Modelling and simulation of electronically controlled diesel

injectors

Author:
Tran, Xuan-Thien
Publication Date:
2003
DOI:
https://doi.org/10.26190/unsworks/21180
License:
https://creativecommons.org/licenses/by-nc-nd/3.0/au/
Link to license to see what you are allowed to do with this resource.

Downloaded from http://hdl.handle.net/1959.4/19278 in https://

unsworks.unsw.edu.au on 2022-06-08
 THE UNIVERSITY OF NEW SOUTH WALES

SCHOOL OF MECHANICAL AND MANUFACTURING

ENGINEERING

MODELLING AND SIMULATION OF

ELECTRONICALLY CONTROLLED DIESEL

INJECTORS

by

Xuan-Thien Tran

A thesis submitted in fulfilment of

the requirement for the degree of
Master of Engineering

Supervisor : Dr. Michal Tordon

Co-supervisor : Prof. Brian Milton

August 2003, Sydney, Australia

 ABSTRACT

The study presents a one-dimensional, transient and compressible flow models of a

commercial Common Rail Injector (CRI) and a prototype of a single-fuel

Hydraulically actuated Electrically controlled Unit Injector (HEUI) developed at the

University of New South Wales (UNSW) in conjunction with local industry. The

unique feature of the UNSW HEUI is the fact that it uses diesel fuel as the driver for

pressure amplification within the unit injector. The work undertaken is part of a

wider study aimed at optimization of the design of diesel injectors for dual-fuel

systems to reduce green house gas emissions. The contribution of this thesis is the

development of the model of the UNSW HEUI injector, which can be used to

investigate possible modifications of the injector for its use in dual-fuel injection

systems.

The developed models include electrical, mechanical and hydraulic subsystems

present in the injectors. They are based on Kirchhoff’s laws, on the mass and

momentum conservation equations and on the equilibrium of forces. The models

were implemented in MATLAB/SIMULINK graphical software environment, which

provides a high degree of flexibility and allows simulation of both linear and

nonlinear elements.

The models were used to perform sensitivity analysis of both injectors. The

sensitivity analysis has revealed that the temperature of the solenoid coil is one of the

critical parameters affecting the timing and the quantity of the fuel injection of both

i
 injectors. Additional critical parameters were found to be the dimensions of the

piston of the CRI, the stiffness of the needle spring of the HEUI and the dimensions

of the intensifier of the HEUI. The models also revealed that in the case of pilot

injections the speed of the solenoid is the major limiting factor of the performance.

The developed models provide better understanding of the issues and limitations of

the injectors. They give detailed insight into their working principles. The

investigations of the models permit making quantitative analysis of the timing of the

HEUI solenoid and to evaluate the proposed change of the direction of the pressure

acting on the HEUI solenoid plunger.

ii
 ACKNOWLEDGEMENT

I would like to express my grateful thanks to my supervisor, Dr. Michal Tordon, and

my co-supervisor, Prof. Brian Milton, for their consistent guidance, permanent

encouragement and great support through my study in the University of New South

Wales.

I would like to thank my fellow students, Tim White, Canh-Quang Nguyen,

Slobodan Ilic, Daniel Hanafi, Seong-Pal Kang for sharing ideas, knowledge and

funny minutes.

A number of laboratory staff who helped me complete a series of experiments must

be thanked: Tim Anderson and John Dodd of Internal Combustion Engine

laboratory, George Otvos of Electronics laboratory, Alfred Hu of Mechatronic

Systems laboratory and other members of Mechanical School workshops.

Finally I wish to thank my family and friends for their continuous support,

understanding and encouragement.

iii
 CONTENTS

Abstract .................................................................................................................. i

Acknowledgement.................................................................................................. iii

Contents ................................................................................................................. iv

List of figures.......................................................................................................... vii

List of tables............................................................................................................ ix

List of charts............................................................................................................ x

Nomenclature ......................................................................................................... xi

Acronyms ................................................................................................................xiii

CHAPTER 1: INTRODUCTION ........................................................................... 1

Aim of the study.............................................................................................. 4

CHAPTER 2: CRI AND HEUI DESCRIPTION…................................................ 5

CRI ................................................................................................................. 5

HEUI .............................................................................................................. 8

CHAPTER 3: LITERATURE REVIEW ............................................................... 16

CHAPTER 4: SYSTEM IDENTIFICATIONS ..................................................... 26

Fuel characteristics......................................................................................... 26

Spring stiffness coefficient's measurement ................................................... 29

Measurement of small size holes .................................................................. 30

iv
 Friction coefficient ....................................................................................... 30

System of units .............................................................................................. 31

CHAPTER 5: MODELLING COMMON RAIL INJECTOR ............................... 32

The solenoid .................................................................................................. 33

The hydraulic part ......................................................................................... 46

Model implementation in Simulink .............................................................. 52

CHAPTER 6: MODELLING UNSW HEUI INJECTOR ......................................57

The solenoid .................................................................................................. 57

The hydraulic components............................................................................. 58

Hydraulic differential valve, HDV ...................................................... 60

Intensifier.............................................................................................. 63

Injection needle..................................................................................... 65

Model implementation in Simulink .............................................................. 68

CHAPTER 7: EXPERIMENTAL SETUP AND VALIDATION .........................70

CRI ................................................................................................................70

HEUI .............................................................................................................77

CHAPTER 8: SIMULATION ...............................................................................87

CRI ................................................................................................................87

Investigation of multipulse regime...................................................... 87

Sensitivity analysis of the model ........................................................ 93

HEUI .............................................................................................................96

v
 Investigation of the forces of the solenoid ........................................ 96

Sensitivity analysis of the model ........................................................ 99

CHAPTER 9: CONCLUSIONS ........................................................................... 102

Model conversion to C program .................................................................. 104

REFERENCES ...........................................................................................….......105

APPENDIX A: The 2003 IEEE /ASME conference paper ...........................….... A

APPENDIX B: CRI model program ...............................................................…... B

APPENDIX C: HEUI model program ........................................................…....... C

vi
 LIST OF FIGURES

Figure 1. Smoke comparison Common Rail and Conventional Fuel Injection

Equipment............................................................................................... 7

Figure 2. Diagram of CRI .................................................................................... 8

Figure 3. Caterpillar HEUI .................................................................................10

Figure 4. Diagram of the UNSW HEUI ..............................................................13

Figure 5. Spring force coefficient test rig ........................................................... 29

Figure 6. Schema of solenoid force test .............................................................. 37

Figure 7. Force coefficient test rig ...................................................................... 39

Figure 8. Wartsila solenoid ................................................................................. 42

Figure 9. Nippondenso solenoid ......................................................................... 43

Figure 10. Bosch solenoid design ....................................................................... 43

Figure 11. CRI hydraulic part ............................................................................. 47

Figure 12. VCO nozzle ...............................................................................….... 51

Figure 13. Simulation scheme based on derivative blocks ................................. 53

Figure 14. Lord Kelvin’s simulation scheme ..................................................... 54

Figure 15. Block-diagram of CRI model ............................................................ 55

Figure 16. Schema of the HEUI solenoid ........................................................... 57

Figure 17. Simplified schema of the UNSW HEUI .…...................................... 59

Figure 18. Poppet valve of the HEUI ..............…............................................... 62

Figure 19. Block-diagram of the HEUI model ................................................... 68

Figure 20. CRI test rig schema ........................................................................... 71

Figure 21. CRI test rig in the UNSW ................................................................. 73

vii
 Figure 22. The detector location .................................................................….......73

Figure 23. The test of detector delay .......................................................................74

Figure 24. Comparison of timing for the 2ms control pulse ...................................75

Figure 25. Comparison of timing for the 4ms control pulse ...................................75

Figure 26. Simplified diagram of test engine setup for the HEUI ......................... 79

Figure 27. HEUI test engine .................................................................................. 80

Figure 28. Cylinder No 1 .........................................................……...................... 80

Figure 29. Superimposed injection rate traces of the UNSW HEUI experimental

responses with fixed fuel delivery and varied actuating pressure

(adopted from [31]) ............................................................................... 81

Figure 30. The experimental cylinder pressure waveforms …….............................82

Figure 31. The simulated responses with the line pressure of 10 MPa ....................83

Figure 32. The simulated responses with the line pressure of 15 MPa ...................84

Figure 33. The simulated responses with the line pressure of 20 MPa ...................84

Figure 34. CRI response (5ms control signal)....................…............................…. 88

Figure 35. Pilot injection (control time interval of 1 ms) ....................................... 89

Figure 36. Pilot injection (control time interval of 0.9 ms) .................................... 90

Figure 37. Pilot injection (control time interval of 0.8 ms) .................................... 91

Figure 38. Pilot injection (control time interval of 0.7 ms) .................................... 92

Figure 39. Direction of the force of the pressure is against the solenoid

opening ................................................................................................. 97

Figure 40. Direction of the force of the pressure is "helping" the solenoid

opening ................................................................................................. 98

viii
 LIST OF TABLES

Table 1. Distillate characteristics ...........................................................................27

Table 2. Main units …............................................................................................31

Table 3. Comparison of the injection quantities ....................................................76

Table 4. Comparison of fuel delivery ....................................................................85

Table 5. Comparison of duration of injection ........................................................85

Table 6. Sensitivity analysis of the CRI model ......................................................94

Table 7. Sensitivity analysis of the UNSW HEUI model .....................................100

ix
 LIST OF CHARTS

Chart 1. Flow coefficient ................................................................................. 28

Chart 2. Force vs displacement of the plunger (for fixed currents).................. 40

Chart 3. Force vs current (for fixed displacements of the plunger).................. 40

x
 NOMENCLATURE1

A - effective cross sectional area of the solenoid;

d - diameter
e - control voltage of the solenoid;
eb - back emf voltage;
F - force;
f - function (no suffix);
f - coefficient of friction (with various suffixes);
g - the gravitational acceleration;
i - current through the coil;
K I - flow coefficient from inlet line to working chamber;
K O - flow coefficient from HDV to return line;
K Z - flow coefficient of the injection flow;
K b - back e.m.f. coefficient of the solenoid;
K f - force constant of the solenoid;
k - stiffness of the springs (with various suffixes);
L - inductance of solenoid coil;
l - length;
m - mass of the components (with various suffixes);
N - number of turns of the solenoid coil;
PL - pressure (with various suffixes);
qU - injection flow rate;
QU - injection amount of one injection;
R - solenoid coil resistance;
S - cross-sectional area (with various suffixes);
S 4 - effective area of the needle;
S 5 - effective area of the nozzles;
S I - equivalent area of the inlet orifice to working chamber;
S O - equivalent area of the outlet orifice from working chamber;
T - temperature;
t - time;
u , U max - displacement and maximum displacement of the needle;
u 0 - preload of the needle spring;
V - volume (with various suffixes);
x , X max - displacement and maximum displacement of the solenoid plunger;

1
Except in Chapter 3 and where otherwise specified.

xi
 x0 - preload of the solenoid spring;
y - displacement of the intensifier;
y 0 - preload of the intensifier spring;
z , Z max - displacement and maximum displacement of the HDV;
z 0 - preload of the HDV spring;
α - angle;
β - bulk modulus of the diesel fuel;
δ - pressure coefficient of bulk modulus
µ 0 - permeability of free space;
µ r - relative permeability;
Φ - variable
Ψ - variable
Ω - variable

xii
 ACRONYMS

bmep – brake mean effective pressure

CRI – Common Rail Injector

DF – Dual Fuel

ECU – Electronic Control Unit

ECM – Electronic Control Module

EUI – Electronic Unit Injector

HEUI – Hydraulically actuated Electrically controlled Unit Injector

HDV – Hydraulic Differential Valve

I/O – Input/Output

LCD – Liquidized Crystal Display

MOS – Metal Oxide Semiconductor

NG – Natural Gas

PC – Personal Computer

PDE – Pumpe-Düse-Einheit (unit injector)

PLD – Pumpe-Leitung-Düse (unit pump)

PID – Proportional-Integral-Derivative

PWM – Pulse Width Modulation

VCO – Valve Covered Orifice

VGA – Video Graphics Array

VMA – Versa Module Eurocard

xiii
 CHAPTER ONE

INTRODUCTION

Nowadays throughout the world, the calls for a cleaner environment can be heard

everywhere. The legislators in most countries continuously issue tougher and tougher

requirements related to the car industry in particular. Similar procedures are now

being implemented for trucks and buses. This forces automotive producers to apply

new technologies for the purpose of improving efficiency and reducing harmful

exhaust gases. As an important component of the engine industry, diesel engine

producers are also looking towards solving this problem. Although diesel engines are

considered more efficient than spark ignition engines due to their higher brake mean

effective pressure (bmep), there are many tasks for the producers to undertake in

order to meet these tough legislations.

Presently, the main direction towards solution of the problems is the creation of a

new generation of highly effective fuel injectors based on the latest technologies that

incorporate electronic control. There are two recent examples of these new

generation injectors: the Common Rail Injector (CRI) and the Hydraulically actuated

Electrically controlled Unit Injector (HEUI). These injectors are built on a base that

is a combination of a traditional mechanical injector and an electronic control unit

(ECU). The ECU calculates the optimal quantities for combustion parameters, such

as air volume, amount of fuel, ignition timing, injection pressure and exhaust gas

recirculation. It manages the combustion so that it fulfills the driver’s intention in

1
 accordance with road and vehicle conditions. The ECU allows these high-

technology injectors to inject specified amount of fuel at the required instant of time,

independent of the engine speed and crankshaft position. These exclusive features

make CRI and HEUI the most effective diesel injectors so far.

