effects of charge motion control during cold start of si

Effects of Charge Motion Control

During Cold Start of SI Engines

by

Dongkun Lee

B.S., Mechanical and Aerospace Engineering Seoul N ational University, 2003

Submitted to the Department of Mechanical Engineering in partial fulfillment of the requirements for the degree of

Master of Science in Mechanical Engineering

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

September 2005

© Massachusetts Institute of Technology 2005. All rights reserved. )

Signature redacted Author ............. ~-.;'-"; ........................................... .

C,... Department of Mechanical Engineering August 19, 2005

Signature redacted Certified by ......................... ' .................. . , . ':' .' ...... '"':' .. .

John B. Heywood Professor of Mechanical Engineering

~ Thesis Supervisor

Signature redacted Accepted by ........................................................ .

Lallit Anand Chairman, Department Committee on Graduate Students

MASSACHUSETTS INS OF TECHNOLOGY ' ___ -NOV 0 7 2005

LIBRARIES

To my mother

3

Effects of Charge Motion ControlDuring Cold Start of SI Engines

by

Dongkun Lee

Submitted to the Department of Mechanical Engineeringon August 19, 2005 in partial fulfillment of the

requirements for the degree ofMaster of Science in Mechanical Engineering

Abstract

An experimental study was conducted to investigate the effects of various intakecharge motion control approaches on the cold start-up process of a port fuel injected SI en-gine. Engine experiments were performed to assess the effectiveness of enhanced chargemotion on mixture preparation, combustion, and hydrocarbon (HC) emissions. Differentcharge motions were produced by three different shapes of charge motion control valves(CMCV), which block off 75% of the engine's intake ports.

Cold-fluid steady state experiments were carried out with the CMCV open and closed.Increased charge motion with the CMCV closed was found to shorten the combustion dura-tion, which caused the 50% mass fraction burned to occur up to 50 CA earlier for the samespark timing. By the use of the CMCV, significant improvements in combustion stability(up to 1.5% reduction in COV) and fuel efficiency (up to 8% reduction in ISFC) wereachieved with increased levels of spark timing retard due to enhanced air-fuel mixing andrelatively faster burning.

Engine start-up experiments were conducted with three different geometric chargemotion control valves. The CMCV improved mixture preparation due to increased swirland tumble intensities, which enhanced fuel distribution and evaporation. Moderate sparkretard (AOs, -5') was found to reduce the engine-out HC emissions during the enginestart-up process. Peak engine-out HC emissions with the CMCV closed occurred in the ear-lier stages of engine start-up process relative to the CMCV open case. Greater fuel vapori-zation and faster burning rate with the CMCV closed allowed reduced fuel injection andadditional spark retard, resulting in significantly reduced HC emissions. The engine-out HCemissions were reduced by 18% during the 0-3 seconds and by 7% during the 3-20 seconds.

Thesis Supervisor: John B. HeywoodTitle: Professor of Mechanical Engineering

5

Acknowledgments

I am ever grateful to God, the Creator and the Guardian, to whom I owe my very exis-tence. He has assured me that He will be with me and pointed out to me that no one hasknowledge, power, or importance apart from him.

There have been a number of people without whom this thesis could not have beencompleted. My deepest gratitude goes out to my supervisor, Professor John Heywood. Ithas been an honor to be mentored by a leader of such great character, knowledge, and in-tegrity. He has tolerated an incalculable number of questions from me and has weatheredthat storm beautifully. He has not only taught me about internal combustion engines butalso showed me the way of thinking to conduct research.

I would like to thank all of the members of the MIT Engines and Fuels Research Con-sortium (General Motors, Ford, Daimler Chrysler, Delphi, and Saudi Aramco) for theirvaluable feedback and support. Special thanks to Dr. Ron Herrin and Dr. Norm Maasshoffof General Motors, who provided the engine, charge motion control devices, and flow test-ing results. It has been a pleasure working with them, and I hope to continue our workingrelationship in the future.

