hole at 24 mm from the inlet

3. There will be 12 holes at the throat circumference

4. A 2 mm drill size was chosen to be used to make the holes at the

throat

5. The throat size was finalised at 19 mm from CFD simulations.

After CFD simulation, the optimised model was made into two prototypes.

The prototypes were made using two different materials which are:

1) Perspex prototype

2) Aluminium prototype

For the first prototype, material was chosen to make a transparent mixer.

Figure 4.10 shows the mixer which was fabricated using Perspex. The Perspex

prototype was used to validate the mixing of gases in the mixer.

Figure 4.10 Perspex model for flow testing

64

Using the same dimensions, an aluminium model was made. The properties

for the chosen material are shown in Appendix E. The fabricated mixer was shown in

Figure 4.11 in assembled view and the detailed view of all the components are shown

in Figure 4.12.

Figure 4.11 Assembled view of Aluminium mixer

Throat

Mixeroutlet

CNG inlet

Mixerinlet

Figure 4.12 Components of Aluminium mixer

65

4.3 Validating the Mixer Design

The main concern of the experiments was to get the pressure drop and air fuel

ratio values and compared these values with results of CFD. Three sets of

experiments were conducted with the prototype of the device:-

i) Smoke mixing in Perspex prototype

ii) Air fuel ratio testing of mixer

iii) Pressure drop testing of mixer

4.3.1 Smoke Mixing in Perspex Prototype

This was the first test done on the mixer. The transparent mixer shows the

accumulation of smoke in the mixer throat due to suction at the CNG inlet. Only

three engine speeds were used. They were 1000 rpm, 2000 rpm and 3000 rpm. The

reason for this was that at higher speeds, the smoke could not be seen as it was

diluted in the air stream.

Figure in Appendix C and Figure 4.13 and shows the pictures that were taken

of the smoke and CFD simulation results. From here the difference between the

simulated condition and the experimental results at the applied engine air speed was

analysed.

66

1000 rpm

2000 rpm

3000 rpm

Figure 4.13 Simulation of smoke at 1000 rpm, 2000 rpm and 3000 rpm air speed

Most of the results obtained were in the range near the stoichiometric value.

This was as predicted from the simulation. Some results were scattered in the plot.

From the experimental results, a best line was drawn onto the graph. This was based

on an average of the points obtained from the experiment. This line was compared to

the simulation results.

After the relevant similitude analysis was done to the results obtained from

the gaseous flow meter, the results were compared to simulation results. Overall, the

simulation result gives a richer value compared to experiments for the AF ratio tests.

This was shown in Figure 4.14. From the results, a maximum of 4.79% AF ratio

error was calculated using the mixer. This was obtained by averaging all the errors

obtained from comparing the points of simulation and experiments.

4.3.2 AF Ratio Testing of Mixer

Based on the colour contour, simulation results show a flow pattern as air

flow speed increases. Mixing of fuel was more at high speeds in the mixer outlet; the

light blue colour shows this mixing. In the experiment, the white colour intensity of

the smoke leaving the mixer (mixed flow) was lesser than in the CNG inlet pipes.

This shows that the air and smoke mixing. Some fuel was divided into two by the

rapid air movement causing the flow to be pushed near to the wall. This was true as

the experiment had smoke taking shape from the mixer walls when leaving the

mixer. The simulation validation proves that the software can be used to analyse the

flow field qualitatively.

67

68

141516171819202122

010

0020

0030

0040

0050

0060

0070

0080

0090

00

AF=

-5%

AF=

17.

23

AF=

+5%

E

xper

imen

t

Sim

ulat

ion

Fig

ure

4.14

Exp

erim

ent a

nd s

imul

atio

n re

sults

of

AF

ratio

Eng

ine

Spee

d (r

pm)

AF ratio

69

The difference of simulation to experiment occurred due to ideal conditions

assumed by the simulation. In real condition, some of these ideal conditions would

not occur. One of these assumptions was that the atmospheric pressure at the CNG

inlet hose was not exactly at the atmospheric condition. The reason being the hose

length connecting the gaseous flow meter has given some restriction in the fuel flow.

From the results of AF ratio in Figure 4.14, it shows that the mixer was

capable of giving an overall stoichiometric AF ratio. The 19 mm mixer was slightly

leaner at low speeds. This condition was not seen very critical for the mixer as in real

life the bi-fuel kit was used to compensate the idle condition until 2000 rpm by

adjustments to the pressure regulator (Landirenzo, 2003).