In addition, another approach for the solution of the problems is the application of

alternative fuels for vehicles. Natural gas (NG) is targeted thanks to its clean burning

and plentiful supply. One alternative solution to achieve a highly environmentally

friendly diesel engine is to convert existing truck/bus diesel engines to dual-fuel (DF)

systems using diesel and NG. Full NG operation requires a change to a spark system

with high conversion costs and loss of route flexibility due to the current (and for the

foreseeable future) inadequate NG fuelling infrastructure. Partial NG operation

requires dual-fuelling (DF) where the gas fuel is ignited by a pilot diesel spray. A

well controlled system would use a low diesel percentage where possible but would

have the ability to revert to 100% diesel fuel when NG is not available. To date, the

gas has been generally introduced in the inlet manifold. While this overcomes the

cost and route problems, some combustion problems remain. These can be overcame

by direct injection of NG with the diesel into the engine cylinder. A combined NG,

diesel injector with good control features is therefore required. Both NG and diesel

would have to have high injection pressure and short, precise injection times.

In-cylinder dual-fuel systems require fast acting injectors with precise control of the

combined DF injectors. Processes in modern high-pressure electronically controlled

diesel injectors are complex and involve interactions of electrical, mechanical and

hydraulic systems. Development of the combined dual-fuel injector is a complex

2
 processes requiring detailed understanding of both the diesel and gas components of

the dual fuel injector.

At the University of New South Wales (UNSW), a compact, flexible HEUI diesel

injector has been undergoing development, in conjunction with local industry. The

unique feature of the UNSW HEUI is the fact that it uses diesel fuel as the driver for

pressure amplification within the unit injector. The HEUI under development has

been tested in an engine to an injection pressure of 230 MPa. It can maintain a high

injection pressure for fuel deliveries down to about 1% of the maximum. The

characteristics of this injector make it a good candidate for the diesel components of

an in-cylinder DF (diesel-gas) system. Work is in progress to develop such a DF

(diesel-gas) injector based on the UNSW HEUI [39].

A parallel research in the UNSW is to examine the use of a CRI as the diesel

component for another DF (diesel-gas) injector. CRI is also targeted because of its

current widespread use and its relatively simple construction compared to the HEUI,

as well as its ability to work in the multi-pulse regime. The multi-pulse regime is one

of the promising methods to reduce NOx, smoke and noise in diesel engines. For the

best economical results, further work is required to modify a CRI and HEUI for their

use in the dual-fuel system. It is important to stress that planned DF injectors should

be able to be fitted into existing diesel engines without any significant engine

modification.

3
 AIMS OF THE STUDY

As a part of efforts to explore fuel injection-combustion systems and to modify

diesel injectors’ characteristics and geometrical dimensions, a number of studies

have been undertaken to develop detailed computer models of the CRI systems. The

existing models of CRI in most cases present only the hydraulic part of the injector.

They do not target the detailed modelling of the injector internal components. Due to

relatively short existence of HEUI, to our knowledge, no computer models for them

have been published.

The aim of the present study is to build a one-dimensional, transient and

compressible flow model of a conventional CRI and the first computer model of the

UNSW HEUI. The developed models should include electrical, mechanical and

hydraulic subsystems present in the injectors. They should be based on Kirchhoff’s

laws, on the mass and momentum conservation equations and on the equilibrium of

forces. The models should be implemented in MATLAB/SIMULINK graphical

software environment, which provides a high degree of flexibility and allows

simulation of both linear and nonlinear elements. They should provide a better

understanding of the issues and limitations of the injectors. The models will provide

the tools to evaluate the suitability of the CRI and the UNSW HEUI for the DF

system. They will also help to study any possible characteristic and detail

modifications required before practical implementations of a DF injector.

4
CHAPTER TWO

CRI AND HEUI DESCRIPTION

CRI

It is known that although diesel engines are famous for their efficiency, their exhaust

particulates (smoke) and NOx emission are their significant drawbacks. One way to

make diesel burning better is to improve the atomization of the diesel fuel within the

cylinder. One way of achieving this target is to provide higher fuel injection

pressures than those commonly used in conventional diesel engines.

The first development in that direction was the creation of the PDE injector (Pumpe-

Düse-Einheit – unit injector). The PDE combines pump and injector into one housing

and the generation of pressure is mechanical by means of a cam operating the pump

plunger. There are no high pressure fuel lines to inhibit the use of high pressure and

these systems are capable of injecting at up to 160 MPa. The other variation of the

PDE is the PLD system (Pumpe-Leitung-Düse -unit pump), where the pump and the

injector are separated. In this system, the fuel line is very short and injection pressure

is possible up to 150 MPa. But these injectors, in conjunction with conventional

injectors, are fully mechanical so the injection amount and duration are dependent on

the engine speed and crank angle. Further development along these lines was the

electronic unit injector (EUI) that use a solenoid valve on a spill line to control the

5
 length of the injection process. However, these parameters still remained controlled

mechanically as before.

The turning point was the application of the Common Rail Injector, which was

initially developed by Bosch and Nippondenso. Electronic controls, which allow

greater flexibility in the setting of individual parameters, and reliable sensors, which

give engine control systems increased information, have been combined with

improvements in materials, manufacturing techniques and computing power to

produce a fully flexible fuel injection system whose settings change throughout the

steady state and transient operating regimes. With common rail technology, the

quantity, timing, and pressure of the injection are controllable separately. The high

fuel pressure is supplied by a separate high pressure pump while an Electronic

Control Unit (ECU) calculates and sends a signal to the Common Rail Injector,

which opens and shuts a control solenoid to obtain the required amount of fuel to be

injected at the desired instant of time. By application of the ECU, the fuel injection is

no longer dependent on the position of crankshaft. This feature allows fuel to be

injected as required, which means that a more precise amount of injected fuel can be

metered giving higher diesel efficiency. On the other hand, the fuel pressure is

controlled by the pump from the pressure transducer signal so that it remains stable

at a high pressure of approximately 160 MPa. Higher fuel pressure results in better

fuel atomization, which leads to higher efficiency and less emissions. Lastly, an

electronically controlled device permits a combination of many more related engine

signals than just crankshaft angle and engine speed in the mechanically controlled

counterpart, which leads to much more precise injection. Figure 1 shows how

significantly the CRI reduces exhaust smoke, especially in low load regimes.

6
 Figure 1. Smoke comparison Common Rail and Conventional Fuel Injection Equipment

(http://www.bba-reman.com/crdp.htm)

Figure 2 shows the Bosch CRI. The Bosch CRI , which is under investigation in the

University of New South Wales, works as follows: The injector is connected directly

to the high pressure fuel accumulator (rail) and the high pressure exists permanently

at the needle seat chamber, which allows the injection timing to be controlled

accurately. In order to control the opening and closing time of the needle, a small

chamber (thereafter the chamber is called working chamber) with pressurized fuel is

present on the top of the needle. The chamber is connected to the rail through a small

orifice, which ensures that same pressure exists between the nozzle and the working

chamber when the needle is closed. A solenoid balanced valve, which receives an

electrical signal at an accurately specified time, is located on the top. When the

solenoid valve is open, it creates a pressure drop in the working chamber so this will

cause a negative force which overcomes the force of the needle spring and initiates

7
 the injection. As soon as the solenoid valve closes, the pressure in the working

chamber rises again and forces the needle seat back.

Figure 2: Diagram of CRI [19]

8
HEUI

The HEUI Fuel System represents one of the most significant innovations in diesel

engine technology in the last decades of the previous century. An HEUI system

overcomes many of the limitations of conventional diesel injectors, and achieves the

highest standards for fuel injection efficiency, reliability and exhaust gas control.

HEUI injectors use a liquid in a common rail accumulator at moderate pressure and

amplify this within the injector to give a high injection pressure. One of the biggest

diesel engine producers, Caterpillar, has manufactured HEUI injectors since the mid-

1990s. A feature of the Caterpillar HEUI is that pressure intensification in these

injectors is achieved by using the engine's lubricating oil. That is, it is a two-fluid

injection system. The Caterpillar HEUI, shown in Figure 3, consists of four basic

components [17]:

* HEUI injector, which uses hydraulic energy from pressurized engine lube oil for

injection. The pressure of the incoming oil (5.6MPa to 23 MPa) controls the rate of

injection, while the amount of injected fuel is determined by the Electronic Control

Module.

* Electronic Control Module (ECM). This sophisticated on board-computer precisely

manages fuel injection and other engine systems. The HEUI injector solenoid is

energized by an electronic signal generated in the ECM. Using input from multiple

sensors, the ECM’s dual microprocessors use proprietary software and customer-

9
 supplied performance parameters to produce maximum engine performance under

any conditions.

* High Pressure Pump. The variable displacement axial pump features a built-in

reservoir to immediately supply oil at cold start.

* Injector Actuation Pressure Control Valve. This electronically operated valve

controls the oil pump output and injection pressure.

Figure 3. Caterpillar HEUI.[17]

10
Recently, a group of researchers in the University of New South Wales (UNSW) in

conjunction with local industry is developing new kind of HEUI called UNSW

HEUI. The significant difference of the UNSW HEUI compared to Caterpillar

counterpart is that the UNSW HEUI uses a single-fluid system. This term means that

the UNSW HEUI employs diesel fuel for the driving side of the pressure

intensification in the injector, avoiding the use of the engine's lubricating oil. The

UNSW HEUI is a state-of-the-art injector, having three main significant advantages

over the Caterpillar HEUI:

1. The Caterpillar HEUI system injects at 140 to 160 MPa pressure. The UNSW

HEUI injects at pressures up to 230 MPa.

2. The UNSW HEUI is much smaller and less complicated because the use of the

single-fluid system.

3. The UNSW HEUI can be fitted to standard engines without any changes because

of its smaller diameter while the Caterpillar requires specially designed engines. The

Caterpillar HEUI requires a supply of oil (as a driver) as well as fuel to be delivered

to the engine. This makes it physically larger than conventional mechanical injector

and sometimes difficult to fit to new engines that are not purposely designed to

receive it. The two-fluid systems are larger than single-fluid ones and therefore

present more difficulties in this respect.

11
 The description of the UNSW HEUI presented here is compiled from [31]. HEUI

injectors use hydraulic rather than mechanical energy to raise the pressure of the fuel

to a level suitable for direct injection. This is achieved with a differential piston

inside the unit injector. The injection timing, duration and thus quantity is controlled

by a solenoid. The HEUI injectors are fed by a common rail. However, since the

injector contains its own pressure-amplifier, the rail pressure is significantly lower

than the injection pressure. The controlling solenoid works against this lower

pressure.

A particular benefit of the UNSW HEUI is that it also exhibits a very high turndown

ratio to less than 2% of the maximum delivery [42] whilst maintaining high injection

pressure.

12
 Figure 4: Diagram of the UNSW HEUI [38]

The design of the UNSW HEUI also allows high injection pressures and better

control of the fuel flow than the Caterpillar since the solenoid control valve is

subjected to a much lower fluid pressure. Injection pressures of up to 230MPa have

13
 been achieved whilst the solenoid may have to cope with only a little over 20MPa

from the accumulator. Figure 4 shows how this HEUI uses diesel fuel as the driver

for the amplification instead of an alternative oil supply.

The UNSW HEUI works as follows: In the initial position, the solenoid valve is not

energized and the connection via the throttling hole from the control chamber to the

return line is closed. The hydraulic differential valve (HDV) is closed and the line

pressure in the control chamber forces the intensifier into its bottom position. The

pressure in the needle chamber combined with the force of the spring keeps the non-

return valve closed. This prevents leakage into the combustion chamber of the

engine, which could otherwise occur due to worn nozzles.

When the solenoid is energized, it opens the port and allows the fuel to flow from the

working chamber to the return line. The flow of the fuel is restricted by the poppet

valve. When the pressure in the working chamber has decreased to a specific level,

the intensifier will start to move upwards, driven by the pressure difference between

top and bottom of the intensifier. This intensifier movement causes pressure to

decrease in the needle chamber, thus initiating the opening of the non-return valve.

This is the fuel-metering phase of the operation, called preliminary metering.

The fuel-metering phase is terminated by the de-energizing (closure) of the solenoid.

After this closure the pressure difference between the inlet line and working chamber

forces the HDV to move downwards. This allows the fuel to enter the control

chamber. The pressure in the control chamber increases and causes the intensifier to

move down, thereby compressing the fuel in the needle chamber. The compressed

14
 fuel overcomes the force of the needle spring, opens the nozzle and initiates the

injection of the fuel into the combustion chamber.

This is the design of the UNSW HEUI, which is to be investigated and modeled. As

mentioned earlier, the UNSW HEUI is in the prototype stage so it has a number of

slightly different-from-each-other models. The HEUI described above is the basic

design of the UNSW HEUI so it is the one used for modeling. Additional sub-models

for added features in the HEUI design could be easily modified from the basic

model.

15
 CHAPTER THREE

LITERATURE REVIEW 2

In a study of CRI system, Woermann et al. [40] describe a high-resolution real-time

model of pump, rail, control valve and injector, which results in an approximation of

the dynamic characteristics of pressure and mass flow of the fuel. The model is

combined with a steady-state model of a diesel engine. Input variables are the engine

crank angle φ, the signal for the pressure control valve ιc and the control signal ιI for

the injector. It calculates the rail pressure pR, the amount of injected fuel mp, mI and

the start of injection φp, φI for pre- and main injection. These values feed a steady

state model triggered every ignition stroke to calculate the torque ME of a diesel

engine with engine speed ω.

The model was built based on a number of approximations and assumptions. For

example, the injector control signal is first processed by a delay line to approximate

the time delay Td caused by hydraulic effects in the injector. The mass flow mF

through the injector is assumed to be proportional to the relative opening sI of the

needle and the rail pressure pR . The opening of the control solenoid valve sC is

assumed to be proportional to the current iC of the solenoid, and the valve opening is

assumed to be at maximum for all iC higher than a certain value of current (ic,max).

