Heavy-Duty Spark-Ignited Single Cylinder Engine Fueling System Rohan Sharad Kittur

(1)

Heavy-Duty Spark-Ignited Single Cylinder Engine Fueling System

Rohan Sharad Kittur

Master of Science Thesis MMK TRITA-ITM-EX 2018:512 KTH Industrial Engineering and Management

Machine Design SE-100 44 STOCKHOLM

(2)

(3)

i

Examensarbete MMK TRITA-ITM-EX 2018:512 Bränslesystem för encylindrig motor

Rohan Sharad Kittur

Godkänt

2018-06-26

Examinator

Dr. Andreas Cronhjort

Handledare

Dr. Ola Stenlåås

Uppdragsgivare

Scania CV AB, Södertälje

Kontaktperson

Dr. Ola Stenlåås

Sammanfattning

Forskning inom motorutveckling bedrivs för att möta kommande emissionskrav och samtidigt minska bränsleförbrukningen. Kommande förbud mot dieseldrivna fordon planeras i flera städer runt om i världen. Alternativa bränsle som exempelvis naturgas ses som en lovande ersättning även för tunga fordon. Metan som är huvudkomponenten av naturgas har en fördelaktigt förhållande mellan väte och kol vilket gör den attraktiv för CO2-reducering. Hur som helst, bränslets låga cetantal och den höga aktiveringsenergin som krävs för att tända naturgas förutsätter tändstiftsantändning.

En fördel av att använda en encylindrig motor inom forskning är möjligheten att studera fenomen utan negativa gasväxlingsinteraktioner från intilliggande cylindrar. Jämfört med en fullmotor möjliggörs även ett snabbare utbyte av motordelar samt lägre bränsleförbrukning.

Fokus för detta examensarbete var genomförandet av ett flexibelt bränslesystem för en tändstiftsantänd encylindrig motor. Motorn är en tändstiftsantänd Scania 9 liters som modifieras för encylinder körning. Flexibilitet som t.ex. laddningshomogenitet, selektiv fyllning av inloppsporter och förberedelser för direktinsprutning av flytande bränsle realiserades. För enkel användning är motorn styrd av en eftermarknadsmotorstyrenhet som använder ett användarvänligt grafiskt gränssnitt för ändring av driftsparametrar. Säkerhetshänsyn vid blandning av gasformiga bränsle och luft långt innan inloppsporterna har implementerats.

Nyckelord: bränslesystem, tunga fordon, tändstiftsantändning, encylinder motor, eftermarknadsmotorstyrenhet, bränsle-luft blandningssäkerhet

(4)

ii

(5)

iii

Master of Science Thesis MMK TRITA-ITM-EX 2018:512

Heavy-Duty Spark-Ignited Single Cylinder Engine Fueling System

Rohan Sharad Kittur

Approved

2018-06-26

Examiner

Dr. Andreas Cronhjort

Supervisor

Dr. Ola Stenlåås

Commissioner

Scania CV AB, Södertälje

Contact person

Dr. Ola Stenlåås

Abstract

Most of the fundamental research in internal combustion engines is driven by the ever-increasing stringency of emissions regulations along with the need for increased fuel economy. The proposed ban on diesel vehicles in several cities around the world combined with extensive availability, has made natural gas a promising substitute even for heavy-duty applications. The high hydrogen-to-carbon ratio of methane, the major component of natural gas, makes it attractive from an emissions reduction perspective. CO2 emissions from natural gas combustion are particularly low. However, the low cetane number and high activation energy required to ignite natural gas, requires spark-ignition.

In a research setting, it is often advantageous to have a single cylinder engine. The main benefit is the ability to study phenomena without adverse interactions which multi-cylinder operation may cause. This is especially important for gas-exchange studies. Quicker replacement of parts and lower fuel consumption are secondary benefits.

The focus of this thesis was the implementation of a flexible fueling system for a single cylinder spark-ignited engine. The engine is a Scania 9-liter spark-ignited engine modified for single cylinder operation. Flexibility in terms of charge homogeneity, selective intake port filling and provisions for liquid fuel direct injection have been provided. For ease of use, the engine is controlled by an aftermarket engine control unit with a graphical user interface for configuration.

Safety considerations when mixing gaseous fuels and air well upstream of the intake ports have been implemented.

Keywords: Fueling system, heavy-duty, spark-ignited, single cylinder engine, aftermarket engine control unit, fuel-air mixture safety.

(6)

iv

(7)

v

FOREWORD

I would like to express my gratitude to all the people who have offered their time and support to help me see this project to the end.

Foremost, I thank Dr. Ola Stenlåås (Scania CV AB, Södertälje) for motivating me to step outside my comfort zone so that I could learn something. I also thank him for helping improve my technical writing skills. My thanks to Dr. Andreas Cronhjort as well for being a constant source of support.

It has been a pleasure working with Richard Adolfsson (NMEG). I am extremely grateful to him for sharing his vast practical knowledge of engines. For being accommodating of my lack of experience and patiently teaching me practical techniques to gain the required experience, I thank him very much.

I would like to thank the entire Otto engine pre-development and optimization teams for accepting me as their colleague and for tirelessly answering each of my questions. My particular gratitude goes to Arne Burger and Mats Janzon (NMEG) for their kind nature, encouragement and enlightening discussions over lunch.

I extend my special thanks to Tomas Karlberg (NLRD) for patiently answering all my questions related to the single cylinder test cells. Pål Axelsson and Björn Lind (KETE) have been extremely helpful with the flame arrestor for which I thank them. I also thank Ludvig Adlercreutz (AVL, Södertälje) for his support.

My heartfelt thanks go to Philip Thörhammar (KETL), Peter Olsson (KETL) and Timo Lohilahti (NLWM) for sharing their practical knowledge and tools. I am also grateful to all the other workshop personnel for their patience and assistance.

My fellow students of the IC engines program at KTH deserve special mention as well, for all the fun times we had over the course of two years. Particular thanks to Andreas Lius for allowing me to learn more about engines by assisting him with his thesis.

Lastly, I thank my parents for believing in me and being my pillars during all of life’s highs and lows.

Rohan Sharad Kittur Stockholm, June 2018

(8)

vi

(9)

vii

NOMENCLATURE

The notations and abbreviations used in this report are as follows.