In real operation, the mixer was estimated to operate within the city driving

cycle at conventional engine speeds, which rarely exceeds 6000 rpm. From the graph

in Figure 4.14, at speeds exceeding 6000 rpm, the mixer is providing a rich mixture

for the engine to combust. The operation speed of 6000 rpm can be neglected as the

engine rarely achieves this high speed in normal driving.

4.3.3 Pressure Drop Testing of Mixer

The test results were obtained with the flow test rig using the Aluminium

prototype and the digital manometer. The tests were done from the same flow rates

that were used in the simulation. The overall pressure drop was obtained by taking

the results shown in the digital manometer.

In the experiment, the result does follow the trend of increasing pressure drop

as seen in the simulations. The overall simulation results were giving higher pressure

drop than experimental data as shown in Figure 4.15. A best fit line was again drawn

As seen by the results the points plotted for the experimental reading are

fluctuating more at higher speeds. The reason was that the higher rotation speed of

the blower uses more electric power. The increased electric power usage also

amplifies the small errors producing more fluctuations to the results.

The reason for the difference was that the simulation ideal assumption did not

actually occur at real condition. During the experiment, the blower causes the

pressure drop reading to fluctuate slightly. The fluctuation was seen when the

pressure reading were taken using the manometer. The error was caused by small

fluctuation from the electric power supply, which was unavoidable in a high

electricity usage facility such as the ADC lab.

to see the errors of simulation and experiments. A maximum 4.96% error in average

of all points was found while comparing simulation and experiment results.

70

71

0

1000

2000

3000

4000

5000

6000

010

0020

0030

0040

0050

0060

0070

0080

0090

00

E

xper

imen

t

Sim

ulat

ion

Fig

ure

4.15

Sim

ulat

ions

and

exp

erim

ent p

ress

ure

drop

Eng

ine

spee

d (r

pm)

Pressure (Pa)

CHAPTER 5

CONCLUSION AND RECOMMENDATION

5.1 Conclusion

In the process of designing the CNG mixer, it can be deduced that the design

has reached the objectives set. A venturi burner type of CNG mixer was designed

and fabricated for a 150cc two stroke engine application. A mixer capable of

providing a stoichiometric AF ratio condition and low pressure drop was created.

CFD was a good tool to analyse the mixer and is a form of cost saving. From

the simulation, dimensions of the design were finalised. A conceptual design was

obtained from CFD analysis and developed into the prototype of the final product.

The two validations of pressure drop and AF ratio showed the accuracy of the

simulation results in real life application. Validation of the mixer by comparing the

CFD result and experimental results showed some errors. The errors were minimised

by doing a repetition of the experiments. The overall pressure drop and AF ratio had

recorded errors of 4.96% and 4.79% for the validation of air fuel ratio and the

validation of pressure drop. A less than 5% error was considered as a good prediction

that the simulations has achieved. The experiments proved that the flow did follow

similar trend as in the simulations. With this design process via simulation, the mixer

was found to have a general specification as shown in Table 5.1.

73

Table 5.1 Specification of the mixer designed

Length 60 mm

Inlet and outlet angles 60°

Location of hole 24 mm from inlet

Number of hole 12

Size of hole 2 mm

Size of throat 19 mm

Maximum pressure drop 4400 Pa

AF ratio range 5% 17.23 (Stoichiometric)

5.2 Recommendation

A better design of the mixer would follow a venturi shape. This was because

the shape does provide lesser losses in term of pressure drop. The assumption of

symmetrical inlet and outlet angles was done to minimise the length to fit the mixer

to the engine. A venturi specification can be followed if there are more spaces to fit

the mixer in the two stroke engine.

To get a more accurate test, a direct CNG test can be done by taking the

extreme safety precautions when using the fuel. A design of spark timing and heating

unit can be used to give timing reading and heated water to the bi-fuel conversion kit.

Actual test in engine would be the ultimate test that should be done in order

to validate the mixer further. This validation should be done by proper mounting and

sealing of the mixer to the engine. The air flow and fuel flow to the engine must be

measured directly. Other parameters such as fuel consumption, power and efficiency

compared to petrol usage can also be focused in actual engine test.