These approximations and assumptions may be too restrictive. This approach does
2
The symbols in this chapter may be different from those of the Nomenclature as we keep the
symbology of the referred articles.

16
 not allow observing and quantitatively calculating the influence of a specific

injector component on a specific injector delay.

The diesel engine simulator used in this study consists of a multiprocessor VME-bus

system as the real time unit. System is very powerful and enables real time

simulation of their relatively simple model.

The simulated and experimental results show good agreements of dynamic

characteristics so for the calibration purpose it was easy to optimize parameters of

the control algorithm implemented in the ECU. It is useful to test and verify the

implementations under real-time conditions.

The model has proved itself to be a useful tool for parameter optimization for the

ECU control algorithm, but a number of assumptions and assumed constant delays

make the model not suitable for detailed optimization of the injection system. The

model building does not allow prediction of system characteristic changes caused by

each injector detail so it cannot be useful for the purpose of injector’s detailed

geometric optimization.

Arcoumanis et al. [3] present a one-dimensional, transient and compressible flow

model that was used in order to simulate the flow and pressure distribution in

advanced high-pressure fuel injection systems. These include electronic distributor-

type pumps with either axial or radial plungers and a common-rail system. Different

types of injector nozzles have been used to validate the model. The model is targeted

17
 to reveal the important parameters of the injector which are difficult to obtain by

experiments, such as the actual injection pressure at the nozzle tip, the effective hole

area at the hole exit due to the presence of hole cavitation, the fuel injection

temperature and the hole-to-hole variability of inclined nozzles. In the model, the

flow directions are assumed to coincide with the pressure variation along the fuel

injection system, this being justified on the basis that in most engine systems the

length of the pipe is much longer than their diameter. The flow elements used in this

study consist of a cam plate profile, pressure profile, pipes, valve, volume, chamber,

slots and nozzle tip. In addition to the above elements, various flow phenomena have

been considered to enhance the model predictions. The flow phenomena considered

include cavitation, variable discharge coefficients, pressure losses, pipe wall

flexibility and leakage. Fuel viscosity, compressibility, density and other fuel

properties such as surface tension, latent heat of vaporization, specific heat are also

simulated and assumed to be functions of temperature, local pressure and vapor

volume.

The model simulation and validation show a good match between simulated and

experimental results. But the model does not include a solenoid valve into its

simulation so a number of important delays and responses related to the solenoid are

not examined.

Yudanov [42] describes the key design factors of an HEUI which determine its

injection rate, pressure, energy efficiency and accuracy of fuel delivery, and the

correlations between these design factors and performance parameters. The study

18
 describes two types of HEUI. The first type has a built-in small accumulator and the

second uses a common external accumulator. The second type obtains significant

advantages because it allows independent control of fuel delivery and injection

pressure as well as the ability of obtaining a desired shape of an injection pressure

curve. This second type is the base for the development of the UNSW HEUI used in

our study.

As stated in [42], the significant feature of the HEUI is the application of an internal

hydraulically driven plunger, which makes the HEUI significantly different from

conventional mechanically driven fuel injection systems or from the CRI. An

important characteristic is that the injection pressure is practically independent of the

2
⎛D⎞
intensifier’s diameters d and D if the value M = ⎜ ⎟ is kept unchanged. This
⎝d⎠

means that the HEUI pressure amplification can be achieved using a much smaller

diameter for the plunger than in the case of mechanically driven injection systems.

This important feature allows the injector size minimization and an improvement of

the stability of the fuel delivery in consecutive injections.

Another unique feature of the HEUI is that the correlation between the injection

pressure and flow area (µƒnz) of the nozzle can be very small while, in the

mechanically driven injector, the injection pressure is roughly inversely proportional

to the flow area. It is one more advantage of the HEUI because changes in the flow

area due to wear or coking of the nozzle’s orifices will not significantly affect

injection atomization.

19
 The paper also presents two methods of fuel metering in the HEUI. The preliminary

metering method is chosen because it avoids the drawbacks of the direct metering

method. For all electrically controlled injectors, the correlation between the duration

of the electric control pulse applied to the solenoid and the actual fuel delivery is an

important factor. For the injector with the direct metering method, this correlation

changes according to changes in the total flow area (µƒnz) of the nozzle’s orifices.

The HEUI with preliminary metering is free from this drawback so the control is

more precise and independent of the condition of the nozzle.

The paper also presents the working principle of the HEUI and supplies important

equations for the calculations of the injector characteristics, but it does not present a

mathematical model of the HEUI. To understand the complex processes of the

HEUI, a mathematical model needs to be developed.

Milton et al. [31] describe the research for development of a tuneable diesel engine

injection system. The research was undertaken in the University of New South

Wales. The project used a medium pressure Accumulator Fuel System because a

high pressure Accumulator Fuel System has a number of disadvantages. The goals

for development of a new HEUI system are:

- Injection pressure should be able to reach at least 200 MPa;

- The HEUI should be able to fit on conventional fuel systems without any design

changes;

20
 - The whole accumulator fuel system should be installable on a conventional diesel

engine with no alterations to the engine design;

- An electronic device for injection shape control should be feasible.

The HEUI used in this study was manufactured in the University of New South

Wales in conjunction with local industry. The maximum possible injection pressure

of this HEUI was quantitatively determined based on other injector characteristics:

Pac M
PF =
1 + M E ( µf nz / µf in max ) 2
3

where PF is maximum possible injection pressure, M is geometrical intensification

factor of the injector, i.e. the ratio of the areas of the intensifier's ends, µf nz is the

effective flow area of the nozzle, µf in max is the maximum effective flow area of the

inlet valve, E is a coefficient depending mainly on the friction losses within the

HEUI, and Pac is the hydraulic accumulator pressure.

The results achieved in this study show significant advantages of the UNSW HEUI

system compared to the standard Bosch fuel system in the same engine under

identical working conditions. For example:

- The smoke emission is reduced by more than 53.3%

- The NOx emission is reduced by more than 3.2%.

21
 The research also suggests a number of methods of engine parameter optimisation.

The methods of optimisation of the injection timing and injection pressure are

presented using the computer-based controller. There are also suggestions related to

improvement of the injector and the controller so that the system can be finally

commercialized.

The research is also an experimental approach to improve the HEUI performance. It

is focused on building a precise, high-level Test Controller and performing an

analysis of the HEUI performance based on received measurements. It does not

present a theoretical base to obtain a dynamic model for the HEUI. Our study would

build mathematical and computer models for this HEUI.

Schechter [35] describes a new concept of a fast response multipole solenoid and

discusses factors leading to faster solenoids. The factors, which affect the response of

a solenoid to an electrical signal, are the time constant of the solenoid coil and the

ratio of the magnetic traction force to the moving mass. The time constant

determines the time delay involved in building up the magnetic force to the required

magnitude, while the force to mass ratio represents the acceleration of the moving

mass.

Analysis of mathematical relationships between various parameters of a solenoid coil

indicated that the time constant T can be expressed by a simple equation as follows:

22
 2 Fl
T=
P

where F is a single magnetic pole, l is the initial air gap, and P is the power input.

Because the magnetic pole and initial air gap are usually fixed for a solenoid, the

time constant of a particular solenoid depends on the input power only. A fast

response solenoid is a high energy solenoid and must have a high power to force

ratio, at least during the activation period. The force to the moving mass ratio

declines with increase in the force and size of the coil.

The author also presents a new concept of a fast response solenoid, which is a ring-

shaped multipole solenoid where a multitude of magnetic poles of alternating

polarity appear on the traction surface of the solenoid core. A list of equations, which

describe the main solenoid parameters, such as current, voltage, force, acceleration,

velocity, travel and travel time, is presented. Schechter made the conclusion that

multipole solenoids having very large forces can be designed with the fast responses

usually associated with a small solenoid coil.

Although the study shows a good theoretical base for solenoid calculations, it

requires an enormous experimental facility to determine a number of the coefficients

and parameters. A separate study is required to build a solenoid model based on the

equations, which are presented by Schechter. In our approach we are trying to build a

simple model for the small size type of solenoids.

23
 Cheung [7] presents a project which has the purpose of building control models for

solenoids to evaluate the feasibility of using them in force and position control. Two

models of solenoids are presented.

The first model is based on linear magnetic principles. The equations presented in the

model assume that saturation does not occur; hysteresis and eddie current effects are

also omitted. A software simulation package SIMNON is used to perform the

dynamic simulation, which simulates the dynamic response of the plunger when a

constant voltage source is applied to the solenoid. The simulated results illustrate a

significant drawback of the model showing large differences between simulated and

experimental results (The simulated plunger travel time is 18 ms vs. 32 ms actual

time).

The second model is non-linear and considers that the relative permeability and

inductance are functions; the flux linkage of the reluctance has non-linear relations

with current as well as with air gap distance. For the second model, induced flux is

measured to establish the magnetic characteristics of the solenoid. Also hysteresis

loops at different plunger positions and current levels are obtained from the

measured data. The simulated results of the overall system show a better match with

actual solenoid responses. Position, velocity and current curves are approximately

the same shape as the measured dynamic profiles. However, the predicted travel time

of 69 ms is much shorter then the measured travel time of 101 ms.

The inaccuracy occurs because the equations do not reflect the reality in some

sensitive instances, such as when the solenoid begins moving or stopping. The

24
 models also do not include hysteresis and eddie current into the calculation.

Moreover, the effect of temperature is not considered in the models, and it may cause

changes in the solenoid electrical characteristics. The models are difficult to apply to

a miniature solenoid like the CRI or the HEUI solenoids, in which the very short

movement makes the above-mentioned magnetic characteristic measurements

practically impossible.

25
 CHAPTER FOUR

SYSTEM IDENTIFICATION

FUEL CHARACTERISTICS

The system uses distillate as fuel. For all calculations, the main characteristics of

distillate were taken from the Table 1.

In Table 1, the fuel density is shown at 15oC. It is well known that fluid density

heavily depends on fuel pressure. It also depends on fuel temperature but in a smaller

scale. Therefore, the fuel density will be calculated with a thermal expansion

coefficient of 480⋅10-6 K-1 (Source: from the Internet http://imartinez.etsin.upm.es/)

and pressure coefficient of 4⋅10-7 kg/(m3.Pa) (Source: from the Internet

http://www.amesim.com). The density is assumed to be identical for every particular

element volume, in which it occurs.

26
 CIMAC Requirements 1990 for Distillate Fuels for Diesel Engines

Characteristics Limit CIMAC CIMAC CIMAC CIMAC

DX DA DB DC
Related to ISO 8217 Specification ISO-F ISO-F ISO-F ISO-F
DMX DMA DMB DMC
Residual Inclusion None None Some Allowed
Trace
Density at 15°C, kg/m³ max. - 890 900 920
Kinematic Viscosity at 40°C, cSt (1) max. 5.5 6.0 11 14
min. 1.4 1.5 2.5
Flash Point, °C min. 43 60 60 60
Pour Point °C Winter (5) max. - -6 0 0

°C Summer max. - 0 6 6
Cloud Point, °C max. -16 - - -
Carbon Residue

Ramsbottom on 10% res, % m/m max 0.20 0.20

Microcarbon % m/m max 0.25 0.25
Ash, % (m/m) max. 0.01 0.01 0.01 0.03
Sediment by Extraction, % (m/m) max. - - 0.02 -
Total Sediment, % (m/m) max - - - 0.05
Water, % (v/v) max. - - 0.30 0.30
Cetane Number (4) min. 45 40 35 35
Visual Inspection Clear (3) Clear (3) May be May be
Black Black
Sulphur, % (m/m) max. 1.0 1.5 2.0 2.0
Vanadium, mg/kg max - - - 100
Aluminium plus Silicon, mg/kg max. - - - 25

Table 1. Distillate characteristics.

(Source: http://www.plazamarinefuel.com/cimac1990distillatefuel.htm)

There are some difficulties in determining the Reynolds number for different flows

inside the injectors. Because the injection occurs over a very shot duration of time

and under high pressure, it is assumed that all flows are turbulent and therefore the

27
 d
Reynolds number is bigger than 3000 for all of flows. The relationship (where d
l

is the diameter of the orifice and l is the length of the orifice) applicable for all

orifices in injectors is less than 0.1 so the flow coefficient for flow calculations in the

models can be taken as equal to 0.6, according to Chart 1.

Chart 1. Flow coefficient [34]

28
SPRING STIFFNESS COEFFICIENT'S MEASUREMENT

All springs used in the two high-technology injectors are miniature but they are stiff

enough to bear high pressure forces in the injectors. This feature causes a need to

manufacture a special tool for the measurement of the spring stiffness coefficients.

Figure 5 shows this special tool under working conditions.

Figure 5. Spring force coefficient test rig.

It is a piston-cylinder system, with the clearance between piston (1) and cylinder (2)

being 0.05 mm. It is lubricated by oil and has a tray (3) to place a weight (4) in. The

spring is placed inside the cylinder and when a certain weight is put on the tray, it

29
 moves downwards. The displacement of the tray is recorded by two micrometers (5)

and (6). The micrometers are first placed bottom up, and then top down. The mean

value of these measurements will eliminate the influence of the micrometer spring’s

force on the measurement system.

MEASUREMENT OF SMALL SIZE HOLES

For measuring the very tiny holes of these high-technology injectors, the Nikon

Shadowgraph model 6C is employed to measure holes with diameters less than 1

mm. The weakness of the shadowgraph is that, if the measured hole is not cylindrical

and access to the hole is available from just one end, the result is only the diameter of

the accessible end. This drawback may cause some errors in measurement and

therefore it may result in some lack of precision in the models.