Notations

Symbol Description

bar(a) Absolute pressure bar(g) Gauge pressure

Abbreviations

BDC Bottom Dead Center

CFD Computational Fluid Dynamics

CI Compression-ignition/ignited

CNG Compressed Natural Gas

ECU Electronic Engine Control Unit FAME Fatty Acid Methyl Ester

HVO Hydrogenated Vegetable Oil

Hythane Hydrogen-methane blends (% composition may vary)

ICE Internal Combustion Engine

LNG Liquefied Natural Gas

PWM Pulse Width Modulation

RON Research Octane Number

SI Spark-ignition/ignited

TDC Top Dead Center

(10)

viii

(11)

ix

LIST OF FIGURES

Figure 1. Top view of the engine illustrating reasons for choosing to retain cylinder 1 as the

active cylinder ... 8

Figure 2. Laboratory gas supply schematic ... 11

Figure 3. Gas panel for CNG engines ... 13

Figure 4. Illustration of the recommended gaseous fueling system layout ... 14

Figure 5. (a) Proposed orientation of the high-pressure fuel rail, needing 45 mm spacers between the manifold and cylinder head; (b) Proposed orientation of the high-pressure fuel rail demonstrating the use of the existing high-pressure connector without changing any bend radii ... 17

Figure 6. Typical solenoid current build-up, injector valve lift and corresponding fuel flow rate for (a) voltage-controlled injector and (b) current-controlled injector. ... 21

Figure 7. Lambda window for highest TWC conversion efficiency ... 23

Figure 8. Typical narrow-band O2 sensor output voltage as a function of lambda ... 24

Figure 9. (a) Construction and (b) characteristic curve of a typical wideband lambda sensor. 25 Figure 10. Illustration of intake damping tank (a) Alternative 1 with baffle; (b) Alternative 2 with larger volume ... 28

Figure 11. Illustration of the method used to adapt the intake manifold for single cylinder operation ... 30

Figure 12. Illustration of alternative 2 for variable intake port filling; (a) flow directed towards swirl port and (b) flow directed towards filling port ... 31

Figure 13. Illustration of the exhaust side damping tank with rupture disc for overpressure protection. ... 33

Figure 14. Crimped metal-ribbon flame arresting element ... 35

Figure 15. Crimped metal-ribbon flame arrestor (a) element and (b) assembly with housing 37 Figure 16. Schematic drawing of the rupture disc with holder and installation flanges ... 38

Figure 17. Intake side damping/mixing tank with overpressure rupture disc installed ... 39

(16)

xiv

Figure A-1. Spacer between the intake manifold and cylinder head with half the stock filling port area ... A-1 Figure A-2. Spacer between the intake manifold and cylinder head with half the stock swirl port area ... A-2 Figure A-3. Spacer between the intake manifold and cylinder head with blocked filling port ... A-3 Figure A-4. Spacer between the intake manifold and cylinder head with blocked swirl port A-4 Figure A-5. Spacers with stock port areas for unused cylinders ... A-5

(17)

xv

LIST OF TABLES

Table 1. Common SI engine fueling methods ... 2

Table 2. Alternative methods for well-homogenized premixture of gaseous fuel and air ... 10

Table 3. Components of the laboratory fueling setup from tanks to the engine ... 11

Table 4. Gas panel components and their respective functions ... 12

Table 5. Gas panel pressure regulator preset pressure settings ... 13

Table 6. Components of the fueling system assembly within the test cell ... 15

Table 7. XPI system modifications for use with alcohols ... 16

Table 8. Comparison metrics for ECU selection ... 19

Table 9. Sensors necessary for basic engine operation ... 20

Table 10. Alternative damping tank designs ... 27

Table 11. Intake manifold modifications ... 30

Table 12. Alternatives for varying port filling ... 31

(18)

xvi

(19)

1

1 INTRODUCTION

This chapter gives a brief introduction to internal combustion engines, the purpose of this project followed by the deliverables and delimitations.

1.1 Internal Combustion Engines – A Brief Overview

Internal combustion engines (ICEs) date back to the late 19^th century when Nikolaus Otto developed the spark-ignition (SI) engine and Rudolf Diesel invented the compression-ignition (CI) engine. ICEs have played a dominant role in the fields of energy and propulsion. However, the last three or so decades have seen vast developments in engine research and development to tackle the issues of pollution, fuel cost and market competitiveness (Heywood, 1988).

Engines can be classified in a variety of different ways, but for the purposes of this report, they shall only be classified as being either of the SI type or the CI type. This classification is based on the method used to provide the activation energy required to initiate the combustion process.

Both methods are applicable to fuels possessing different qualities. Spark ignition is used for fuels that have high auto-ignition temperatures. As the name suggests, the ignition is initiated by a spark plug which provides a high voltage electric arc. On the other hand, certain fuels can be ignited by simply compressing them in a quasi-adiabatic process (Heywood, 1988).

A majority of ICEs operate on what is known as the four-stroke cycle. Each of the four strokes takes place during half a rotation of the crank or 180°. Thus, one complete cycle requires that the crank rotate through a total of 720°. The four strokes that make up the cycle are as follows (Heywood, 1988):

1. The intake stroke which starts at Top Dead Center (TDC) and ends at Bottom Dead Center (BDC) during which the charge is inducted into the engine via the intake valves. The exhaust valves remain closed.

2. The compression stroke during which both the exhaust and intake valves are closed. The charge is compressed to a fraction of its initial volume by the upward motion of the piston form BDC to TDC. Combustion is initiated toward the end of this stroke and there is a further increase in the cylinder pressure.

3. The expansion stroke results from the expanding high-temperature gases which are products of the combustion. This is the only stroke during which work is extracted from the fuel. The expansion causes the piston to move from TDC to BDC. The exhaust valves open toward the end of this stroke.

4. The exhaust stroke during which the exhaust valve is fully open and the upward motion of the piston from BDC to TDC driven by inertia, causes the exhaust gases to be evacuated from the cylinder. The intake valves open toward the end of this stroke to allow fresh charge to be inducted for the next cycle.

1.1.1 Commonly used fueling methods:

SI engines employ four main fueling methods (see Table 1) (Heywood, 1988):

(20)

2

Table 1. Common SI engine fueling methods

Item number

Method Advantages Disadvantages

1 Carburetion – makes use of the Venturi effect to draw fine droplets of fuel into the intake air stream

• Well-mixed charge induction ⟹ no large rich/lean pockets in the cylinders and associated impact on emissions

• Can cause imbalance in cylinder filling depending on the design of the inlet manifold. In case of a log-style manifold, the cylinder closest to the throttle could run rich while the farthest cylinder could run leaner.

• Slower response to load changes due to the large volume between the point of mixing and the cylinders.

• Result in the intake manifold being filled with an inflammable mixture which is undesirable in the event of a backfire.