REFERENCES

Andreas N. Alexandrou (2001). Principles of Fluid Mechanics. Prentice Hall. New

Jersey.

Baert R. S. G., Beckman D. E., Veen A. (1999). Efficient EGR technology for future

HD diesel engine emission targets. TNO Road Vehicles Research Institute.

SAE 1999-01-0837.

Bryan Willson. (2002). Direct Injection as a Retrofit Strategy for Reducing

Emissions from 2-Stroke Cycle Engines in Asia. Hong Kong.

Ferguson, C.R (2001). Internal Combustion Engines- Applied Thermo-sciences. John

Wiley & Sons. Canada.

Gan L.M., (2003). Design and Development of Two Stroke Engine Using Blower

Mechanism. UTM, Thesis.

Gas Malaysia Sdn. Bhd. (2003). Natural Gas in Malaysia. Gas Malaysia

Heywood J.B (1988). Internal Combustion Engines Fundamentals, Mc Graw Hill

International Edition. Automotive Technologies Series

Jitendra (Jitu) Shah, N.Harshadeep (2001), Urban Pollution from Two Stroke Engine

Vehicles in Asia, Regional Workshop on Reduction of Emissions from 2-3

Wheelers, September 5-7, 2001– Hanoi, Vietnam.

Landirenzo, (2003). TN-SIC CNG Regulators. Installation Manual. Landirenzo

S.p.A. Italy

Lenz, H.P, (1992). Mixture Formation in Spark-Ignition Engines. SAE Inc. New

York.

Luiz Henrique Borges, Carlos Hollnagel and Wilson Muraro. (1996). Development of

Mercedes-Benz Natural Gas Engine M 366 LAG with a Lean Burn System.

SAE Brasil 1996. 962378 E

Maxwell T.T. and Jones J.C. (1995). Alternative Fuels: Emissions, Economics and

Performance. USA Society of Automotive Engineers: SAE Inc.

75

Mardani Ali Sera, Rosli Abu Bakar, Sin Kwan Leong. (2003). CNG Engine

Performance Improvement Strategy through Advanced Intake System.

Universiti Teknologi Malaysia. JSAE 20030229. SAE 2001-01-1937. Japan.

Mikio Furuyama, Bo Yan Xu. (1998). Mixing Flow Phenomena of Natural Gas and

Air in the Mixer of a CNG Vehicle. SAE 981391. Chiba University. Japan.

Mohamed Maurie Bundu. (1998). Investigation of the Performance of A Spark

Ignition Engine with Gaseous Fuels. Dalhouse University. Canada

Poulton M.L. (1994). Alternative Fuels for Road Vehicles. Computational Mechanics

Publications. Southamton. UK and Boston. USA. Pg 99-121.

Rosli Abu Bakar, Azhar Abdul Aziz and Mardani Ali Sera. (2002a). Effect of Air

Fuel Mixer Design on Engine Performance and Exhaust Emission Of A

CNG Fuelled Vehicles, 2nd World Engineering Congress Sarawak,

Malaysia,22-25 July 2002

Rosli Abu Bakar, Mardani Ali Sera, Sin Kwan Leong. (2002b). Design and

Development of New Compressed Natural Gas (CNG) Engine. IRPA Vot

72351. UTM.

Rosli Abu Bakar, Devarajan Ramasamy, Gan Leong Ming. (2004). Design of

Compressed Natural Gas (CNG) Mixer Using Computational Fluid

Dynamics. 2nd BSME-ASME International Conference on Thermal

Engineering. 2-4 January 2004. Dhaka

Rosli Abu Bakar, Devarajan Ramasamy, Chiew Chen Wee, (2003). Effects of Port

Sizes in Scavenging Process on New Two-Stroke Engine Using Numerical

Analysis. 3rd International Conference on Numerical Analysis in Engineering,

Batam View Beach Resort, 13-15 March.

Sierra Instruments, (1994). Top-Trak Mass Flow Meters. Instruction manual.

California. USA.

Taib Iskandar Mohamad, Mark Jermy, Matthew Harrison, (2003).Direct Injection of

Compressed Natural Gas in Spark Ignition Engines. ICAST 2003.

Willard W. Pulkrabek, (1997). Engineering Fundamentals of the Internal

Combustion Engine. Prentice Hall.