FRICTION COEFFICIENT

Friction coefficient (the ratio of the force resisting motion to the applied load) is a

key design parameter in all machinery with moving parts. Frequently published data

for friction coefficients is difficult to find. Since friction is very dependent on the

contacting materials and the state of the interface (i.e. lubricated, dry, wet,

contaminated etc.), it is usually best to measure friction under conditions as close to

those of the operation as possible.

30
 Because of time limitation of this study, the friction coefficient is taken as constant

for all contacts inside the injectors. Its value is assumed to be equal to 0.16 for well

lubricated surfaces of the steel parts (source: From the Internet

http://www.carbidedepot.com/formulas-frictioncoefficient.htm )

SYSTEM OF UNITS

The system of units used in this study is SI Sytem of Units. The units for the main

quantities are as follows:

Quantity Unit

1 Pressure pascal

2 Temperature degree of Kelvin

3 Weight kilogram

4 Length meter

5 Square square meter

6 Volume cubic meter

7 Velocity meter per second

8 Flowrate cubic meter per second

9 Voltage volt

10 Current ampere

11 Electric resistance ohm

Table 2. Main units

31
 CHAPTER FIVE

MODELLING THE COMMON RAIL INJECTOR

As mentioned earlier, the development of a new injection system requires a high

level of understanding of the injectors in general and their parts in detail. One of the

possible methods of investigating the characteristics of injectors is an examination of

their computer models. Such modeling should be a fast and cheap way to determine

the potential for detailed modifications and to predict the results.

While several CRI injection systems are present in the market, the only commercial

HEUI is that of Caterpillar. The UNSW HEUI is in the prototype stage and huge

efforts are being made to perfect it. Moreover, the application of dual-fuel injection

systems and the multi-pulse injection regime require further modification and

improvement of both the CRI and HEUI types.

For the purpose of modification of the CRI and HEUI systems to fit them in the new

designs, the traditional “trial and error” method is not considered to be useful.

Because they are working under extremely high pressure, the mechanical parts of

these injectors have to be made with extra high precision. For example, the tolerance

of the HDV piston diameter of the UNSW HEUI is less than 10 µm. These tolerances

make the “trial and error” method especially time consuming and financially costly.

32
 A computer model may overcome these drawbacks of the “trial and error” method

and can be used for investigation of the characteristics from further geometrical

modification of the injectors. This is also the aim of this study, which aims at

building a proper model of these high-technology injectors. The chosen software is

MATLAB SIMULINK for a number of reasons. Firstly, SIMULINK is strong in

modeling complex nonlinear systems. Any system, which can be defined by

continuous differential equations, could be modeled in SIMULINK. It also allows

adding non-linear features into the model, which gives a high level of flexibility to

the model. Secondly, in SIMULINK it is easy to build and analyze the model.

SIMULINK permits the placement of Scopes at any point in the model so that, not

only behavior of the whole system can be observed, but also that the response of any

part of the system can be analyzed. This feature permits an observation and effect

valuation of any change in system characteristics or the detailed dimensions on the

overall system behavior as well as on each individual component. Thirdly, models in

SIMULINK can be easily converted to C code. This is a convenient way to obtain C

code for further real-time control system testing. Finally, the availability of

MATLAB SIMULINK in the University environment makes the model easy to be

built and to be shared.

THE SOLENOID

The solenoid is important part of both the CRI and HEUI. The solenoid receives an

electrical signal from the ECU and opens port(s) in the CRI or HEUI to begin the

injection process. Because of the short duration of the injection time, especially in

33
 high-speed engines, these solenoids must be extremely fast. In addition, the

dimensions of the solenoid must be reduced to a minimum to fit on top of the

injectors. It is therefore difficult to measure the solenoid force and magnetic

characteristics.

Cheung et al. [8] describe the dynamic behaviour of the solenoid based on a linear

magnetic circuit model. This model is relying on standard magnetic circuit principles

where the force and the inductance of the solenoid can be expressed as:

µ 0 N 2 Ai 2
F ( x, i ) = (5-1)
x2

µ0 N 2 A (5-2)
L( x) =
l
+x
µr

where x is the plunger position (displacement), N is the number of turns of the

solenoid coil, A is the effective cross-sectional area, l is the effective return

magnetic path, µ 0 and µ r are permeability of free space and relative permeability

respectively.

The voltage equation can be represented by:

d ( L( x).i )
e= + Ri (5-3)
dt

therefore the dynamic equation of a solenoid can be expressed as:

34
 d 2x (5-4)
m1 . = F ( x, i ) − m1 g − k1 .x
dt 2

where k f is the spring constant, m1 is the mass of the plunger, and g is the

gravitational acceleration constant.

The problem of the above model is the fact that it does not include the phenomenon

of the back e.m.f. voltage generated in the electrical circuit of the solenoid due to the

movement of the plunger. This was also pointed out by Cheung et al. [8] when they

showed large discrepancies between the model and experimental results.

They have subsequently introduced the non-linear magnetic model which included

an extra term in the voltage equation (5-3). The term represented the back e.m.f.

voltage generated by the movement of the steel plunger. The voltage equation of the

solenoid was expressed in a new form:

d ( L( x ).i )
e= + Ri + eb (5-5)
dt

with the back e.m.f. voltage eb is:

dx (5-6)
eb = K b
dt

It should be noted that the force generated by the solenoid is proportional to the

current via the force constant K f and can be expressed by the equation:

35
 F ( x, i ) = K f ( x).i (5-7)

Although functionally the force constant (parameter) K f ( x ) and back e.m.f. constant

(parameter) of the solenoid K b ( x ) are two separate parameters, they are closely

related and for the SI unit system they are numerically equal:

K f ( x) = K b ( x) (5-8)

This relationship is similar to that one of back e.m.f. constant and torque constant of

a DC motor which are also equal for the SI unit system as shown in [26].

In further development of models of the CRI and the UNSW HEUI we will be using

the generic solenoid model described by equations from (5-4) to (5-8).

There are principle difficulties in measurement of K b ( x ) parameter of the model.

The use of equation (5-8) allows establishment of K b ( x ) through measurement of

K f ( x ) . For the measurement of K f ( x ) we have made an experimental setup. The

working principle of this setup is shown in Figure 6.

The force generated by the solenoid coil is determined by two sets of experiments.

36
 Figure 6. Schema of solenoid force test.

In the first set of experiments:

• The solenoid plunger is scaled with a weight. The solenoid is located

vertically, bottom down (see Figure 6).

37
 • DC current is applied to the coil. An amperemeter is connected to measure

the current. A resistor is connected to adjust the current.

• A piece of paper, the thickness of which is precisely measured by a

micrometer, is put on top of the solenoid plunger.

• The solenoid plunger with the weight are hooked to the solenoid and kept

fixed in position under the force of the solenoid coil.

• The current is reduced gradually, with low speed, until the plunger with

weight falls down.

At this instant, when the plunger with weight falls down, the force caused by the

solenoid coil is assumed to be equal to the force of gravity of the plunger-weight

system. The value of current applied to the solenoid is recorded.

This step is repeated several times to obtain an average value. After individual

measurements the solenoid was demagnetized to avoid residual magnetism.

In the second set of experiments, the current remained constant while the weight was

gradually changed to find the force that correlates with plunger displacement. The

experiment was performed to check the results of the first set of experiments and to

establish the relationship between the force generated by the current through the coil

of the solenoid and the individual displacements of the plunger.

38
 Figure 7 shows the test setup in the laboratory to measure these characteristics of

the solenoid. The solenoid (1) holds the weigh (2) under a current supplied by the

power supply (3). The multimeter (4) measures the current through the solenoid coil.

Figure 7. Force coefficient test rig.

A sample of the experimental results of the solenoid force measurement is presented

in Chart 2.

39
 35
30
25 I=1 A
Force (N) 20 I=2 A
15 I=3 A
10 I=4 A
5
0
0 0.2 0.4 0.6
Displacement of the plunger
(mm)

Chart 2. Force vs displacement of the plunger(for fixed currents)

2
y = 0.3866x - 0.9771x + 0.578
3
2.5
2
X=0.5 mm
Force (N)

1.5
1 Poly. (X=0.5
mm)
0.5
0
-0.5 0 2 4 6
Current (A)

Chart 3. Force vs current (for fixed displacements of the plunger)

40
 For the purpose of modeling, it is useful to obtain analytical expression for the

dependence of the force of the solenoid on the displacement of the plunger. When a

polynomial of the fourth order is applied to the experimental data, the approximate

functions F = f (x ) , where F is the force of the solenoid and x is displacement of

the plunger, are established as shown in Chart 3. Based on experimental results we

have establish 5 functions F = f (x ) namely for measured current dependences for

current 1,2,3, 4 and 5 amperes.

The force F ( x, i ) is calculated by the experimental formulas as explained above,

these being determined for discrete values of current ii . Inside the interval ii - ii +1 , the

force is assumed to be directly proportional to the increment of ii , i.e.

ij
F j ( x, i ) = f i ( x). (5-9)
ii

where ii ≤ i j ≤ ii +1 .

The solenoid is located in the hot compartment of a diesel engine so the effect of the

temperature on the coil resistance must be included [8]:

R = [1 + α ⋅ (T − 20)].R20 (5-10)

41
 where R20 is the coil resistance at 20oC (297 K), α is the temperature coefficient,

and T is the actual temperature of the solenoid coil. It could be taken equal to the

temperature in the engine compartment.

There are several designs of the solenoid “actuation method”. For example, the

Wartsila solenoid has a plunger as a spool valve [16]:

Figure 8. Wartsila solenoid[16].

while Nippondenso [18]:

42
 Figure 9. Nippondenso solenoid[18].

and Bosch [19] use a ball valve as an actuation device. The solenoid plunger keeps a

ball underneath and close to the orifice which connects the return line to the working

chamber. When the solenoid is energized, the plunger rises allowing the pressurized

fuel to force the ball up to open the port. The opened port allows fuel to flow through

the orifice and relieve the pressure in the working chamber.

Figure 10. Bosch solenoid design[19].

43
This design of the valve has a number of advantages:

• The design of the valve with a ball eliminates the friction between the

plunger and its liner such as that which exists in the case of the spool valve.

• The direction of force of solenoid coil is in agreement with the direction of

force of the pressurized fuel to open the port. This provides shorter opening

time.

So the movement of the solenoid plunger for the CRI injector is modeled by the

following equation:

d 2x dx
m1 . 2
= K f .i + PL S1 − m1 g − k1 .( x + x 0 ) − f 1 . (5-11)
dt dt

where m1 is the mass of the solenoid plunger, x is the plunger movement, x max is

the maximum plunger movement, PL is the pressure in the working chamber, S1 is

the cross-sectional area of the outlet orifice (see Figure 11), k1 is the stiffness of the

solenoid plunger spring, g is the gravitational acceleration, and f 1 is the friction

coefficient between the plunger and the stationary parts of the solenoid.

The driving force for this movement of the solenoid plunger is the current of the

solenoid coil helped by pressure in the working chamber and opposed by solenoid

spring force, friction force and the force of gravity.

44
There are a number of assumptions for the hydraulic part of the solenoid model

which are applied because they do not greatly affect the results while making the

model much less complicated:

• Firstly, the coil inductance L is assumed to be constant. This assumption is

made because the plunger travels outside the internal part of the solenoid coil

and the travel itself is small (up to 0.5 mm).

• Secondly, the weight of the ball is ignored (actually it weighed 0.0097 gram,

small enough to be ignored).

• Thirdly, the force of the pressurized fuel in the working chamber, which acts

on solenoid plunger, is assumed to be inversely proportional to the plunger

movement, e.g. it obtains its maximum value when the valve is closed and

reduces to zero when the plunger reaches the top upper position.

• Lastly, any possible friction due to bulging of the solenoid spring is ignored.

Even though the spring is long enough, the movement of the plunger is

relatively small and the system is submerged in fuel so it is assumed that the

spring does not bulge sideways. Even if this happens, it is small enough to be

ignored.

45
 The above equations present a mathematical model of the Common Rail Injector’s

solenoid. It is also applicable to other designs of solenoids, for example, that of the

HEUI solenoid as mentioned later.

THE HYDRAULIC COMPONENTS

The equations used for building mathematical models of the injectors, as mentioned

earlier, are based on well-known equations of fluid mechanics, which are presented

in [34].

When the pressure in the working chamber (see Figure 11) drops, the force of the

pressurized fuel in the needle chamber overcomes the force of the compressed needle

spring and causes the piston and the needle to move up, opening the nozzle and

letting an injection occur. Although the piston and the needle are designed as

separate parts, in operation they act as a united part so in the modeling they are

considered as a single component.

For an hydraulic system such as this CRI, the system is considered as a spool valve

system. The rate of flow of fluid through such a valve depends on the spool

displacement from the null position (valve closed) and on the pressures upstream and

downstream of the valve.

46
 Figure 11. CRI hydraulic part

For the model of the hydraulic components of the CRI, the following assumptions

are applied:

• Speed of the pressure propagation is infinite, meaning related events are

simultaneous. The assumption is based on the fact that the distances inside

the injector are sufficiently short (maximum distance is not over 0.1 m while

the pressure wave requires just 1ms to travel a distance of 1.5 m).

47
 • Fuel leakage between the components of the injector is ignored.

x
• The outlet orifice’s cross-sectional area is proportional to .
X max

• The pipe wall expansion also is not taken into calculation. The actual inside

diameter is 4.3 mm while the wall thickness is about 10 mm, which ensures

that the metal expansion can be ignored.

• It is assumed that no air is present in the fuel.