2 Throttle-body

injection using a fuel injector array upstream of the throttle valve in place of a carburettor

• Well-mixed charge induction ⟹ no large rich/lean pockets in the cylinders and associated impact on emissions

• More precise fuel

metering than

carburetion

• Same as those for carburetion

3 Port fuel injection • Eliminates the issue of

uneven charge

distribution

• Can premix fuel and air sufficiently well depending on runner design and injection timing

• Slight loss in volumetric efficiency when using gaseous fuels since part of the air in the ports and runners is displaced by fuel

• Port and runner length becomes important to avoid spill-over of injected fuel into runners of adjacent cylinders

(21)

3

• Effects of wall- wetting need to be accounted for when using liquid fuels 4 Direct injection – uses

one injector per cylinder located in the cylinder heads

• Eliminates the problem of reduction in volumetric efficiency inherent to port injection of gaseous fuels

• High injection pressures are needed

• In-cylinder fuel and air mixing problems could arise due to low penetration of gaseous fuels in air

1.1.2 CNG as a fuel

Natural gas is a hydrocarbon fuel, the majority composition of which is methane. The methane composition of natural gas can range from 75% to over 95% depending on the region of origin (Min et. al., 2002). Compressed Natural Gas (CNG), as the name suggests, is natural gas compressed to 200 to 250 bar. The simple chemical composition of methane, consisting of only one carbon atom per molecule, results in over 20% reduction in CO2 emissions compared to other hydrocarbon fuels. Experiments have shown an average decrease in developed power of 9% while running on CNG and it has a molar heating value that is 4.6 times less than that of conventional gasoline (Douailler et. al., 2011). However, the sources of natural gas are more plentiful than the sources of oil (Einewall, 2005).

Natural gas cooled cryogenically to around -160 °C is called Liquefied Natural Gas (LNG).

Liquefying natural gas and storing it at 10 bar (approx.) increases its volumetric energy density to nearly two and a half times that of CNG at 200 bar. However, LNG has its own drawbacks that prevent it from being as economical as CNG for commercial use. Firstly, the energy required to liquefy natural gas is much more than that required to compress it to 200 or 250 bar at room temperature. It also needs special insulated equipment such as containers for transportation and bulk storage, pumps for refueling vehicles and insulated on-board fuel tanks (Kumar et. al., 2011).

CNG is also widely available from sources that are not related to food in any way. Other fuels such as ethanol and vegetable oils used to create fatty acid methyl esters (FAME) or hydrogenated vegetable oils (HVO) either use edible matter directly, or are farmed on land which could be utilized for food production instead.

Ever increasing restrictions on emissions, scarcity of fossil fuels, questionable renewability of certain alternate fuels combined with low emissions has made CNG an attractive alternative to conventional oil based fuels.

1.2 Purpose

The ever-increasing demand from automakers to reduce vehicular emissions has made Compressed Natural Gas (CNG) one of the more promising options. The 4:1 hydrogen-to-carbon ratio of methane compared to the higher 2:1 (approx..) of typical diesel or petrol is advantageous for reduced CO2 emissions (Reif and Dietsche, 2011). High quality CNG

(22)

4

with a high percentage of methane also provides a virtually soot-free operation (Reif and Dietsche, 2011). Moreover, CNG has an Research Octane Number (RON) of approximately 130 which makes it more knock-resistant (Reif and Dietsche, 2011). This makes it possible to use higher compression ratios compared to typical gasoline engines. The higher knock-resistance of CNG also means that the engines running on CNG can be turbocharged to higher pressures for improved indicated efficiency. The decrease in volumetric efficiency due to displacement of air by fuel can also be addressed through forced induction.

In a research setting where one intends to study various phenomena concerning engine operation, it is often desirable to eliminate or isolate as many variables as possible in a controlled manner. In a multi-cylinder engine, flow and/or electrical phenomena occurring in one cylinder could possibly affect those in another cylinder. This makes it difficult to perform a controlled study of the phenomena of interest. Nevertheless, a single cylinder engine permits the study of these phenomena of interest without disturbances due to adjacent cylinders.

1.3 Task definition

The project aims at implementing flexible fueling of a series production multi-cylinder heavy-duty CNG fueled engine modified for single cylinder operation using an aftermarket electronic engine control unit (ECU). The modified engine is intended for further research and will be installed in one of the test cells at the Internal Combustion Engines laboratory at KTH Royal Institute of Technology, Stockholm. This thesis is part of the bigger Project CLOCK (Closed Loop Otto Combustion and Knock control). The participating partners are Scania CV AB, KTH Royal Institute of Technology, Sweden, and the Swedish Energy Agency. Although the first customer of the engine once installed at KTH will be Project CLOCK, future projects could potentially involve research on different fuels and fueling methods.

1.3.1 Research objectives

The main research objectives intended to be fulfilled through this project are:

1. Evaluating the best aftermarket ECU taking the extent of controllability offered into account while remaining easy to use in an academic setting

2. Making provisions for variable homogeneity of gaseous fuel-air mixture with due consideration to safety

3. Making provisions for selective intake port usage for future studies on its effect on in-cylinder flows

4. Making hardware provisions for implementation of liquid fuel direct injection

1.3.2 Project outline

The project involves the following steps/stages:

1. Selection of a suitable engine for modification

2. Selection of a suitable aftermarket ECU to control the fuel system

(23)

5

3. Selection of required sensors and actuators needed to implement the control

4. Making required mechanical modifications on the engine from a fueling system perspective

5. Installing the control hardware and configuring the control system as required 6. Delivering a working engine

The engine uses a port-injected CNG fueling system. The parameters of the fueling system that are expected to be controlled are: Mass flow rate of fuel, injection pressure and start of injection.

1.4 Deliverables and delimitations

This thesis project spanned over a duration of 20 weeks, was limited in the deliverables. The following deliverables and delimitations were formulated after mutual agreement between the author and the project commissioner.

1.4.1 Deliverables

1. The fuel injectors will be controlled with respect to start/end of injection and mass flow rate of CNG.

2. A conventional speed-density based open loop and lambda sensor-based closed loop control of the injectors will be implemented.

3. Additional injector(s) (capable of injecting a different gaseous fuel) will be installed in the intake air path thus varying homogeneity (and composition) of the resulting charge.

4. Provisions will be made for higher or lower flow rate through either intake port.

5. The selected aftermarket ECU should be capable of liquid fuel direct injection, if possible.

6. The ECU should be able to control Exhaust Gas Recirculation (EGR).

7. Safety devices such as flame arrestors and vent valves will be installed at appropriate locations

8. Documentation for future users will be created for all the non-standard auxiliaries mounted on the engine.

1.4.2 Delimitations

1. Mechanical modifications for conversion to single cylinder if using a multi-cylinder engine will not be the main focus although assistance for the same will be provided.