Yeap Beng Hi, Azeman Mustafa, Zulkefli Yaacob. (2002). Computational

Investigation of Air-Fuel Mixing System for Natural Gas Powered

Motorcycle. 6th Asia-Pacific International Symposium on Combustion and

Energy Utilization 20 – 22 May 2002, Kuala Lumpur

76

ISBN 983-52-0244-3

Yusoff Ali and Zailani Muhammad (2003). The Issues Promotion of the Use of

Natural Gas in Automotive the New Trend. ICAST 2003.

Appendix A

Thesis Gantt Chart

MO

NT

HT

ASK

DIV

ISIO

N

1- J

2- F

3- M

4- A

5- M

6- J

7- J

8- A

9- S

10-

O

11-

N

12-

D

13- J

14-

F

15-

M

16-

A

17-

M

18- J

19- J

20-

A

21-

S

22-

O

23-

N

24-

D

1L

ITE

RA

TU

RE

RE

VIE

W:

Tw

ost

roke

eng

ine,

CN

G a

s fu

el, b

i fue

lkits

,m

ixer

oper

atio

ns, a

nd p

ress

ure

in v

entu

ri.

2M

IXE

R D

ESI

GN

:A

ir r

equi

rem

ents

of

two

stro

kes,

con

cept

mod

el, t

hroa

t siz

ing.

3C

FD

Ana

lysi

s: M

esh

stud

y, f

low

sim

ulat

ion,

air

fue

lrat

io a

naly

sis

and

pres

sure

drop

ana

lysi

s in

CFD

4F

AB

RIC

AT

ION

PR

OC

ESS

: Pr

otot

ype,

tech

nica

l dra

win

gs, m

achi

ning

, fi

nish

ing

and

asse

mbl

y.

5T

EST

RIG

: T

echn

ical

dra

win

gs,

oper

atio

n of

blo

wer

,ada

pter

des

ign

for

mix

er f

ittin

g

6T

EST

ING

:A

ssem

bly

of c

ompo

nent

s,A

Fra

tiote

stin

g, p

ress

ure

drop

test

ing

7R

EP

OR

T W

RIT

ING

: C

ompi

ling

pape

rs, t

ypin

g, e

ditin

g an

d bi

ndin

g.

The

pro

ject

is c

arri

ed o

ut o

n a

two

year

(24

mon

th)