With these assumptions, a one-dimensional, transient and compressible flow model

of the hydraulic component of the CRI is built based on mass conservation, force

equilibrium equations. These equations describe flows through an orifice, and take

into consideration the fuel compressibility. Although the fluid compressibility is

often ignored in steady state flow, for transient flows, especially in high pressure

regimes, it may be extremely large. Transient flowrate associated with fuel

compressibility are likely to be significant compared with other flowrates in a

system. This transient flowrate may be expressed as:

V dP
qc = . (5-12)
β dt

where V is the volume of the fuel involved, β is the effective bulk modulus and

dP
is the rate of change of pressure of the fuel.
dt

48
The fuel flow through the outlet orifice when the solenoid valve is opening and

closing ( q Of 5 = K I .S1 . PL − PW ) causes a flow out through outlet orifice

x V w 0 dPW
( q Of 6 = K O .S O . . PW − PR ), acts on the fuel volume ( . ) and moves the
X max β dt

dy
piston system ( S 2 ).
dt

Hence the mass conservation equation for the working chamber can be expressed as

follows:

dy V w0 dPW x
S2 − . = K I .S1 . PL − PW − K O .S O . . PW − PR (5-13)
dt β dt X max

where V w 0 is the volume of the working chamber, β is the bulk modulus of the

fuel, PW is the pressure in the working chamber, PL is the pressure in the line

(accumulator), PR is the pressure in the return line. Also, S 0 is the cross-sectional

area of the outlet orifice, S1 is the cross-sectional area of outlet orifice, S 2 is the

cross-sectional area of the piston, K I is the inlet flow coefficient, and K O is the

outlet flow coefficient.

In this equation, the fuel bulk modulus β is a function of pressure, which is

determined specifically for a particular fuel. Therefore, β is calculated by an

equation adopted from the curve in [15]:

49
 β = β 0 ⋅ [1 + δ ⋅ ( P − P0 )] (5-14)

Here β 0 is the bulk modulus for distillate at 10 MPa, δ is a coefficient, P is the

current pressure in MPa, and P0 is the base pressure (10 MPa).

The force acting on the piston system is the force of the pressurized fuel in the needle

chamber. This overcomes the force of needle spring, the force of the pressurized fuel

in the working chamber and the force due to friction. The actuating force causes the

piston system to move upwards, open the nozzle and provide an injection.

The piston system motion equation is:

d2y dy
m 2 . 2 = PL ⋅ S N − PW .S 2 − k 2 .( y + y 0 ) − f 2 . (5-15)
dt dt

where m 2 is the weight of the piston system, y is the piston movement, y 0 is the

pre-compressed spring movement, S N is the effective cross-sectional area of the

needle, k 2 is the spring force coefficient, and f 2 is the friction coefficient in the

system piston-liner.

It is useful to note that the nozzle type is valve covered orifice nozzle (VCO-nozzle)

(see Figure 12). This design of nozzle stops the injection flow immediately after the

needle regains its seat. It eliminates the main drawback of the sac-hole nozzle. Sac-

50
 hole nozzles keep the hole-to-hole variation low but cause more fuel to dribble after

the end of injection.

Figure 12. VCO nozzle[13]

Based on the “flow through an orifice” theory, the injection flow rate is calculated by

the following equation:

y
q I = K I .S I . . PL − PW (5-16)
Ymax

where q I is the injection flow rate, K I is the flow coefficient through the nozzles,

S I is the effective nozzle area, Ymax is the maximum movement of the needle.

For the VCO-nozzle, there is an assumption that the actual effective nozzle area is

y
directly proportional to , i.e. the injection flows through "changing" orifices,
Ymax

51
 y
which is equal S I . Therefore, the orifice effective area changes from 0 when
Ymax

the needle is shut to S I when the needle is fully open.

Note that the application of a sac-hole nozzle may cause inaccuracy in the results

because the model for a sac-hole nozzle is not yet implemented.

Because the needle chamber of the CRI is directly connected to the high-pressure

accumulator, the line pressure is assumed to be constant during injection. Also, it is

assumed that the temperature of the fuel remaining unchanged inside the injector.

Lastly, the injection quantity is the integral of the injection flow rate for every shot:

t
Q = ∫ qI (5-17)
0

MODEL IMPLEMENTATION IN SIMULINK

As mentioned earlier, MATLAB SIMULINK is chosen as the software program to

build the model because of its strength in dealing with nonlinear systems. Presented

in a block scheme, a SIMULINK program provides a more intuitive understanding of

the physical phenomenon than does a set of mathematical equations.

The first requirement for a computer program is stability and high speed. To achieve

these targets, a number of techniques are applied. Behind the clear program setup or

52
 appropriate organization of look-up tables, Lord Kevin’s simulation scheme is also

used. The basic principle of the scheme is presented according to [6] as follow.

The direct method of resolving a differential equation, for example:

••
Φ+ Φ = f (5-18)

is building a simulation scheme as follow:

Figure 13. Simulation scheme based on derivative blocks [6]

The significant weakness of this approach is that every signal is corrupted by noise.

When such a differentiated signal is differentiated again, the output could be out of

range and cause the simulation to be stopped.

The idea of using integrator blocks as basic method for simulating such scheme is

presented by Lord Kevin. According to the method, the differential equation

d nΦ d n −1Φ dΦ
n
= f ( n −1 ,..., , Φ , Ω, Ψ ) (5-19)
dt dt dt

53
can be solved using a scheme deputed in Figure 14:

Figure 14. Lord Kelvin’s simulation scheme [6]

This approach gives the scheme stability and flexibility, these being very essential

for such a complicated program. The integration method is set to auto, which allows

an automatic choice of an appropriated integration method from the MATLAB six

basic integration methods and a speeding up the simulation. All of the techniques are

used in the model for this CRI.

Based on the mathematical model above, the computer program has been built in

MATLAB SIMULINK for the purpose of simulation. Figure 15 shows the block

schema of the model of CRI built in SIMULINK. The Solenoid block and Hydraulic

component block are separated and a number of scopes allow observation of the

behavior of the solenoid as well as details of the hydraulic part. The advantage of this

54
 structure is that it allows observation of inputs and outputs of individual blocks,

which helps in the debugging stage of the program development. It also allows

independent evaluation and modification of individual blocks without changing the

structure of the whole model. The use of strategically placed scopes permits tracing

of specific signals and thus gives a valuable graphical insight into the internal

working of the CRI.

Figure 15. Block-diagram of the CRI model

The non-linearities and discontinuities of the system are resolved by placement of

MATLAB “Fcn” blocks. These blocks are functions written in MATLAB and

present non-linearities and discontinuities, such as that plunger hitting the upper end

or the needle reaching the spacer.

By implementing the program, different scopes, which are acting as oscilloscopes,

show the behaviors and responses of every detail of the injector as well as the

changes in the electrical and hydraulic characteristics of the system. The flexibility in

the scope placement could help in the understanding of the responses and behaviors

55
 of any targeted detail, which may be impossible to achieve by experiments. The

observed results could provide a base for further justification of a development in the

injector’s modifications.

The listing of the complete program is presented in Appendix B.

56
 CHAPTER SIX

MODELING THE UNSW HEUI INJECTOR

THE SOLENOID

The UNSW HEUI solenoid is slightly different from the CRI solenoid. The biggest

difference is in the direction of the line pressure on the solenoid plunger. In the CRI

case, the line pressure “helps” open the solenoid valve while in the UNSW HEUI

case, the line pressure resists the opening (see Figure 16).

Figure 16. Schema of the UNSW HEUI solenoid

57
 This difference causes the only change in the HEUI solenoid model (compared to

that of the CRI solenoid model). This is that the sign in the force equilibrium is

changed from plus to minus, i.e.:

d 2x dx
m1 . 2 = K f .i + m1 g − PL S1 − k1 .( x + x 0 ) − f 1 . (6-1)
dt dt

Other equations applied for the HEUI solenoid model are similar to those of the CRI

solenoid model (equations (5-4) to (5-10)).

THE HYDRAULIC COMPONENT

The hydraulic component of the HEUI is much more complicated than the CRI

because it also includes valves and a pressure amplification device. The significant

difference is that the HEUI includes a pressure amplifier, which can increase the

pressure up to 9 times. This feature ensures an extremely high injection pressure

while keeping the inlet fuel at a modest level of pressure. It helps to avoid using a

high pressure pump, which requires energy and can cause fuel leakage.

The UNSW HEUI is still in the prototype stage so the design is still being changed

slightly from model to model. In this study, the model which is taken in the

investigation has a simplified schema as presented in Figure 17.

58
 Figure 17. Simplified schema of the UNSW HEUI

The UNSW HEUI hydraulic component consists of three main elements: the

hydraulic differential valve (HDV), the intensifier and the injection needle. Each of

these has a separate input and output so it is possible to establish a mathematical

model for every element separately. This method simplifies the whole model and

makes the programming easier.

59
 This approach also allows modifying the model according to possible changes in

injector design. As mentioned earlier, the UNSW HEUI is in prototype stage so its

design is not finalized. The main target of the HEUI improvements is the intensifier

with the goal of achieving faster metering and a better injection shape. So this

approach allows changes in the intensifier model to adapt design modifications while

keeping other parts unchanged.

Like the CRI model, the UNSW HEUI hydraulic component model is a one-

dimensional, transient and compressible flow model. It is built based on well-known

fluid mechanical equations, which are presented in [34]. The model also includes a

number of non-linearities, which represent specific design features of the injector.

HYDRAULIC DIFFERENTIAL VALVE, HDV

When the control current is applied to the solenoid valve, it opens and allows the fuel

to flow from the working chamber to the poppet chamber, then further to the control

chamber and finally out to a return line. There is a restriction to this flow by the

poppet, which causes an hydraulic force to act on the HDV in the direction of the

flow holding the HDV closed with the additional assistance of the HDV spring.

This outflow reduces the pressure in the working chamber and causes the intensifier

to move upwards, which begins the metering phase. The solenoid closing causes this

flow to stop and the pressure difference between the working chamber and the inlet

port then makes the poppet valve open.

60
The movement of the poppet valve is performed in very short time, so it is not

precise if a steady-state model of a poppet valve is included. The model plays an

important role in the precision of the whole model, because the poppet opening

determines the time during which the intensifier is pressurized.

When the pressure difference between the inlet port and outlet port of the poppet

valve overcomes the spring force, the poppet valve begins to open. With an

acceptable accuracy, the flow through the valve opening can be considered to take

the form of a jet (of annular section) which follows the face of the poppet. The jet

causes a reaction force which equals the rate of change of momentum of the flowing

fuel. The inlet port chamber of the poppet valve is much larger than the opening port

volume (due to the very short movement of the poppet valve disk) so the velocity of

the fuel in the inlet port can be ignored when it is compared to the fuel velocity of the

jet.

The equivalent orifice area of the poppet valve when it opens with a movement in the

axial direction z [29] (as shown in Figure 18) is:

S P = π .z. sin α (d − z. sin α cos α ) (6-2)

where d is the seat diameter.

61
 Figure 18. Poppet valve of the HEUI

Then, the flow force of the jet can be expressed as:

q2
Fj = (6-3)
K jS j

where the flow through the poppet valve is calculated using the formula for a flow

through an orifice:

q = K j S j PL − PW (6-4)

After summarizing all the possible forces acting on the poppet valve disk, the

equation presenting the motion of the poppet valve (it is also the HDV of the HEUI)

can be written in a form as follows:

d 2z dz
m 2 . 2 = PL .S 21 + m 2 g − k 2 .( z + z 0 ) − F j − S 22 .PW − f 2 . (6-5)
dt dt

62
 where z is the movement of HDV piston, z 0 is pre-compressed amount of the

HDV spring, m 2 is the mass of the HDV, k 2 is the force coefficient of the HDV

spring, f 2 is the friction coefficient between the HDV and the liner, S 21 is the cross-

sectional area of the working chamber side of the HDV, and S 22 is the cross-

sectional area of the poppet chamber side of the HDV.

The equation describing mass conservation in the working chamber end of the HDV

is presented under the next heading, “Intensifier”.

INTENSIFIER

All equations in the reviewed literature simply illustrate how much the intensifier

could amplify pressure. They cannot be used in the calculation because the

phenomenon is much more complicated in transient conditions. The fuel is

compressed in both end compartments of the intensifier. Moreover, there are flows in

and out of the compartments, for example, the injection flow. This flow will occur

only when the fuel pressure in the needle chamber, which is directly connected to

high-pressure compartment of intensifier, is intensified up to the level which

overcomes the force of the needle spring. At the same time, the fuel continues to

flow from the pressure accumulator into the low-pressure compartment. Based on

these justifications, the equations applied for the intensifier are written as mass

conservation equations for both compartments of the intensifier – i.e. the working

and needle chambers; and the motion equation for the intensifier itself.

63
When the pressure in the working chamber has decreased to a certain level, the

intensifier begins to move up under pressure difference between the two ends. The

movement releases the non-return valve spring, which was holding the non-return

valve closed. It is assumed that the non-return valve does not cause any pressure loss

downstream; it is working entirely as an open-shut gate.

The motion equation for the intensifier can be expressed as:

d2y dy (6-6)
m3 . 2
= PN .S 32 + k 3 .( y − y 0 ) − m3 g − S 31 .PW − f 3 .
dt dt

where m 3 is the mass of the intensifier, y is the movement of the intensifier, y 0 is

the pre-compressed motion of the intensifier spring, k 3 is the spring stiffness

coefficient of the intensifier spring, S 31 and S 32 are the cross-sectional areas of the

low-pressure chamber and the high-pressure chamber of the intensifier respectively,

and f 3 is the friction coefficient between the intensifier and its cylinder.

The flow entering the working chamber through the annular control orifice of the

poppet valve causes flow out to return line through the solenoid valve and makes the

intensifier move and compress the fuel in the chamber. The equation is:

dy Vw0 dPW x z
S 31 . − . = K O .S O . . PW − PR − K j .S j . . PL − PW (6-7)
dt β dt X max Z max

64
 where V w 0 is volume of the working chamber, PW is the pressure in the working

chamber, K j is the flow coefficient of the flow into the working chamber, K O is the

flow coefficient of the flow out from the working chamber, S j and S O are the

effective areas of the inlet and outlet ports of the working chamber respectively, and

PR is the pressure in the return line.