2. Cycle-to-cycle injector control will not be carried out.

3. The ECU will not be programmed at machine level but only via the bundled software GUI.

(24)

6

4. Ignition control remains outside the scope of this project. It shall be the focus of a concurrent thesis project carried out by a fellow student.

5. Controlling the liquid fuel direct injection fueling will be excluded.

6. Further components in the intake air path upstream of the damping tanks were considered to belong to the test cell and were excluded from the scope of this project.

Mechanical modifications to the engine for single cylinder operation were carried out as part of a separate thesis project concurrent to this one. The author has assisted in this work.

(25)

7

2 ENGINE SELECTION

The first section of this chapter presents the rationale behind using a single cylinder engine for research. The final choice of engine based on certain requirements has also been discussed.

2.1 Single cylinder research engine

Internal combustion engines used for research can be either single cylinder or multi-cylinder based on the phenomenon of interest being studied. When the phenomena of interest are combustion, combustion chamber designs, in-cylinder flows, flows across valves etc., it is prudent to use a single cylinder engine.

A single cylinder engine has the inherent advantage of not being affected by the adverse flow interactions due to other cylinders. It is also logical to use a single cylinder engine to reduce fuel consumption. The ability to facilitate easy replacement of internal components suitable for the type of tests intended to be run is another advantage.

2.2 Requirement specification and alternatives

The most important requirement regarding the engine was the availability of spare parts in the future. The engine was a contribution from Scania CV AB towards Project CLOCK. At the beginning of the project, there were three tentative engine options to choose from:

1. A custom-built single cylinder engine based on the 13-liter Scania DC13 engine 2. A 5-cylinder Scania OC09 9-liter Otto engine

3. A 6-cylinder Scania OC13 13-liter Otto engine

Unlike options 2 and 3, the custom-built single cylinder engine would not require mechanical modifications to run it as single cylinder. However, most of the parts for such an engine were custom-made. From the perspective of ready availability of spare parts, it was decided that the series-produced multi-cylinder engines would be a more practical choice. Additionally, the multi-cylinder engines were already balanced and would not vibrate as much even when only one cylinder was fired.

2.3 Implementation

The availability of a run-in and tested 5-cylinder, 9-liter engine, made it the default choice. All requisite modifications for conversion to single cylinder operation were done by another student, assisted by the author. The details of these modifications can be found in a separate report (Lius, 2018).

It was decided to use cylinder 1 (first cylinder from the front of the engine) as the active cylinder because of the following considerations:

1. The open end of the intake manifold lies adjacent to cylinder 1, meaning that the existing flange on the manifold could be used without subjecting the manifold to machining

(26)

8

2. Future researchers would want to use the engine for optical studies as well, so cylinder 1 provides better access for the optical diagnostic instruments

3. Setting up ignition timing for cylinder 1 would be easier

Figure 1. Top view of the engine illustrating reasons for choosing to retain cylinder 1 as the active cylinder

The selected engine in its unaltered state, was a port fuel-injected spark-ignited natural gas engine with two intake and two exhaust ports.

(27)

9

3 FUELING SYSTEM

The following chapter is a discussion on the fueling of the engine, the challenges associated with gaseous fueling and final implementation.

3.1 Fuel-air mixture formation

This section discusses the main challenge involved in mixing two fluids of comparable densities. The fluids of interest to this project are CNG and air. Researchers have reported turbulence, volume and residence time available for mixing affect the homogeneity of the mixture. Methods to promote the turbulence intensity and to vary the residence time to achieve varying mixture homogeneity were implemented.

3.1.1 Gaseous fuel-air mixture

Turbulence caused by the transfer of momentum from the periphery of a jet to the surrounding fluid provides the necessary shear forces to generate eddies and vortices that promote mixing.

The appreciable difference in the densities of liquid fuel droplets and air, create significant turbulence at the jet-air interface. Vaporization of the droplets creates a vapor trail behind them which, combined with the momentum induced small-scale vortices, accelerates the diffusion rate. This process helps create a more uniformly distributed mixture (Khan et. al., 2009).

However, the shear forces between a gas jet injected into another quiescent gas, are low and mixture formation may become diffusion dominated, which is a relatively slow process. The comparable densities of gaseous fuels and air, do not allow the utilization of jet penetration for large scale mixing throughout the volume. Low penetration is characteristic of gas jets irrespective of injection pressure. Research has shown that mixing can be improved by inducing large-scale turbulence (Herrera et. al., 2012; Khan et. al., 2009).

3.1.2 Effects of fuel injector location and injection timing

The location of the fuel injectors determines the residence time available for the fuel to form a homogeneous mixture. Researchers have reported that injecting fuel just downstream of the throttle body provides a greater time for homogeneous mixture formation owing to the length of the intake pipe from the injector(s) to the cylinders. The turbulence prevalent after the throttle valve also serves to promote mixing (Soanes and Koopmans, 2012; Puzinauskas et. al., 2000).

In the case of port fuel injection, variation in mixture homogeneity was reported to be sensitive to the injection timing. The decrease in mixing length makes the mixing dependent on the inlet valve opening and pulsations in the manifold (Soanes and Koopmans, 2012). It is believed that injecting against a closed inlet valve can provide enough residence time for mixture formation.

However, port injection into an open inlet valve was reported to aid turbulence effects and promote in cylinder mixing. Residual mixture in the port from the previous injection cycle was observed to reduce the effect of any stratification in the cylinder. However, this strategy was not found to produce any prominent effects at high loads (Czerwinski et. al., 1999).

(28)

10

3.2 Charge homogeneity variation

As discussed in Section 3.1.2, the location of the injectors determines the homogeneity of the charge. To be able to induct a well-homogenized mixture into the cylinder, it was necessary to provide as much turbulence and residence time as possible.

The following table provides a comparison of the various methods for creating well-homogenized mixtures (see Table 2):

Table 2. Alternative methods for well-homogenized premixture of gaseous fuel and air

Item number

1 Rapidly pulsing one or more injectors installed well upstream of the intake manifold

• Potential to create evenly spaced fuel and air pockets of small volume if executed properly

• Needs long length of pipe before the manifold to provide highest degree of homogeneity

• Injector solenoids could overheat due to rapid pulsing 2 Using MAN gas-mixer • Good mixture

formation

• MAN spare part availability and lead time uncertain

• No performance testing done by Scania, hence no assurance on performance from Scania engineers 3 Using the Scania Euro

5 gas-mixer

• Excellent mixture formation assured

by Scania

engineers

• Standard Scania part still in production ⟹ quick replacement possible if needed

• Bulkier compared to the MAN design

For studies such as the effects of in-cylinder flows on homogeneity, less homogeneous mixtures were needed.