tim

e pe

riod

78

Appendix B

CFD Analysis

80

B1 Inlet and Outlet Angle Analysis

Figure B1 Angle 60 °

Figure B2 Angle 70 °

Figure B4 Angle 90 °

Figure B3 Angle 80 °

81

B2

Hol

e Si

ze a

t cir

cum

fere

nce

sim

ulat

ion

seen

at 8

000

rpm

whi

ch is

the

high

est s

uctio

n th

at w

ill b

e gi

ven

by th

e m

ixer

1mm

1.5

mm

Fig

ure

B5

1 m

m a

nd 1

.5 m

m h

ole

size

82

2 m

m

2.5

mm

Fig

ure

B6

2 m

m a

nd 2

.5 m

m h

ole

size

83

Fig

ure

B7

3 m

m h

ole

size

84

B3

Thr

oat

AF

rat

io s

imul

atio

n

1000

rpm

2000

rpm

Fig

ure

B8

16 m

m th

roat

siz

e at

100

0 rp

m a

nd 2

000

rpm

85

3000

rpm

4000

rpm

Fig

ure

B9

16 m

m th

roat

siz

e at

300

0 rp

m a

nd 4

000

rpm

86

5000

rpm

6000

rpm

Fig

ure

B10

16

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

87

7000

rpm

8000

rpm

Fig

ure

B11

16

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

88

1000

rpm

2000

rpm

Fig

ure

B12

17

mm

thro

at s

ize

at 1

000

rpm

and

200

0 rp

m

89

3000

rpm

4000

rpm

Fig

ure

B13

17

mm

thro

at s

ize

at 3

000

rpm

and

400

0 rp

m

90

5000

rpm

6000

rpm

Fig

ure

B14

17

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

91

7000

rpm

8000

rpm

Fig

ure

B15

17

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

92

1000

rpm

2000

rpm

Fig

ure

B16

18

mm

thro

at s

ize

at 1

000

rpm

and

200

0 rp

m

93

3000

rpm

4000

rpm

Fig

ure

B17

18

mm

thro

at s

ize

at 3

000

rpm

and

400

0 rp

m

94

5000

rpm

6000

rpm

Fig

ure

B18

18

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

95

7000

rpm

8000

rpm

Fig

ure

B19

18

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

96

1000

rpm

2000

rpm

Fig

ure

B20

19

mm

thro

at s

ize

at 1

000

rpm

and

200

0 rp

m

97

3000

rpm

4000

rpm

Fig

ure

B21

19

mm

thro

at s

ize

at 3

000

rpm

and

400

0 rp

m

98

5000

rpm

6000

rpm

Fig

ure

B22

19

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

99

7000

rpm

8000

rpm

Fig

ure

B23

19

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

100

1000

rpm

2000

rpm

Fig

ure

B24

20

mm

thro

at s

ize

at 1

000

rpm

and

200

0 rp

m

101

3000

rpm

4000

rpm

Fig

ure

B25

20

mm

thro

at s

ize

at 3

000

rpm

and

400

0 rp

m

102

5000

rpm

6000

rpm

Fig

ure

B26

20

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

103

7000

rpm

8000

rpm

Fig

ure

B27

20

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

104

1000

rpm

2000

rpm

Fig

ure

B28

21

mm

thro

at s

ize

at 1

000

rpm

and

200

0 rp

m

105

3000

rpm

4000

rpm

Fig

ure

B29

21

mm

thro

at s

ize

at 3

000

rpm

and

400

0 rp

m

106

5000

rpm

6000

rpm

Fig

ure

B30

21

mm

thro

at s

ize

at 5

000

rpm

and

600

0 rp

m

107

108

8000

rpm

Fig

ure

B31

21

mm

thro

at s

ize

at 7

000

rpm

and

800

0 rp

m

7000

rpm

Appendix C

Apparatus and Experiments

110

Fig

ure

C1

Flow

Tes

t Rig

Ele

ctri

cal D

iagr

am

111

Figure C2 Laminar Flow Element

Figure C3 Digital Manometer (DP-Calc)

112

Figure C4 Sierra Top Trax Flow Meter

Figure C5 Pressure Difference At Inlet And Outlet Setup

113

Figure C6 Digital Manometer used to measure pressure difference, Sierra Flow

meter used to measure flow rate

Figure C7 Gas Connection

114

Figure C8 Flow test rig in suction condition

Figure C9 Flow test rig in blowing condition

115

G

21 3 414

B 5Y

No Part No Part1 3 Stage Pressure

Regulator8 Spark Timing Advance Processor

(STAP)2 Pressure Gauge 9 Battery 12VDC3 Gasoline Solenoid Valve 10 Ignition Key 4 Refueling Valve 11 Gasoline-CNG Fuel Switch 5 High Pressure Pipe 12 Mixer6 CNG Tank 13 Power Valve 7 Engine Ignition Coil 14 Low Pressure Pipe

Color Code R-Red wire

G-Green wire

B-Blue wire

Br-Brown wire

W-White wire

Y- Yellow wire

8

R Br

G

G

R12

13

11

10

9 7

6

-+

STAP

W G

Figure C10 Schematic diagram of bi-fuel system

116

Figure C11 Smoke flow at 1000 rpm

Figure C12 Smoke flow at 2000 rpm

Figure C13 Smoke flow at 3000 rpm

Appendix D

Technical Drawings

A A

38

44

34

61

67

0

10

70

80

A-A

(1

: 1)

DRA

WN

APP

V'D

UN

LESS

OTH

ERW

ISE

SPEC

IFIE

D:

DIM

ENSI

ON

S A

REIN

MIL

LIM

ETER

SFI

NIS

H:

DEB

UR A

ND

BREA

K SH

ARP

EDG

ES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

WEI

GHT

:

UNIV

ERSI

TI TE

KNO

LOG

I MA

LAYS

IA

TITLE

:D

WG

NO

.:SC

ALE

:1:2

SHEE

T 1 O

F1

A4

SURF

AC

E FI

NIS

H:

TOLE

RAN

CES

:(U

NLE

SS S

PEC

IFIE

D)

LIN

EAR:

AN

GUL

AR:

QTY

.:

DEV

ARA

JAN

A/L

RA

MA

SAM

Y05

/01/

2005

Flow

Tes

tRig

Ad

apt

er

FAKU

LTI K

EJUR

UTER

AA

NM

EKA

NIK

AL

PRO

F. M

AD

YA D

R. R

OSL

I

118

720

850

1123

.50

167

93245

3.50

273.