The high-pressure compartment of the intensifier is directly connected to the needle

chamber via a hole, the diameter of which is big enough to assume that the high-

pressure compartment of the intensifier and needle chamber can be considered as a

united chamber. Thereafter, this is called the needle chamber so the mass

conservation equation for the needle chamber, which is presented in the next

heading, is also for the high-pressure compartment of the intensifier.

INJECTION NEEDLE

The next event is that the non-return valve opens and transmits the pressure from the

inlet port to the needle chamber. When the intensifier is in the metering phase, the

pressure in the needle chamber is equal to the pressure in the line accumulator (as

with an assumption that the non-return valve does not cause a pressure loss). When

the intensifier is in the pressing phase, the pressure goes up and, at the moment when

the pressure reaches a certain level, injection occurs.

The movement of the needle can be expressed by the force equilibrium as follows:

65
 d 2u du
m4 . 2
= PN .S 4 − m 4 g − k 4 .(u + u 0 ) − f 4 . (6-8)
dt dt

where m 4 is the mass of the needle, u is the movement of the needle, u 0 is the pre-

pressed motion of the needle spring, k 4 is the spring stiffness coefficient of the

needle spring, S 4 is the effective areas of the needle , and f 4 is the friction

coefficient between the needle and its cylinder.

The mass conservation equation for the needle chamber can be written as:

dy Vn 0 dPN
S 31 . − . = qU (6-9)
dt β dt

where V n 0 is the volume of the needle chamber, PN is the pressure in the needle

chamber or the injection pressure, and β is the bulk modulus of the fuel.

It is also noted that the bulk modulus is changing according to pressure, i.e.

β = f ( P ) as mentioned earlier. The bulk modulus is also dependent on temperature

but not to such a degree as with pressure, so for the calculation in needle chamber,

the temperature effect on bulk modulus is ignored.

In the above equation, qU is the injection flowrate caused by the fuel compressed in

the needle chamber. In turn, this can be presented as a flow through orifice-nozzle

holes as follows:

66
 u (6-10)
qU = K Z .S 5 . . PN − PC
U max

where K Z is the flow coefficient of the injection flow, S 5 is the total effective area

of the nozzle holes, , and U max is the maximum movement of the needle.

As stated in the CRI chapter, a VCO type of injector nozzle is used so it is assumed

u
that the injection flow at every position u of the needle is proportional to S 5 . .
U max

Beside that, the value of pressure in the cylinder is inserted manually. It is also

considered unchanged during the injection because there is insufficient data for the

cylinder pressure during the injection. For a future model, an approximate trace of

cylinder pressure may be included to provide a more precise pressure difference

between the pressure inside the needle chamber and the cylinder.

Like the CRI model, the injection amount is the integral of the injection flow rate for

the time period from t 0 to t :

Q = ∫ qI (6-11)
t0

The volume flow rate is then calculated according to fuel properties in standard

conditions (20oC, 105 Pa) to obtain the quantity of the fuel delivery in kg (mg).

67
 MODEL IMPLEMENTATION IN SIMULINK

Based on the same approach as applied for building the model for the CRI injector,

the model of HEUI is built in blocks. A simplified diagram of model is presented in

Figure 19.

Figure 19. Block-diagram of the HEUI model

68
The program consists of four main blocks: Solenoid, HDV, Intensifier and Needle.

The output of every block is connected to the Main Scope, which allows observation

of the complete system's characteristics on a single monitor.

A number of M-functions are created to implement non-linearities of the system.

These M-functions slow down the speed of simulation, so their number is kept to a

minimum. A number of programming techniques are also applied to avoid Algebraic

loops, which may also slow down the simulation.

The listing of the complete program is presented in the Appendix C.

69
 CHAPTER SEVEN

EXPERIMENTAL SETUP AND VALIDATION

CRI

For the purpose of validating the model of the CRI and for further CRI

investigations, an injector test rig has been built to directly examine the injector

characteristics. The test rig allows measurement of rail pressure, engine speed, and

injection quantity as well as the duration of control signals. The test rig also allows

photographing the form of the injection flow leaving the nozzle. The diagram of the

test engine rig is presented in Figure 20.

The main components of the test rig, which are used for the purpose of obtaining

experimental results for this study, are:

1. A swash-plate variable displacement pump, which has a PWM

(Pulse Width Modulation) pressure regulator mounted at the end of the rail. The

pump can create pressures up to 135 MPa and deliver up to 60 cubic millimeters of

fuel with each revolution of the pump.

70
 Figure 20. CRI test rig schema

2. An injection control unit, which consists of an electronic board based on a

PowerMOS transistor BUK9524-55, a DC power source and a Pulse Generator GPG-

8018G. This injection control unit can supply electric signal of 24 V with a drain

current up to 45 A with a range of time duration from 10-5 seconds to 0.5 seconds.

3. A Max 213 series Flow Meter together with a Micromann Universal Frequency

Transmitter UHZ/A. The Flow Meter is a positive displacement piston type unit

which is capable of very high accuracy over a wide range of flow rates with a K-

factor (number of pulses, which the flowmeter sends out when 1 cubic centimeter of

71
 fuel flows through it) of 115. The movement of the piston is converted to a circular

motion at the central crankshaft, which is coupled to a magnet in the flow meter.

This motion is sensed by an external electronic transmitter which converts the flow

into a pulse train. The Universal Frequency Transmitter displays the number of

pulses received and converts them into an output voltage.

4. A Tektronix Digital Oscilloscope TDS3TRG. It is four-channel oscilloscope with

500 MHz bandwidth equipped with a full VGA Color LCD. It has a plug-in Printer

for portable documentation of results and a built-in Floppy Disk drive for easy

storage and documentation. These features allow the storage of experimental results

on a PC for easy comparison with computer simulated results.

5. An injection detector is based on a SY-509 Photo-Interrupter. It is a high

reliability, high speed photo sensor, which has average rise and fall time of less than

5 µs. The sensor gives a signal in the instant when the injection spray reaches the

sensor passage, and shuts down the signal when the injection spray passed the

sensor. The signal is then captured on the oscilloscope to record the instants of the

beginning and the end of the injection.

Figure 21 shows the picture of the test rig located in the Internal Combustion Engine

laboratory in the UNSW. Figure 22 shows the detail view of the CRI and the

injection detector.

72
Figure 21. CRI test rig in the UNSW.

Figure 22. The detector location.

73
 Due to the size of the injection detector used, it is detecting the injection at the

distance of 2.8 mm from the nozzle. The speed of the fuel departing the nozzle is

estimated to be over 300 m/s, which would cause negligible time delay. A number of

additional measurements were performed for the distances of the detector from 3 to 6

mm as shown in Figure 23. The results confirm that these different gaps between the

nozzle and the detector cause no noticeable difference between the received signals.

Figure 23. The test of detector delay.

Using the test rig described in Figure 20 we have performed measurements of

injections for the solenoid control pulses of 15 VDC with the durations of 2,4,6, and

8 ms. The results of measurements together with simulated responses for 2 ms and 4

ms control signals are shown in Figure 24 and 25 respectively. These figures show a

very good match between the experimental and simulated results. The actual timing

errors between simulated and experimental results are below 3.5 % (2.4 % for 2 ms

control signal and 3.2 % for 4 ms control signal).

74
Figure 24. Comparison of timing for the 2ms control pulse.

Figure 25. Comparison of timing for the 4ms control pulse.

75
 The quantities of injections were also compared. To obtain the average value of

injection quantity from every particular shot, a number of the injector shots were

performed. For the simulated results, the temperature of the solenoid coil is set to

40oC. Table 3 presents the results received from the experimental and simulated

tests. The calculated errors between experimental and simulated injection quantities

are below 5 %, which was considered to show acceptable agreement between

experiment and simulation.

Average Simulated
measured quantity of
Duration Number Total
Pressure quantity an injection Error
of pulse of quantity
(bar) of an (mg) (%)
(ms) injections (g)
injection
(mg)
4 100 500 11.4 114 115.819 1.6

6 100 500 17.2 172 169.017 -1.7

8 100 500 22.7 227 222.402 -2.0

10 50 500 14.2 284 275.825 -2.8

Table 3. Comparison of the injection quantities

The results show that the CRI model can be, with a good degree of accuracy, used

for the purpose of further improvements of the CRI system. Such improvements

could include injector size reduction for DF injection systems or the investigation of

possible use of the injector for multipulse (pilot injection) regime.

76
 HEUI

The engine used for validation was a 6.7 litre 6-cylinder engine D7B230 donated by

the Volvo Truck Corporation (Sweden), which was used for the development of a

tuneable diesel engine injector system for engine calibration and optimisation [31].

The engine was then modified to adapt to the requirements of this project. The No 1

cylinder manifolds are separated from the remaining five cylinders to create a single

cylinder test engine. The No 1 cylinder has its own inlet air supply from a pressurized

air accumulator which can be controlled to any desired value. The remaining five

cylinders operate normally and help stabilize the system under test conditions. A 400

kW eddy current dynamometer provides the power absorption. The engine is

equipped with Accumulator Fuel System which supplies the fuel to the UNSW HEUI

in No 1 cylinder. The other five cylinders remain on the existing conventional pump-

line-nozzle fuel system. The exhaust gas path of the cylinder No 1 is also separated

from the main exhaust gas pipe. Therefore all the operating conditions of the cylinder

No 1 (intake air pressure and temperature, exhaust pressure, load and speed) can be

set individually, and all of the cylinder No 1 performance parameters (indicated mean

effective pressure, power, fuel consumption, and exhaust emissions) can also be

measured separately.

The fuel pressure in the accumulator is controlled by a PID controller. The PID

controller controls the opening of a by-pass valve from the outlet of the pressure

pump (fitted to the engine) back to the suction line. The test engine is equipped with

two PCs to control the system. The first PC is used for programming, storing data

about the duration and timing of the HEUI control pulses, and for actuating the

77
 pressure specific to particular tests. The other PC provides synchronization of the

in-cylinder pressure measurement as well as synchronization of the HEUI control

pulses with the position of the engine's crankshaft.

The HEUI injection flowrate is measured by a Max 213 series Flow Meter together

with Micromann Universal Frequency Transmitter UHZ/A and transmitted to the

computer. Tektronix Digital Oscilloscope TDS3TRG is used for storing experimental

results on the PC and for making comparisons with computer simulated results.

Figure 26 shows a simplified diagram of the engine setup. Figure 27 shows the HEUI

test engine in the laboratory. Figure 28 presents the location of fuel supply lines for

the Cylinder No 1, which was used to test the UNSW HEUI.

78
Figure 26. Simplified diagram of test engine setup for HEUI

79
Figure 27. The HEUI test engine.

Figure 28. Cylinder No 1

80
 Due to breakdown of the prototype of the UNSW HEUI we were unable to perform

full validation of the HEUI model. Partial validation of the model was based on the

results obtained from earlier experiments on the UNSW HEUI [31] as shown in

Figures 29 and 30.

Figure 29. Superimposed injection rate traces of the UNSW HEUI experimental responses with fixed

fuel delivery and varied actuating pressure (adopted from [31]).

The experimental results of the UNSW HEUI were obtained from the running test

engine. The UNSW HEUI injected fuel into the Cylinder No1, in which the pressure

81
 was changing according to crank angles. To match the simulated conditions with

the working conditions of the engine, the cylinder pressure in the model was set

based on a cylinder pressure waveform in Figure 30 (obtained from [31]).

Figure 30. The experimental cylinder pressure waveforms [31].

82
 We have simulated the performance of the HEUI for line pressure values of 10, 15

and 20 MPa.

Figure 31. The simulated responses with the line pressure of 10 MPa..

Figures 31, 32 and 33 shows the simulated injection rate curves and injection

quantities for the line pressures 10, 15 and 20 MPa respectively. All simulations

were performed with fixed injection quantities of approximately 120 mm3/stroke. It

should be noted that we have tried to match in simulation experimentally obtained

quantities of fuel delivery.

83
Figure 32. The simulated responses with the line pressure of 15 MPa..

Figure 33. The simulated responses with the line pressure of 20 MPa..

84
 The unusual forms of the simulated injection rate curves could be explained by the

use of the experimental cylinder pressure waveform.

Table 4 shows experimental and simulated results for the fuel delivery quantities and

Table 5 shows the results for the duration of the injections.

Trace Line pressure Experimental fuel Simulated fuel Error

No (MPa) delivery (mm3/stroke) delivery (mm3/stroke) (%)

1 10 122.2 112 -8.3

2 15 121.9 132 8.2

3 20 119.8 129 7.7

Table 4. Comparison of fuel delivery

Trace Line pressure Experimental duration Simulated duration

Error
No (MPa) of injection (ms) of injection (ms)

1 10 1.8 1.57 -12.7

2 15 1.43 1.36 -4.9

3 20 1.29 1.35 4.6

Table 5. Comparison of duration of injection

The largest errors of –12.7 % and –8.3 % are observed for the trace No 1

corresponding to the line pressure of 10 MPa. While the magnitude of the errors is

large, the shorter simulated duration is, with high probability, linked to the smaller

quantity of the delivered fuel. Similar argument can be applied to the trace No 3

where increased duration of simulated injection corresponds to increased simulated

85
 fuel delivery. The largest discrepancy can be observed for the trace No 2 where

shorter simulated duration of injection is linked with larger fuel delivery. Taking into

account the magnitudes of errors are relatively small (8.2 % and –4.9 %) it is

considered that the simulated results show acceptable agreement of the UNSW HEUI

and its model.