(29)

11

3.3 Laboratory fueling setup

The fueling system in previous projects on CNG engines at Scania have used the following setup (see Figure 2) (Scania, 2018):

Figure 2. Laboratory gas supply schematic (Scania, 2018)

The various components comprising the laboratory fueling setup are as listed in Table 3:

Table 3. Components of the laboratory fueling setup from tanks to the engine

Item number

Component Function description

1 Fuel tanks with electronic and manual shut-off valves for safety

To store CNG under a pressure of up to 250 bar.

2 Primary fuel pressure regulator fed by high pressure lines

Expands the fuel from the feed pressure to 50 bar. Fuel is supplied to the various test cells at 50 bar.

Remark: The expansion process causes a considerable temperature-drop in the regulator. Freezing is prevented by electrically heating the regulator unit.

3 Secondary fuel pressure regulator Expands the fuel from 50 bar to 10 bar.

4 Gas panel with electronic and manual shut-off valves for safety

Maintains the supply pressure such that there is a constant 7 bar pressure difference across the injectors.

(30)

12

5 Fuel rail Temporary fuel storage before

injection.

Remark: Another function of the fuel rail especially in a multi-cylinder setup, is to ensure that each injector receives fuel at the same pressure and that none of the injectors are starved.

6 Fuel injectors connected to the fuel rail

Injecting fuel in the intake ports to induct a premixed charge into the cylinders

Safety devices such as flame traps and shut-off valves (both manual and automatic) are located at various points in the circuit

3.3.1 Gas panel

Installed within the test cell between the secondary pressure regulator and the fuel rail, is a gas panel which consists of the following components (see Table 4 and Figure 3):

Table 4. Gas panel components and their respective functions

Item number

Component Function

1 Reference pressure line to intake manifold

To maintain a constant pressure difference of 7 bar across the injector irrespective of intake manifold pressure

2 Pressure limiting valve To prevent fuel overpressure which could impact injector opening time and current

3 Pressure regulator To reduce the gas from tank pressure (~ 200 bar) to injection pressure (~ 7.3 bar)

4 Pressure sensor For diagnostics

5 Solenoid valve For automatic gas supply switching on/off on the low-pressure side when the power is switched on/off

6 Manometer For manual inspection of supply pressure from the tank

7 Fuel filter To remove oil droplets possibly introduced during compression, and other impurities from the gas

8 Manual shut-off valve To manually turn on/off low-pressure gas supply

(31)

13

9 Coolant in/out To prevent freezing of the pressure regulator due to expansion of the gas from tank pressure to injection pressure

Figure 3. Gas panel for CNG engines (Scania, 2018)

The gas pressure regulator has the following preset pressure settings (see Table 5):

Table 5. Gas panel pressure regulator preset pressure settings

Outlet pressure 7.3 bar(g)

Inlet pressure 17.2 bar(g)

Pressure relief valve range 10.3 to 12 bar(g)

The regulator has an operating temperature range of - 40° C to 120° C. This gas panel is also a standard assembly on production trucks.

(32)

14

3.4 Installation assembly

A fueling system like the one described in Section 3.3 was recommended to be installed at KTH. The gaseous fuel-air path inside the test cell was designed as a modular unit as shown in Figure 4.

Figure 4. Illustration of the recommended gaseous fueling system layout

(33)

15

The components of the fueling system along the intake air path are as follows (see Table 6):

Table 6. Components of the fueling system assembly within the test cell

Item number

Component Function

1 Damping/mixing tank with overpressure rupture disc

• The baffle in the tank provides damping of pressure pulses propagating from the engine;

enables more accurate air mass flow measurement

• Mixing chamber for gaseous fuel and air when the gas-mixer is installed upstream

2 Gas-mixer (Scania Euro 5 design)

• To generate turbulence and introduce the gaseous fuel into the air stream for good mixing

3 Throttle • To control the flow of air/charge mixture into the engine

4 Flame arrestor • To prevent flame propagation upstream from the intake manifold in case of backfire

5 Flexible pipe • To minimize the vibrations due to single cylinder operation from oscillating the upstream components, particularly the heavy flame arrestor assembly

6 Fuel injectors upstream the intake manifold

• To be able to induct a semi-homogeneous charge mixture into the engine if future studies call for it 7 Port fuel injectors • Stock setup to inject gaseous fuels into the air

stream in the intake runners

As described in item 1 of Table 6, the damping tank can be used to create the most well-mixed charge possible if the gas-mixer is installed upstream of the tank. The turbulence created by the mixer combined with the greater residence time creates a very homogeneous mixture.

The flame arrestor and the overpressure rupture disc to vent the damping/mixing tanks on the intake and exhaust sides have been discussed in detail in Chapter 6.

3.5 Liquid fuel direct injection

Liquid fuel direct injection was decided to be included in the fueling system for the engine with the aim to offer fueling flexibility.

3.5.1 Assignment

The engine would eventually replace the existing engine in Test Cell 2 at KTH. The existing engine is based on a Scania 12-liter, 6-cylinder diesel engine and has been modified to provide optical access on cylinder 1. The new engine would be more beneficial for future projects if

(34)

16

provisions are made for flexible fueling. A requirement from the commissioner was to be able to directly inject liquid fuels such as diesel and alcohols.

3.5.2 Implementation

To fulfill this requirement, it was decided to use the Scania XPI direct injection fueling system.

Apart from the extremely high pressure (~ 1500 bar) fuel injection capability of the system, the cylinder heads are designed to accommodate the XPI injectors. The engine also had the gear-train to drive the high-pressure fuel pump pre-installed. These features necessitate no extensive machining to be able to implement liquid fuel direct injection.

3.5.3 Modifications

Some modifications needed to the XPI system are listed in Table 7:

Table 7. XPI system modifications for use with alcohols

Item number

Modification Purpose Method

1 Re-calibrating the high-pressure fuel pump output

The pump needs to supply fuel only to one injector in a single cylinder engine

One of the two pumping barrels of the pump need to be plugged shut

2 Plugging of unused outlets on the high- pressure fuel rail

The stock fuel rail is meant for a five/six-cylinder engine and will have four/five unused outlets respectively when used with a single cylinder engine

Plug the unused outlet ports with standard plugs

The high-pressure line from the fuel rail to the high-pressure connector of injector number 1 needed to be re-oriented. This was necessary since the intake manifold of the stock natural gas engine differs in design from that of the diesel engine. On the natural gas engine, the intake manifold occupies the location of the diesel fuel rail and hence the need for the re orientation.