50275

280

212

117

A90

Ben

d P

VC

Pip

e

Pers

pe

x A

da

pte

rD

igita

l Dis

pla

y U

nit

Lam

ina

r Flo

w E

lem

en

t

Ce

ntr

ifug

al B

low

er

645

1397

450

300

220

DET

AIL

ASC

ALE

1 :

5

Mix

er

DO

NO

T SC

ALE

DR

AW

ING

Flo

w T

est

Rig

SHEE

T 1

OF

1

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:SC

ALE

: 1:1

WEI

GH

T:

REV

DW

G.

NO

.

ASIZ

E

TITL

E:N

AM

ED

ATE

Q.A

.

MFG

AP

PR

.

EN

G A

PP

R.

CH

EC

KE

D

DR

AW

N

FIN

ISH

MA

TER

IAL

INTE

RP

RET

GEO

MET

RIC

TOLE

RA

NC

ING

PER

:

DIM

EN

SIO

NS

AR

EIN

MIL

IME

TER

STO

LER

AN

CE

S:F

RA

CTI

ON

AL

AN

GU

LAR

: MA

CH

BE

ND

TW

O P

LAC

E D

EC

IMA

LT

HR

EE

PLA

CE

DE

CIM

AL

PRO

PRIE

TARY

AN

D C

ON

FID

ENTIA

L

THE

INFO

RM

ATI

ON

CO

NTA

INED

IN T

HIS

DR

AW

ING

IS T

HE

SOLE

PR

OP

ERTY

OF

UN

IVER

SITI

TEK

NO

LOG

IMA

LAY

SIA

(U

TM).

AN

YR

EPR

OD

UC

TIO

N IN

PA

RT

OR

AS

AW

HO

LE W

ITH

OU

T TH

E W

RIT

TEN

PER

MIS

SIO

NO

F U

TM IS

PR

OH

IBIT

ED.

DE

VA

RO

SLI

119

A A

CN

G In

let

19 m

mRi

ng

32

18.0

6

32

45.9

4

60

A-A

(1 :

1)

Mix

er O

utle

tM

ixer

Inle

t

DR

AW

N

AP

PV

'D

UN

LESS

OTH

ERW

ISE

SPEC

IFIE

D:

DIM

ENSI

ON

S A

RE

INM

ILLI

MET

ERS

FIN

ISH

:D

EBU

R A

ND

BREA

K S

HA

RP

ED

GES

NA

ME

SIG

NA

TUR

ED

ATE

MA

TER

IAL:

WEI

GH

T:

UN

IVER

SITI

TEK

NO

LOG

IMA

LAYS

IA

TITL

E:D

WG

NO

.:S

CA

LE:1

:2S

HEE

T 1

OF

1

A4

SUR

FAC

E FI

NIS

H:

TOLE

RA

NC

ES:

(UN

LESS

SP

ECIF

IED

) L

INEA

R:

AN

GU

LAR

:

QTY

.:

DEV

AR

AJA

N R

AM

ASA

MY

1/9

/200

4

Alu

min

ium

Mix

er

FAKU

LTI K

EJUR

UTER

AAN

MEK

ANIK

ALA

SSO

C P

ROF

RO

SLI A

BU

BAKA

R

120

A A

37

32

30°

4

2

29

36

024

9.80

18.06

A-A

(2

: 1)

DRA

WN

APP

V'D

UNLE

SS O

THER

WIS

ESP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLI

MET

ERS

FIN

ISH

:D

EBUR

AN

DBR

EAK

SHA

RP E

DG

ES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

WEI

GH

T:

UNIV

ERSI

TI TE

KNO

LOG

IMA

LAYS

IA

TITLE

:D

WG

NO

.:SC

ALE

:1:1

SHEE

T1

OF

1

A4

SURF

AC

E FI

NIS

H:

TOLE

RAN

CES

:(U

NLE

SS S

PEC

IFIE

D)

LIN

EAR:

AN

GUL

AR:

QTY

.:

DEV

ARA

JAN

RA

MA

SAM

Y1/

9/20

04

Alu

min

ium

Mix

er In

let

FAKU

LTI K

EJUR

UTER

AA

N M

EKA

NIK

AL

ASS

OC

PRO

FRO

SLI A

BU B

AKA

R

121

A A

30°

32

36

37

29

36

44

M10

0

9.94

16

20

27

36.14

41.9443.9445.94

A-A

(2

: 1)