Improved accuracy of the simulated results will require further experimental results

of the UNSW HEUI. This may require modification of some of the parameters of the

model. It is considered useful to employ the partially validated UNSW HEUI model

as a tool for investigation of the internal processes of the injector.

86
 CHAPTER EIGHT

SIMULATION

The two computer models of the CRI and the UNSW HEUI were then investigated in

details. The CRI model was examined from the point of view of its ability to work in

the pilot injection regime. The UNSW HEUI was analyzed from the point of view of

possible change in the solenoid design with the aim of reducing the metering period.

Sensitivity analysis was performed for both models with the aim to find critical

parameters of the models affecting the time delay duration and the amount of fuel

injection.

CRI

INVESTIGATION OF MULTIPULSE REGIME

The investigations of the model were performed to observe the system behaviour.

This is an investigation of the injector's ability to perform pilot injections, which

represent a very important feature of a Common Rail Injector and are indicative of

the speed and accuracy the injector can perform its function.

87
Figure 34 shows the injection rate curve developed from a 5 ms control signal.

Figure 34. CRI responses (5ms control signal)

88
 Figure 35 presents simulated injection rate curves of the CRI when the main control

signal of 5 ms together with the pilot control signal of 1 ms are applied. The interval

between the pilot control signal and the main control signal is also set to 1 ms.

Figure 35. Pilot injection (control time interval of 1 ms)

The responses shown on the simulated Scope indicate that the time interval between

the pilot and the main injection rate curves (i.e. between the pilot and main

injections) is approximately 1.9 ms. This time interval is caused by the time

constants representing activation and deactivation of the solenoid and is influenced

by the pressure in the working chamber of the injector.

89
The time interval between the control signals was then reduced with the step of 0.1

ms to examine what is the minimum time interval between the pilot and main

injections. When the interval between the two control signals is reduced to 0.9 ms,

Figure 36 illustrates that the pilot and main injections are now separated with a time

interval of approximately 1.88 ms, practically no difference from the previous

simulation.

Figure 36. Pilot injection (control time interval of 0.9 ms)

90
 The control time interval was then set to 0.8 ms. Figure 37 shows that the pilot

injection is "dissolved" into the main injection. It is believed that this is a

consequence of the overlap between the closing of the solenoid valve after the pilot

injection and the opening of the solenoid valve for the main injection.

Figure 37. Pilot injection (control time interval of 0.8 ms)

91
 Figure 38 shows that for the control time interval of 0.7 ms, the pilot and the main

injections are totally dissolved. For the shorter time interval (i.e. less than 0.7 ms),

similar responses were obtained.

The simulated responses reveal that the shortest time interval between the pilot and

the main injections (or between any two consecutive injections) is approximately

equal to the sum of the solenoid delays.

Figure 38. Pilot injection (control time interval of 0.7 ms)

92
 The model investigation shows that the shortest possible time interval between the

pilot injection of 1ms and the main injection is approximately 1.9 ms for the working

pressure of 50 MPa. If the injection system requires a shorter time interval then a

faster solenoid will be needed. Increased speed of solenoid, while possible, requires

much more power. This may lead to an undesired increase in the size of the injector

which acts against the requirement to fit it to a standard (unmodified) diesel engine.

If shorter pilot injections are required, the possible alternative would be a

piezoelectric actuator which is known for having far better efficiency and much

faster actuation [20].

SENSITIVITY ANALYSIS OF THE MODEL

We are interested in the investigation of the sensitivity of major output parameters of

the CRI injector, which are:

• Injection delay

• Injection duration

• Injection quantity

Using the developed model we have performed sensitivity analysis of these

parameters for following input parameters:

• Force coefficient of the spring of the solenoid ( k1 ).

93
 • Force coefficient of the spring of the needle ( k 2 ).

• Diameter of the piston ( d p ). The diameter here is understood as a complex

parameter affecting the values of other dependent parameters. These include

the cross-sectional area of the piston, the mass of the piston, and the volume

of the working chamber.

• Temperature of the coil of the solenoid ( Ts ).

The results of the sensitivity analysis are shown in Table 6.

k1 k2 dp Ts

Injection
-0.173 0.085 0.69 1.6
delay

Injection
-0.049 -0.051 -0.05 -0.512
duration

Injection
-0.061 -0.059 -0.068 -0.638
quantity

Table 6. Sensitivity analysis of the CRI model

The results show that the system is relatively insensitive to changes in the force

coefficients of the solenoid and needle springs, but it is very sensitive to the solenoid

94
 temperature and the possible design change in the "overall" diameter of the piston.

According the model investigation, the change in the diameter of the piston, which

consequently causes changes in the related parameters, may significantly affect the

injection delay. The positive sensitivity of 0.69 points out that a possible dimension

minimization may reduce the injection delay. The revealed significant sensitivity of

the injection delay and the injection quantity to the temperature of the solenoid coil

may help the future design of a controller for the DF injection system. The sensitivity

analysis shows that the temperature must be taken into calculations, especially in the

case when the CRI is used as an ignition source for NG.

95
 HEUI

INVESTIGATION OF THE FORCES OF THE SOLENOID

We have used the developed model of the UNSW HEUI to investigate the responses

of the HEUI injector with a focus on the solenoid delay. As mentioned earlier for the

investigated UNSW HEUI, the direction of the pressure force acting on the solenoid

plunger opposes the solenoid opening action. A possible change in the direction of

the pressure force is examined by changing the sign of the force in the solenoid

model. In this arrangement (used in some CRI injectors) the pressure force will

“help” to open the solenoid valve when the solenoid is activated and it will “oppose”

the solenoid valve closing when the solenoid coil is deactivated. The results in Figure

39 and Figure 40 show that the possible solenoid modification may lead to the

following effects:

• The metering phase may begin earlier by 0.26 ms.

• The injection quantity may be higher by 27 % for the same control signal.

• The closing action of the solenoid may be delayed by 0.45 ms.

The benefit of this arrangement is that the injector can deliver required fuel delivery

in shorter time despite the slight delay in solenoid closing.

96
Figure 39. Direction of the force of the pressure is against the solenoid opening

97
 Figure 40. Direction of the force of the pressure is "helping" the solenoid opening

Note that, while the HEUI has the potential for pilot injection and leading edge pulse

shaping, this has not been implemented yet in the models available here.

98
 SENSITIVITY ANALYSIS OF THE MODEL

Similar to the CRI case we have performed sensitivity analysis of the UNSW HEUI.

We used the same major output parameters as for the CRI, i.e.:

• Injection delay

• Injection duration

• Injection quantity

Using the developed UNSW HEUI model we have performed the sensitivity analysis

of the above output parameters to the following input parameters:

• Force coefficient of the spring of the HDV ( k 2 ).

• Force coefficient of the spring of the needle ( k 4 ).

• Diameter of the intensifier ( d i ). The diameter of the intensifier here is also

understood as a complex parameter affecting the values of other dependent

parameters. These include the cross-sectional areas of both ends of the

intensifier, the mass of the intensifier, and the volumes of the working and

the needle chamber.

• Temperature of the coil of the solenoid ( Ts ).

99
 k2 k4 di Ts

Injection delay 4.49e-5 0.088 0.04 -3.575

Injection
-6.556e-4 -0.416 0.497 -0.435
duration

Injection
4.55e-5 -1.59 5.68 -3.510
quantity

Table 7. Sensitivity analysis of the UNSW HEUI model

The investigations show a negligible effect of the stiffness of the HDV spring ( k 2 ).

As indicated in Table 7, the variations in the force coefficient of the spring of the

needle ( k 4 ) may cause significant changes in the duration and quantity of the

injections. The phenomena could be explained by the fact that, compared to the CRI,

the injection pressure in the HEUI is significantly higher and the high pressure

volume is much smaller. In the CRI case, the pressurized injection fuel is fed directly

from much bigger volume – the system accumulator.

The injection quantity is very sensitive to the diameter of the intensifier ( d i ). This

means that the possible intensifier size reduction (e.g. in the case when the HEUI

needs to be combined with a gas injector to form a DF injector) and consequently,

the whole injector minimization will significantly affect the injection duration and

quantity. It indicates that the minimization may lead to the reduction of the injection

100
 quantity, so that the intensification value M [42] must be changed to keep the

injection quantity unchanged.

The other important result of the investigation is that the temperature of the coil of

the solenoid ( Ts ) has also a considerable effect on the injector behaviour. The reason

is that the HEUI solenoid has relatively high inductance (2.3 mH) and relatively high

resistance (2.8 Ohm). Increase in the temperature of the coil can cause significant

change in the current of the solenoid. This may significantly affect the metering

period of the injector and consequently, the injection delay and injection quantity.

Possible solutions of this problem may include an application of a current driven

solenoid or the faster type of the actuating device – the piezoelectric actuator.

101
 CHAPTER NINE

CONCLUSIONS

The study presents a one-dimensional, transient and compressible flow models of a

Common Rail Injector and UNSW Hydraulically actuated Electrically controlled

Unit Injector. Based on electrical, magnetic and fluid hydraulic equations, the models

represent complete dynamic models for both these high-technology diesel injectors.

The significant features of the models, compared to existing computer models, are as

follows:

• The model of a solenoid is present as a submodel in both the CRI and the

UNSW HEUI models. Because these injectors are electrically controlled, the

used electromechanical device (solenoid in our case) connects the electrical,

mechanical and hydraulic parts of the injectors. This allows to include in

models all major electrical, mechanical and hydraulic parameters affecting

behaviour of the injectors.

• The HEUI model is the first computer model for the UNSW HEUI. As

mentioned earlier, the UNSW HEUI is at present the most efficient and the

only single-fuel HEUI in the world. The structured nature of the developed

model of the HEUI allows investigation of possible design changes of the

102
 HEUI. It is relatively easy to add new design features to the model

developed.

The developed models were used to perform sensitivity analysis of both injectors.

The sensitivity analysis has revealed that the temperature of the solenoid coil is one

of the critical parameters affecting the timing and the quantity of the fuel injection of

both injectors. Dimension of the piston of the CRI was also found to be critical. For

the UNSW HEUI, additional critical parameters are the stiffness of the needle spring,

and the dimensions of the intensifier. The models also revealed that in the case of

pilot injections the speed of the solenoid is the major limiting factor of the

performance.

The developed models provide better understanding of the issues and limitations of

the injectors. They give detailed insight into their working principles. A number of

injector parameters, which are especially difficult to measure in practice, can be

observed during simulation. The investigation of the model of the UNSW HEUI

permits making quantitative analysis of the timing of the solenoid and to evaluate the

proposed change of the direction of the pressure acting on the solenoid plunger.

Due to the time limitation and the availability of equipments, this study presents the

models of the CRI and the UNSW HEUI with an acceptable accuracy. For the

purpose of further improvement, a number of additional instruments maybe needed

for more detailed evaluation of the models of the injectors. For example, a needle

motion sensor [31] is needed to obtain more accurate timing on a running engine and

103
 for building a model of a sac-hole nozzle type. A more detailed model for the

solenoid may also be required.

MODEL CONVERTION TO C PROGRAM

The models running in MATLAB/SIMULINK require the time from several seconds

to several minutes to simulate an injection process of some milliseconds. The

simulation speed depends on the size of the program and the speed of the computer,

on which the program is running. A task for future work is to convert the finished

models into C programs, which allow the simulation to be run in real time. The C

models may help in performing data acquisition, in designing and testing a

microprocessor controller for the diesel injection systems.

104
REFERENCE

1. Agarwal A., Assanis D. Multi-dimensional modeling of ignition, combustion

and nitric oxide formation in direct injection natural gas engines. SAE Paper

2000-01-1839, 2000.

2. Ahlin K. Modelling of pressure waves in the Common Rail Diesel Injection

Systems. University of Linkoping, Sweden. December 11, 2000.

3. Arcomanis C., and Fairbrother R.J. Computer Simulation of Fuel Injection

System for DI Diesel Engines. SAE Paper 922223, 1992.

4. Arcomanis C., Gavaises M., Abdul-Wahab E., and Moser V. Modeling of

Advanced High-Pressure Fuel Injection Systems for Passenger Car Diesel

Engines. SAE Paper 1999-01-0910, 1999.

5. Brater E., King H., Lindell J., Wei C. Handbook of hydraulics. Seventh

Edition. McGraw-Hill. 1996.

6. Cavallo A., Setola R., Vasca F. Using MATLAB, SIMULINK and Control

System Toolbox. Prentice Hall, 1996.

7. Cheung N.C. A nonlinear, short stroke proportional solenoid. PhD thesis.

UNSW, 1995.

105
 8. Cheung N.C., Rahman M.F., Lim K.W. Simulation and experimental studies

towards the development of a proportional solenoid. AUPEC’93, 1993.

9. Choi I. S., Milton B. E. Dual-fuel combustion with injected natural gas and

distillate. CHT97, Advances in Computational Heat Transfer, ed. de Vahl

Davis and Leonarde, Begel House, pp166-175, NY, 1998.

10. Douville B., Ouellette P., Touchette A., Ursu B. Performance and emission of

a two-stroke engine fueled using high-pressure direct injection of natural gas.

SAE Paper 981160, 1998.

11. Dumitrescu S., G.Hill P., Ouellette P., Guowei L. Effects of injection changes

on efficiency and emission of a diesel engine fueled by direct injection of

natural gas. SAE Paper 2000-01-1805, 2000.

12. Edwards S.P., Pillay A.D., Michon S. and Fournier G. The optimisation of

common rail FIE equipped engines through the use of statistical experimental

design, mathematical modeling and genetic algorithm. SAE Paper 970346,

1997.

13. Favennec A., Lebrun M. Models for injector nozzles. Proceedings of the

Sixth Scandinavian International on fluid Power, SICF'99. Tampere, Finland,

May 26-28, 1999.