Additionally, the high-pressure fuel delivery lines between the high-pressure pump and the fuel rail need to be elongated to connect with the rail in its modified location.

(35)

17 (a)

(b)

Figure 5. (a) Proposed orientation of the high-pressure fuel rail, needing 45 mm spacers between the manifold and cylinder head; (b) Proposed orientation of the high-pressure fuel rail demonstrating the use of the existing

high-pressure connector without changing any bend radii

The control of the XPI system is intended to be done via the Scania S8 diesel ECU, again due to existing compatibility.

(36)

18

(37)

19

4 ELECTRONIC ENGINE CONTROL

In this chapter, a discussion of the electronic engine control unit used for this project has been presented. Also included, are the most important electronic fueling system components – the injectors and the lambda control.

4.1 Engine control unit

This section provides a brief description of the requirements set on the engine control unit in terms of controllability, followed by selection and implementation of the control unit.

The production version of the original engine is controlled by a Bosch EGC 4 engine management system specific to natural gas engines. However, this control unit contains proprietary programming from Bosch and is not open for modification. In order to set up the engine for the intended tests, a user-configurable ECU was found necessary. Aftermarket ECUs which usually have a Graphical User Interface (GUI) were found suitable. These ECUs, while not programmable at the code level, still offered more flexible configurability compared to the Bosch system. Deciding the most important comparison metrics for selection of the ECU was undertaken as part of this project.

4.1.2 ECU selection criteria

The main criteria for choice of aftermarket ECU for the project were as follows (see Table 8):

Table 8. Comparison metrics for ECU selection

Item number

Criterion Description

1 Configurability and ease of use The chosen unit should be easy to configure, preferably on-the-fly through a Graphical User Interface (GUI). A Swedish manufacturer is preferred so that requests for feature addition/modification can be done quickly.

2 Cost The cost should be as low as possible since

it is possible that the new Scania Otto ECU being developed by NESI could replace the unit eventually.

3 Number of inputs and outputs The ECU should be able to read and control as many parameters as possible.

Inputs: Crank and cam position, intake air temperature and pressure, oil pressure, coolant temperature, exhaust temperature

(38)

20

shall be the bare minimum parameters to be read.

Outputs: Electronic throttle control, drive for at least six low impedance injectors, EGR valve control (if possible) shall be the bare minimum outputs.

4.1.3 ECU selection and installation

The criteria mentioned in the previous section resulted in a choice between either the MaxxECU Race or the NIRA i7x. Recommendations from Scania engineers with previous positive experience with MaxxECU were strongly influential in the decision. The NIRA i7x was also found to lack driver circuits for low impedance injectors. Thus, the MaxxECU Race was opted for and installed.

The main sensors needed for engine operations (see Table 9) were connected to the ECU via the accompanying wiring harness.

Table 9. Sensors necessary for basic engine operation

Item number

Sensor Description

1 Engine speed sensor Senses rotational speed and position of the crankshaft to determine injection and ignition timing.

2 Camshaft position sensor Used in conjunction with the engine speed sensor to determine which of the pistons is towards the end of the compression stroke.

3 Intake manifold combined pressure and temperature sensor

Determines the air mass flow rate into the engine via the Speed-Density algorithm.

4 Wideband lambda sensor Determines the stoichiometry of the previous combustion event and used for feedback control.

5 Knock sensor Senses the occurrence of engine knock so that ignition timing can be retarded.

4.2 Fueling control

The following section describes the hardware and considerations that need to be accounted for to control the fueling system.

(39)

21 4.2.1 Types of CNG injectors

There exist two types of injectors which are suitable for CNG applications namely, the voltage-controlled and the current-controlled types. Both types of injectors have been used in all kinds of injection methods such as throttle-body injection, multi-point injection (sequential and continuous) as well as direct injection. Both types of injectors have electromagnetic solenoids which can be opened and closed by energizing and de-energizing the coils (Reif, 2015).

As the name suggests, the voltage-controlled injectors are supplied with the operating voltage by the ECU injector driver circuit. A corresponding current builds up the magnetic field needed to retract the injector valve and hold it open against the rail pressure (see Figure 6). The finite time needed for the build-up of the magnetic field usually results in slower response of these type of injectors.

(a) (b)

Figure 6. Typical solenoid current build-up, injector valve lift and corresponding fuel flow rate for (a) voltage- controlled injector and (b) current-controlled injector (Reif, 2015).

One situation where this limitation is highlighted is during engine idling. Idling requires that the least amount of fuel be injected to just keep the engine running and engine management systems are usually programmed to lean out the mixture to conserve fuel. The slower response time of this type of injectors inevitably results in richer idling which is not desirable.

The current-controlled injectors on the other hand, are supplied with a peak current which lifts the valve from its seat. Once the valve lifts from its seat, the driver then reduces the current to

(40)

22

the amount needed to hold the valve open. This current reduction is done to prevent the higher peak current from potentially burning the coils (Reif, 2015).

4.2.2 Installation

The injectors used for this application are the standard Bosch NGI2 gas injectors with a nominal impedance of 8.5 Ω. The ECU has eight injector driver outputs. It can supply 8 A of peak current and 4 A of hold current at 12 V per driver output (MaxxECU, 2018). The injectors require 1.3 A of peak current and 0.35 A of hold current. At normal operating conditions, the injectors are restricted to 30 mins at 100% duty cycle. It was therefore decided to install two injectors in parallel per driver output. Reducing the on-time per injector for a given quantity of fuel was the rationale behind this decision.

4.2.3 Verification

The ECU was set up for sequential injection every 720° CAD. However, it was not possible to test the synchronization. Recommendations for testing the injectors and injection timing have been made in Chapter 8. These tests will be carried out once the engine has been installed in the test cell where it can at least be cranked using the dynamometer if not under its own power.

4.3 Lambda control

This section discusses the motivation behind maintaining unit lambda, associated hardware namely the lambda sensor and implementation of lambda control.

4.3.1 Stoichiometric combustion and the three-way catalyst

Lean pre-mixed combustion can reduce the combustion temperature, but prolonged operation in this mode results in an increase in NOx emissions. Extremely lean mixtures beyond the flammability limit result in flame extinction and misfire. Experiments have found that lean burn engines produce very low emissions and lower fuel consumption under low loads. However, at higher loads, the high temperatures and excess air lead to an increase especially in the NOx

component of the emissions (Saanum et. al., 2007; Verhelst et. al., 2006). Stoichiometric combustion on the other hand, produce higher NOx, CO and HC emissions and have slightly higher fuel consumption at low loads. The three-way catalytic converter however, can treat these emissions as long as the combustion is stoichiometric (Saanum et. al., 2007;

Verhelst et. al., 2006).