M36

Thr

ead

Leng

th 5

mm

DR

AW

N

AP

PV'

D

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

AR

EIN

MIL

LIM

ETER

SFI

NIS

H:

DEB

UR

AN

D

BREA

K S

HA

RP

EDG

ES

NA

ME

SIG

NA

TUR

ED

ATE

MA

TER

IAL:

WEI

GH

T:

UNIV

ERSI

TI T

EKNO

LOG

I MA

LAYS

IA

TITL

E:D

WG

NO

.:S

CA

LE:1

:1S

HEE

T 1

OF

1

A4

SUR

FAC

E FI

NIS

H:

TOLE

RA

NC

ES:

(UN

LESS

SPE

CIF

IED

) L

INEA

R:

AN

GU

LAR

:

QTY

.:

DEV

AR

AJA

N R

AM

ASA

MY

1/9

/200

4

Alu

min

ium

Mix

er O

utle

t

FAKU

LTI K

EJUR

UTER

AAN

MEK

ANIK

ALA

SSO

C P

ROF

ROSL

I ABU

BA

KAR

122

A A

30°

19

29

3.50°

12 X

2

Thr

u

0

7.94

5

25

A-A

(2

: 1)

DR

AW

N

AP

PV

'D

UN

LESS

OTH

ERW

ISE

SPEC

IFIE

D:

DIM

ENSI

ON

S A

RE

INM

ILLI

MET

ERS

FIN

ISH

:D

EBU

R A

ND

BREA

K S

HA

RP

EDG

ES

NA

ME

SIG

NA

TUR

ED

ATE

MA

TER

IAL:

WEI

GH

T:

UN

IVER

SITI

TEKN

OLO

GI M

ALA

YSIA

TITL

E:D

WG

NO

.:S

CA

LE:1

:1S

HEE

T 1

OF

1

A4

SUR

FAC

E FI

NIS

H:

TOLE

RA

NC

ES:

(UN

LESS

SP

ECIF

IED

) L

INEA

R:

AN

GU

LAR

:

QTY

.:

DEV

AR

AJA

N R

AM

ASA

MY

1/9

/200

4

Alu

min

ium

19 m

m R

ing

FAKU

LTI K

EJUR

UTER

AAN

MEK

ANIK

ALA

SSO

C P

ROF

RO

SLI A

BU B

AKA

R

123

124

A A

16

18

7

M10

0

22

32

44

48

A-A

(2

: 1)

M10

Thre

adLe

ngth

4 m

m

DR

AW

N

APP

V'D

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

AR

EIN

MIL

LIM

ETER

SFI

NIS

H:

DEB

UR

AN

DBR

EAK

SHA

RP

EDG

ES

NA

ME

SIG

NA

TUR

ED

ATE

MA

TER

IAL:

WEI

GH

T:

UN

IVER

SITI

TEKN

OLO

GIM

ALA

YSIA

TITL

E:D

WG

NO

.:S

CA

LE:1

:1S

HEE

T 1

OF

1

A4

SUR

FAC

E FI

NIS

H:

TOLE

RA

NC

ES:

(UN

LESS

SPE

CIF

IED

)LI

NEA

R:

AN

GU

LAR

:

QTY

.:

DEV

AR

AJA

N R

AM

ASA

MY

1/9

/200

4

Alu

min

ium

CN

G In

let

FAKU

LTI K

EJUR

UTER

AAN

MEK

ANIK

ALA

SSO

C P

ROF

ROSL

I ABU

BA

KAR

Appendix E

Material Selection

126

MatWeb.com, The Online Materials Database

Aluminum 6061-T8

Subcategory: 6000 Series Aluminum Alloy; Aluminum Alloy; Metal; Nonferrous Metal

Key Words: Aluminium 6061-T8; UNS A96061; ISO AlMg1SiCu, AD-33 (Russia); AA6061-T8

Component Wt. %

Al 98

Cr 0.04 - 0.35

Cu 0.15 - 0.4

Component Wt. %

Fe Max 0.7

Mg 0.8 - 1.2

Mn Max 0.15

Component Wt. %

Si 0.4 - 0.8

Ti Max 0.15

Zn Max 0.25

Material Notes:Weldability = A; Stress Corrosion Cracking Resistance = A; General CorrosionResistance = B (A = best; E = worst). General 6061 characteristics and uses: Excellentjoining characteristics, good acceptance of applied coatings. Combines relatively high strength, good workability, and high resistance to corrosion; widely available. The T8 and t9 tempers offer better chipping characteristics over the T6 temper.