106
 14. Ficarella A., Laforgia D. Evaluation of Instability Phenomena in a Common

Rail Injector System for High Speed Diesel Engines. SAE Paper 1999-01-

0192, 1999.

15. From the Internet. http://www.amesim.com/de/pdf/fuel_113.pdf. AMESim

and Diesel Fuel Injection Systems-Technical Bulletin No 113, France, 2000.

16. From the Internet. http://www.wartsila.com/english/pdf/en_commonrail.pdf.

Emission control. The common rail system. Finland, 2000.

17. From the Internet. http://www.caterpillar.com/shared/parts_n_service/

03_engine_ parts/ 03_spec_sheet_library/pdf/pehp9526.pdf PEHP9526-

HEUI Fuel Systems-Caterpillar. USA, 1999.

18. From the Internet. http://www.globaldenso.com/TECHNOLOGY/tec-report/

2000/ pdf/denso2k.pdf. Denso technology 2000. Japan, 2000.

19. From the Internet. http://www.chinamotorcycle.com/digest/gygg.html.

Analysis the mainly components characteristics of common rail fuel injection system. Wuhan,

China, 2000.

20. From the Internet. http://www.osti.gov/hvt/deer2000/hakimpap.pdf. Smart

Materials for Fuel Injection Actuation. Detroit Diesel, DEER workshop,

2000.

107
 21. Gavaises M., Yamanishi M., Yamada S., Nakahira T. Modeling the effect of

an in-line Pump-Based Fuel Injection System Characteristics for Small

Industrial DI Diesel Engines on Spray Development. Seoul 2000 FISITA

World Automotive Congress. Seoul, Korea, June 12-15, 2000.

22. Gebert K., Beck N., Barkhimer R.L., Wong H.C. Strategies to improve

combustion and emission characteristics of dual-fuel pilot ignited natural gas

engines. SAE Paper 971712, 1997.

23. Geiger J., Grigo M., Lang O., Wolters P. Direct Injection Gasoline Engines-

Combustion and design. SAE Paper 1999-01-0170, 1999.

24. Karim G. A., Jones W., and Raine R.R. An Examination of the Ignition Delay

Period in Dual Fuel Engines. SAE Paper 892140, 1989.

25. Kohketsu S., Tanabe K., Mori K. Flexibly Controlled Injection Rate Shape

with Next Generation Common Rail System. SAE Paper 2000-01-0705, 2000.

26. Kuo B. C. Automatic control systems. Sixth edition. Prentice-Hall

International. 1991.

27. Liu Shenghua, Hwang J.W., Kim M.H., Chae J.O., Taeyong Chung.

Simulating the Effects of Fuel Injection Characteristics on DI Diesel Engine

Emissions. Seoul 2000 FISITA World Automotive Congress. Seoul, Korea.

June 12-15, 2000.

108
28. Liu Z. and Karim G.A. The ignition Delay Period in Dual Fuel Engines. SAE

Paper 950466, 1995.

29. McCloy D., Martin H. Control of fluid power. Revised edition. Ellis Horwood

Ltd. 1980.

30. Meyer R., Cole J., Klenzle E., Wells A. Development of a CNG engine. SAE

Paper 910881, 1991.

31. Milton B., Casey R., Behnia M., Yudanov S., and Mitchell W. Development

of a tuneable diesel engine injection system for engine calibration and

optimisation. ARC collaborative research project report, 1998.

32. Nielsen O.B., Qvale B, and Sorenson S. Ignition Delay in the Dual Fuel

Engine. SAE Paper 870589, 1987.

33. Ouellette P., Guowei L., G.Hill P., Dumitrescu S. Optimisation study of pilot-

ignited natural gas direct-injection in diesel engines. SAE Paper 1999-01-

3556, 1999.

34. Roberson J., Crowe C. Engineering fluid mechanics. Sixth edition. John

Wiley & sons, Inc. 1997.

35. Schechter M. Fast Response Multipole Solenoids. SAE paper 820203, 1982.

109
36. Schommers J., Duvinage F., Stotz M., Peters A., Ellwanger S., Koyanagi K.

and Gildein H. Potential of common-rail injection system for passenger car

DI diesel engines. SAE Paper 2000-01-0944, 2000.

37. Talus P., Traver M., Atkinson R., Clark N., Atkinson C. Operation of a

compression ignition engine with a HEUI injection system on natural gas

with diesel pilot injection. SAE Paper 1999-01-3522, 1999.

38. Tran X.T., Milton B, White T., and Tordon M. Modelling HEUI injector in

MATLAB SIMULINK. Proceedings of the 2003 IEEE/ASME International

Conference on Advanced Intelligent Mechatronics (AIM 2003), pp 383-388.

Japan, 2003.

39. White T., Milton B., Behnia M. Mixing of Co-axial Natural Gas / Liquid

Diesel Fuel Sprays. Proceedings of Fourth International Colloquium “Fuel

2003”, pp 307-309, Germany, 2003.

40. Woermann R. J., Theuerkanf H.J. and Heinrich A. A Real-Time Model of a

Common Rail Diesel Engine. SAE Paper 1999-01-0862, 1999.

41. Yang M., Lu Q., Li J, Lu X., Du C. Study of the Injection Control Valve in a

New Electronic Diesel Fuel System. SAE Paper 980813, 1998.

110
42. Yudanov S.V. Development of the Hydraulically Actuated Electronically

Controlled Unit Injector for Diesel Engines. SAE Paper 952057, 1995.

43. Zhang X., Liu J., Wang Q., Hu Z. A study of natural gas fueling of

locomotive engines. SAE Paper 981396, 1998.

111
APPENDIX A

The 2003 IEEE /ASME conference paper

A
A-1
2
A-2
3
A-3
4
A-4
5
A-5
6
A-6
7
APPENDIX B

CRI model program

B
1
B-1
2
B-2
3
B-3
4
B-4
5
B-5
6
 M FUNCTIONS FOR THE CRI MODEL

function F = CRISolenoidForce(u)

global m
i=u(1);
Max_x=u(2);
x=u(3);
if (i<=0)
F=0;
elseif (i>=0 & i<=1)
xx=Max_x-x;
F=(772.5*xx*xx*xx*xx-945.67*xx*xx*xx+403.02*xx*xx-
69.905*xx+4.1427)*i;
elseif (i>1 & i<=2)
xx=Max_x-x;
F=(1894.1*xx*xx*xx*xx-2346.1*xx*xx*xx+1014*xx*xx-
178.65*xx+10.866)*i/2;
elseif (i>2 & i<=3)
xx=Max_x-x;
F=(2396.2*xx*xx*xx*xx-2960*xx*xx*xx+1286.6*xx*xx-
234.99*xx+17.279)*i/3;
elseif (i>3 & i<=4)
xx=Max_x-x;
F=(2338.8*xx*xx*xx*xx-3011.9*xx*xx*xx+1420.9*xx*xx-
302.12*xx+29.075)*i/4;
elseif (i>4 & i<=5)
xx=Max_x-x;
F=(2407.2*xx*xx*xx*xx-3123.4*xx*xx*xx+1522.8*xx*xx-
302.12*xx+42.14)*i/5;
else
xx=Max_x-x;
F=(2407.2*xx*xx*xx*xx-3123.4*xx*xx*xx+1522.8*xx*xx-
302.12*xx+42.14)*i/5;
end
F=F;

function Kb = Kb_function(u)

global m
deri_i=u(1);
F=u(2);
if (deri_i==0)
K=0;
else
K=F/deri_i;

B-6
7
end
Kb=K;

function FF = Force_sum(u)

global m
x=u(1);
x_max=u(2);
Fp=u(3);
Fn=u(4);
if (Fp>=Fn & x==x_max)
Fn=Fp;
FF=0;
elseif (Fp<=Fn & x==0)
Fn=Fp;
FF=0;
else
FF=Fp-Fn;
end
FF=FF;

function F = CRIforce(u)

global m
Fp=u(1);
Fn=u(2);
y=u(3);
y_max=u(4);
if (Fp>=Fn & y==y_max)
Fn=Fp;
FF=0;
elseif (Fp<=Fn & y==0)
Fn=Fp;
FF=0;
else
FF=Fp-Fn;
end
F=FF;

function C = CRIcond(u)

global m
y_max=u(1);
force=u(2);
y=u(3);

B-7
8
 if (force==0)
C=1;
elseif (y==y_max)
C=1;
else
C=0;
end
C=C;

function VV = CRINeedle_velocity(u)

global m
v=u(1);
y=u(2);
y_max=u(3);
force=u(4);
t=u(5);
if (force==0 & y==y_max)
vv=0;
elseif (force==0 & y==0)
vv=0;
elseif (t==1 & v>=0)
vv=0;
else
vv=v;
end
VV=vv;

B-8
9
APPENDIX C

HEUI model program

C
1
C-1
2
C-2
3
C-3
4
C-4
5
C-5
6
C-6
7
C-7
8
 M FUNCTIONS FOR THE UNSW HEUI MODEL

function F = HEUISolenoidForce(u)

global m
i=u(1);
Max_x=u(2);
x=u(3);
if (i<=0)
F=0;
elseif (i>=0 & i<=1)
xx=Max_x-x;
F=(-83.333*xx*xx*xx*xx-7.4074*xx*xx*xx+86.389*xx*xx-
60.958*xx+25.008)*i;
elseif (i>1 & i<=2)
xx=Max_x-x;
F=(-125*xx*xx*xx*xx+13.889*xx*xx*xx+97.083*xx*xx-
68.437*xx+33.512)*i/2;
elseif (i>2 & i<=3)
xx=Max_x-x;
F=(-62.5*xx*xx*xx*xx-92.13*xx*xx*xx+161.6*xx*xx-
85.009*xx+44.014)*i/3;
elseif (i>3 & i<=4)
xx=Max_x-x;
F=(-812.5*xx*xx*xx*xx+726.39*xx*xx*xx-109.79*xx*xx-
64.258*xx+56.03)*i/4;
elseif (i>4 & i<=5)
xx=Max_x-x;
F=(-604.17*xx*xx*xx*xx+471.76*xx*xx*xx-2.1528*xx*xx-
82.446*xx+66.026)*i/5;
else
xx=Max_x-x;
F=(-604.17*xx*xx*xx*xx+471.76*xx*xx*xx-2.1528*xx*xx-
82.446*xx+66.026)*i/5;
end
F=F;

function Kb = Kb_function(u)

global m
deri_i=u(1);
F=u(2);
if (deri_i==0)
K=0;
else
K=F/deri_i;
end

C-8
9
Kb=K;

function FF = Force_sum(u)

global m
x=u(1);
x_max=u(2);
Fp=u(3);
Fn=u(4);
if (Fp>=Fn & x==x_max)
Fn=Fp;
FF=0;
elseif (Fp<=Fn & x==0)
Fn=Fp;
FF=0;
else
FF=Fp-Fn;
end
FF=FF;

function Fp = HEUIForce_poppet(u)

global m

FdownZ=u(1);
FupZ=u(2);
z_max=u(3);
z=u(4);
x=u(5);
if (FdownZ<=FupZ & z==0)
ForceZ=0;
elseif (FdownZ>=FupZ & z==z_max)
ForceZ=0;
elseif (FdownZ<=FupZ & z==z_max)
ForceZ=FdownZ-FupZ;
elseif (x>0)
ForceZ=0;
else
ForceZ=FdownZ-FupZ;
end
Fp=ForceZ;

function C = HEUI_HDVcondition(u)

C-9
10
 global m
y_max=u(1);
force=u(2);
y=u(3);
if (force==0)
C=1;
elseif (y==y_max)
C=1;
else
C=0;
end
C=C;

function V = HEUIVelo_poppet(u)

global m
v=u(1);
force=u(2);
z=u(3);
zmax=u(4);
if (force==0 & z==0)
v=0;
elseif (force==0 & z==zmax)
v=0;
else
v=v;
end
V=v;

function O = HEUI_HDV_openness(u)

global m
x=u(1);
z=u(2);
if (x>0)
O=x+z;
else
O=x+z;
end
O=O;

function P_N = HEUI_Pneedle_chamber(u)

global m
Pline=u(1);

C-10
11
 P=u(2);
velo=u(3);
if (velo>0)
P=Pline;
else
P=P;
end
P_N=P;

function F = HEUIForce_W(u)

global m

Fdown=u(1);
Fup=u(2);
y_max=u(3);
y=u(4);
if (Fdown>=Fup & y==0)
Force=0;
else
Force=Fup-Fdown;
end
F=Force;

function V = HEUIVelo_W(u)

global m
v=u(1);
ymax=u(2);
force=u(3);
x=u(4);
y=u(5);
if (force==0 & x==0)
v=0;
elseif (force==0 & y==0)
v=0;
else
v=v;
end
V=[v];

function V_YT = HEUIVelo_condition(u)

global m

C-11
12
 v=u(1);
if (v>=0)
V_YT=0;
else
V_YT=v;
end
V_YT=V_YT;

function F_t = HEUIForce_needle(u)

global m

Froom=u(1);
Fneg=u(2);
t=u(3);
tmax=u(4);
if (Froom<=Fneg & t==0)
ForceT=0;
else
ForceT=Froom-Fneg;
end
F_t=ForceT;

function C = HEUIneedle_cond(u)

global m
force=u(1);
if (force==0)
C=1;
else
C=0;
end
C=C;

function V_T = HEUIVelo_needle(u)

global m
v=u(1);
tmax=u(2);
t=u(3);
Force_T=u(4);
if (Force_T==0 & t==0)
V_T=0;
elseif (Force_T==0 & t==tmax)

C-12
13
 V_T=0;
else
V_T=v;
end
V_T=V_T;

C-13
14