Einewall et. al. (2005) also report a decrease in NOx emissions by over 99% and hydrocarbon emissions by over 90% by operating stoichiometrically compared to lean burn. However, it was found that EGR inclusion resulted in significantly higher CO emissions along with a slight drop in efficiency. It must also be noted that the trade-off window between NOx reduction and CO oxidation in the catalyst is of the order of 0.01 in terms of lambda (see Figure 7) which requires the lambda control to be very accurate.

(41)

23

Figure 7. Lambda window for highest TWC conversion efficiency (Reif, 2015; Heywood, 1988)

4.3.2 Exhaust gas oxygen sensors/Lambda sensors

Measurement of oxygen concentration in the exhaust stream is done by exhaust gas oxygen sensors. There are two types of O2 sensors namely, the narrow-band and the wideband sensor.

Both types of sensors are essentially galvanic concentration cells (or Nernst cells) consisting of two half-cells. The composition of these half-cells is identical except for the concentration of the electrolyte. This concentration difference is the motive force behind the transport of ions.

Ions are charged particles and their motion is indicative of the concentration of the electrolyte in one half-cell provided the electrolyte in the other half-cell has a known concentration (Marek et. al., 2003; Reif, 2015).

One type of O2 sensors is called switching type or narrow band O2 sensor. Such sensors only sense a deviation in mixture quality from stoichiometric. They do not indicate the extent of the deviation. The narrow-band O2 sensor compares the residual oxygen content in the exhaust gas with the oxygen content in the reference atmosphere. A ceramic element, usually made of zirconia is coated with a noble metal such as platinum on either side, which act as electrodes.

The ceramic element is conductive for oxygen ions. The different concentrations of oxygen on either side of the element generates an electrical voltage across the metal electrodes (Marek et. al., 2003; Reif, 2015). Such a Nernst cell produces a voltage of around 450 mV across its electrodes when there exists no concentration difference across the cell. A rich mixture results in a voltage between 450 mV and around 900 mV, whereas a lean mixture results in a voltage between 450 mV and close to 0 mV (see Figure 8). The noble metal coating also serves to catalytically convert other species in the exhaust gas stream such that only an equilibrium reaction involving oxygen takes place in the cell (Marek et. al., 2003; Reif, 2015)

(42)

24

Figure 8. Typical narrow-band O2 sensor output voltage as a function of lambda (Reif, 2015)

The change in oxygen concentration for a 0.02-unit lambda span, can range from 9×10^-15% vol.

on the rich side to 0.2% vol. on the lean side (Reif, 2015). As a result, the ECU constantly hunts around the stoichiometric point, trying to achieve it but inevitably ending up with an overshoot in either direction. Cycle-to-cycle control of the air-fuel ratio would require the use of a predictive model for the amount of air that will be inducted into the cylinders (Reif, 2015).

The other type of O2 sensor is called the wideband lambda sensor. Such sensors not only indicate deviations from stoichiometry but can also determine to what extent the mixture deviates. This allows for more precise control of the air-fuel ratio without the ECU hunting.

Knowing the extent of the deviation, the ECU can more accurately determine the amount of fuel to inject as a corrective measure.

Wideband O2 sensors consist of two electrochemical cells (see Figure 9(a)) with a sensing gap in between them, where the measurement takes place. The first is a regular Nernst cell which produces a voltage based on whether the sample of exhaust gas being measured is rich, lean, or stoichiometric. The second, is a pumping cell. It is so named, since a voltage can be applied to this cell by a control circuit to enable pumping oxygen ions into or out of the sensing gap (Reif, 2015).

(43)

25 (a)

(b)

Figure 9. (a) Construction and (b) characteristic curve of a typical wideband lambda sensor (Marek et. al., 2003).

If the exhaust gas is rich, the Nernst cell which has one of its electrodes exposed to reference air and the other to the gas sample, puts out a voltage higher than the reference voltage of 450 mV. The controller compares the output to the reference supplies a voltage to the pumping cell electrodes such that oxygen ions are pumped from the exhaust stream into the sensing gap. The oxygen ions pumped into the sensing gap result in the Nernst cell voltage returning to 450 mV (Marek et. al., 2003; Reif, 2015). In case of a lean exhaust gas sample, the pumping voltage is reversed such that oxygen ions are pumped out of the sensing gap to return the Nernst voltage to 450 mV (see Figure 9(b)) (Marek et. al., 2003; Reif, 2015). In either case, the current set up in the pumping cell due to the movement of ions provides the ECU with a quantitative measure of the mixture quality. Stoichiometric mixtures on the other hand, yield zero pumping current (Marek et. al., 2003; Reif, 2015).

The O2 sensors are usually located in the exhaust stream between the turbocharger and the catalytic converter. The choice of location of the sensor is dictated mainly by the temperature restrictions of the sensor. This requires the exhaust gases to traverse a certain volume between the exhaust valve and the sensor. The farther away the sensor is, the larger the volume that the exhaust gases need to traverse before reaching the sensor. This condition results in a limitation on the response time of the feedback control loop. For the quickest response times, the lambda sensor ideally needs to be located in the exhaust port. However, since this is not possible, there will always be a delay between when the sensor senses the oxygen concentration and when the mixture correction is implemented by the injectors. Lambda sensors can also be used to compensate for variations in CNG composition which is usually unknown to the end users of the vehicles.

4.3.3 Implementation

The ECU is equipped with its own wideband lambda controller including an inbuilt PWM heater. The heater is used to heat up the sensor to its optimum operating temperature of over 350°C. The supplied sensor, a Bosch LSU 4.2 wideband lambda sensor, was installed in the exhaust gas stream. Temperature restrictions on the sensor were considered in order to

(44)

26

determine the installation location. The sensor was delivered pre-calibrated. However, the closed-loop lambda control could not be tested as the sensor would remain exposed to ambient air until installed in the test cell with the associated piping in place.

The ECU also features a proprietary Speed-Density algorithm for feed-forward lambda control.

The ECU consists of an inbuilt reference pressure port to which a pressure line was connected.

The other end of this line was connected to the intake manifold. Nevertheless, the ECU does not contain pre-programmed fuel maps for CNG. Therefore, for the open loop control to function, a manual fuel map needs to be programmed into the ECU. For initial engine startup and idle running, it was decided to use a modified gasoline fuel map. The gasoline equivalent CNG mass flow rate was determined assuming an average calorific value of gasoline and CNG.

This fuel mass flow rate has been used to set the injection duration but if this method results in unit lambda, has not been tested.