Uses: Aircraft fittings, camera lens mounts, couplings, marines fittings and hardware,electrical fittings and connectors, decorative or misc. hardware, hinge pins, magnetoparts, brake pistons, hydraulic pistons, appliance fittings, valves and valve parts.

Most data provided by Alcoa.

Physical Properties Metric English Comments

Density 2.71 g/cc 0.0979 lb/in³

Mechanical Properties

Hardness, Brinell 120 120 500 kg load/10 mm ball

Hardness, Knoop 150 150 Estimated from Brinell

Hardness, Rockwell A 46.8 46.8 Estimated from Brinell

Hardness, Rockwell B 75 75 Estimated from Brinell

Hardness, Vickers 136 136 Estimated from Brinell

Tensile Strength, Ultimate Min 310 MPa Min 45000 psi

Tensile Strength, Yield Min 276 MPa Min 40000 psi

Elongation at Break 8 % 8 %

127

Modulus of Elasticity 69 GPa 10000 ksi Average of Tension andCompression. In

Aluminum alloys, thecompressive modulus is

typically 2% greater than the tensile modulus

Poisson's Ratio 0.33 0.33 Estimated from trends insimilar Al alloys.

Machinability 50 % 50 % 0-100 Scale (A=90; B=70; C=50; D=30;

E=10)

Shear Modulus 26 GPa 3770 ksi Estimated from similarAl alloys.

Shear Strength 185 MPa 26800 psi Estimated from ultimatetensile strength

Electrical Properties

Electrical Resistivity 3.7e-006 ohm-cm 3.7e-006 ohm-cm

Thermal Properties

CTE, linear 20°C 23.6 µm/m-°C 13.1 µin/in-°F average over 20-100°C

CTE, linear 250°C 25.2 µm/m-°C 14 µin/in-°F Estimated from trends insimilar Al alloys. 20-

300°C.

Heat Capacity 0.896 J/g-°C 0.214 BTU/lb-°F

Thermal Conductivity 170 W/m-K 1180 BTU-in/hr-ft²-°F

Melting Point Min 582 °C Min 1080 °F Solidus

Solidus 582 °C 1080 °F

Copyright 1996-2003 by Automation Creations, Inc. The information provided by MatWeb is intended for personal, non-commercial use. The contents, results, and technical data from this site may not be reproduced either electronically,photographically or substantively without permission from Automation Creations, Inc. No warranty, neither expressed nor implied, is given regarding the accuracy of this information. The user assumes all risk and liability in connection with the use ofinformation from MatWeb.

Appendix F

Mesh Independent Analysis

129

Mesh Independent Analysis

Before further simulation was done a mesh analysis was carried out to

determine the most accurate mesh. Firstly, the model with diameter 16 mm is meshed

in the CFD software using a very coarse mesh this is level-3 meshing in the software.

The mesh is then increased to level 4, level 5, level 6 and level 8 as shown in Table

F1 until the pressure drop is almost constant between two points in the mixer as

shown in Figure F2. The result is level 6 mesh was chosen for the simulation with

refinements as shown in Figure F1. This mesh was considered as the results were not

varying and did not take too much CPU time to calculate

Level 3 Level 4

Level 8 Level 6

Figure F1 Mesh Levels analyzed level 3, level 4, level 6 and level 8

130

0

2000

4000

6000

8000

10000

12000

14000

0 1000 2000 3000 4000 5000 6000 7000 8000

Engine Speed, rpm

Pre

ssur

e D

rop,

Pa

Level 3 Level 4 Level 6 Level 8

Figure F2 Pressure drop for simulation diameter 16 at different mesh

Table F1 Number of cells for each level for 16 mm diameter simulation

Level 3 Level 4 Level 6 Level 8

Fluid Cells 9274 25461 63312 188536

Solid Cells 12553 23009 55505 111835

Partial Cells 14063 25778 53105 100515

Total Cells 35890 74248 171922 400886

CPU TIme1596s27min

4865s1hour 30 min

17893s4 hours 50min

63365s17 hour 36 min