(45)

27

5 GAS EXCHANGE SYSTEM

A majority of the discussion presented in this chapter pertains to the intake air handling system.

A short discussion has also been presented on the exhaust side damping tank.

5.1 Intake damping/mixing tank

The layout of the air intake system has been illustrated in Section 3.4 (see Figure 4). The engine will be supplied with boosted charge-air from an external compressor belonging to the laboratory. The first component in the air flow path is the damping tank which can also be used for premixing a gaseous fuel and air.

5.1.1 Purpose

The tank has multiple purposes, namely:

1. Dampen pulsations in the air path enabling more accurate air mass flow measurement 2. Preparing well homogenized mixtures of air and gaseous fuels before induction into the

engine

3. For mixing EGR into the intake air stream 5.1.2 Alternative designs

Two alternative designs were considered for the tank and have been described below (see Table 10 and Figure 10).

Table 10. Alternative damping tank designs

Item number

Description Features

1 Alternative 1 • 12 liters in volume

• Contains a baffle in the middle for damping pulses

• Not been used for mixing gaseous fuel and air 2 Alternative 2 • Approximately 40 liters in volume

• The large volume itself is responsible for damping

• Has provisions for injecting fuel into the tank for premixed charge formation

• Used to mix EGR with the intake air

• Turbulence induced by the throttle valve installed upstream of the tank is utilized for improved mixing

(46)

28

(a) (b)

Figure 10. Illustration of intake damping tank (a) Alternative 1 with baffle; (b) Alternative 2 with larger volume

5.1.3 Design choice and implementation

The decision to replicate the Scania design was foremost based on space restrictions in the KTH test cells. A simulation study by Scania has recommended that the volume of the tank be increased to 18 liters instead of the current 12 liters (Scania, 2016). The study states a drop in volumetric efficiency of the engine at high speed and high load during simulations as the reason for the recommendation.

The recommendation to use an 18-liter tank was agreed upon with the commissioner. However, from a functionality perspective, the tank was also intended to be used for well-homogenized

(47)

29

charge mixture formation. At present, no provisions have been made for introduction of EGR into the tank. If needed however, the modification to the tank would be minor.

5.2 Throttle valve

This section discusses the purpose of the throttle valve, choice of size based on control concerns and its implementation.

5.2.1 Purpose

The aim of a throttle valve is to control the amount of air which is inducted into the engine. A throttle valve is of more importance in a stoichiometric SI engine since the amount of air inducted determines the quantity of fuel to be injected. The precision of the injected fuel quantity is critical from an exhaust after-treatment perspective due to limitations of the three-way catalyst.

5.2.2 Throttle sizing

The full engine uses a 57 mm electronically actuated Bosch throttle valve matched to the 9-liter displaced volume. Issues with air mass flow control using a large throttle valve for single cylinder operation were expected. It was expected that the steep initial opening ramp of the throttle plate, could resulting in larger throttle angles than desired at low air mass flow rates.

However, Scania engineers reported that previous tests run in their test cells have used the same throttle. Furthermore, it was reported that the engine runs on wide-open throttle above 40%

load. It is only at lower speeds and loads that the throttle has any function at all. It was therefore decided to retain the throttle from the full engine.

5.2.3 Implementation and verification

The ECU supports a single electronic throttle with two motor control outputs and two throttle position sensors. The two outputs are Pulse-width Modulated (PWM) and the position sensor characteristics were as provided by the manufacturer. The position sensors are potentiometers and moving the throttle plate manually could cause a shift in the zero set-point. This in turn, could cause a difference in the clockwise and counter-clockwise rotation angles of the throttle plate. Throughout the project, care was taken not to move the throttle plate manually. Therefore, the position sensors did not need to be re-calibrated. The diagnostic function of the ECU was used to test the throttle actuation. Also, the throttle is a normally-closed valve and was found to shut-off after power supply was cut. However, no tests have been carried out scoping the position sensor output against the demanded position. Such a test will only be possible after the engine has been connected to a dynamometer with throttle control or at least a throttle pedal to determine the demand.

5.3 Intake manifold modifications and implementation

The intake manifold required some modification due to single cylinder operation and liquid fuel direct injection (see Table 11).

(48)

30

Table 11. Intake manifold modifications

Item number

Modification Reason Method

1 Decreasing the volume

The other cylinders were deactivated by removing the intake valves. The intake manifold was retained to prevent foreign objects from entering the cylinders.

However, it was also not desirable to fill the unused cylinders with premixed fuel and air.

A small section of the stock manifold between the swirl port runner of cylinder 1 and filling port runner of cylinder 2 was sawed off. A solid aluminium plate was welded in place, thereby separating the manifold internally while retaining stock external form (see Figure 10).

2 Moving the

manifold away

from the

cylinder heads

Reorientation of the high-pressure fuel rail (see Section 3.5.2) required the manifold to be moved away from the cylinder heads by at least 45 mm.

Aluminium spacers with holes matching the port holes (see Appendix A). If each spacer cannot be made from a single piece, two 25 mm thick pieces may be used per cylinder with gaskets in between.

Figure 11. Illustration of the method used to adapt the intake manifold for single cylinder operation

5.4 Variable intake port filling

This section describes the need for selective filling of the intake ports, the alternatives considered to achieve the objective and final choice of method.

With potential future projects in mind, a method to vary the amount of charge flowing into either of the two intake ports was requested by the commissioner. The current cylinder head

(49)

31

has a swirl port and a filling port. Increasing the amount of charge flowing into either port strongly affects in-cylinder flows. While more flow through the swirl port promotes swirl, increased flow through the filling port promotes tumble.

5.4.2 Alternative solutions

Three alternative solutions were considered to fulfill the assignment described in Section 5.4.1 (see Table 12).

Table 12. Alternatives for varying port filling

1. Using separate throttle valves of appropriate dimension for each port

• Possibility for electronic control

⟹ good precision and no manual work needed

• Finding appropriately sized throttles

• Modifications to the intake runners with a complex geometry

2. Installing baffles in the manifold to direct the flow towards the desired intake port, see Figure. 11

• Simpler than individual throttles

• Need for extensive machining of the manifold

• Requires at least three sections of manifold; One stock and one each for directing flow towards each port

3. Installing spacers between the manifold and the cylinder head with either port partially closed, see Appendix A.

• Even simpler than baffles to implement

• Requires the manifold to be removed and re-fitted to change the port being restricted

(a) (b)

Figure 12. Illustration of alternative 2 for variable intake port filling; (a) flow directed towards swirl port and (b) flow directed towards filling port

Heavy-Duty Spark-Ignited Single Cylinder Engine Fueling System Rohan Sharad Kittur