The Sloan Automotive Laboratory has been a great place to work, learn, and socializewith fellow students and dedicated MIT employees. Thanks to Michael Gerty and BrianHallgren for their assistance and companionship for the past two years. It is impossible toname everybody, but I would like to thank those who made my life at the lab interestingand helped me with my research.

I am very fortunate to have wonderful friends with whom I have shared many greattimes. I would like to extend special thanks to my dream house members (Soon-Jo Chung,Ki Tae Nam, Hyun Jin In, and Jooeun Ahn) for their friendships which have made my lifeat MIT more enjoyable and truly colorful.

All these years, my family has always been there for me. I would like to express myundying gratitude to my mother, Seok Joong, for her love, prayers, and for providing me amodel of how I should live. I also would like to thank my sister, Jin-Ju, for her support andencouragement. Knowing that I have you standing behind me, I will never be afraid tomarch forward. I love you. Finally, I wish to thank my father, Seung Yeon, who I wouldgive anything to have seen me finish this research project.

Dongkun LeeAugust 2005

7

Contents

Abstract................................................................................................................................... 5

A cknow ledgm ents .................................................................................................................. 7

Contents .................................................................................................................................. 9

List of Figures....................................................................................................................... 13

List of Tables ........................................................................................................................ 17

N om enclature........................................................................................................................ 19

Chapter 1: Introduction..................................................................................................... 21

1.1 Background and M otivation......................................................................... 21

1.2 Previous W ork .............................................................................................. 22

1.3 Research Objectives .................................................................................... 23

Chapter 2 Experim ental Apparatus and Procedures ........................................................ 25

2.1 Experim ental Setup ...................................................................................... 25

2.1.1 Engine and D ynam om eter ................................................................... 25

2.1.2 Engine Subsystem s and Instrum entation............................................. 26

2.1.3 Engine Control System .......................................................................... 27

2.1.4 Charge M otion Control V alve .............................................................. 27

2.1.5 In-cylinder Pressure m easurem ents ..................................................... 28

2.1.6 D ata A cquisition System ..................................................................... 28

2.1.7 Fast-Response Em issions Analyzers ................................................... 29

2.2 Experim ental Conditions and Procedure ..................................................... 31

2.2.1 Steady State Test Conditions and Procedure ........................................ 31

2.2.2 Start-up Test Conditions and Procedure............................................... 32

2.3 A nalysis M ethod............................................................................................ 34

2.3.1 Residual G as Fraction.......................................................................... 34

2.3.2 A ir/Fuel Ratio D eterm ination............................................................... 34

2.3.3 Burn Rate Analysis.............................................................................. 36

9

Chapter 3 : Steady State Experimental Results ................................................................. 37

3.1 Charge M otion Overview ............................................................................ 37

3.2 Combustion Characteristics.......................................................................... 38

3.2.1 Combustion Stability ............................................................................ 38

3.2.2 Bum Rate Analysis............................................................................... 39

3.3 Exhaust Emissions........................................................................................ 41

3.3.1 Exhaust Gas Temperature...................................................................... 41

3.3.2 Exhaust HC Emissions .......................................................................... 42

Chapter 4 : Engine Start-up Experimental Results.......................................................... 47

4.1 Engine Start-up Overview ............................................................................ 47

4.2 Comparison Results of Various CM CVs ..................................................... 49

4.2.1 CM CV Overview................................................................................. 49

4.2.2 M ixture Preparation...............................................................................50

4.2.3 Combustion Characteristics................................................................. 53

4.2.4 Exhaust Emissions................................................................................. 56

4.2.5 Flow Testing Results ............................................................................ 58

4.2.6 Overall Assessment of Different CM CVs ............................................. 60

4.3 Effects of Reduced Fuel Injection ................................................................. 61

4.3.1 M odification of Fuel Injection............................................................. 61

4.3.2 M ixture Preparation............................................................................... 61


4.3.4 Exhaust Emissions................................................................................. 65

4.4 Effects of Retarded Spark Timing ............................................................... 67

4.4.1 M odification of Spark Timing ............................................................. 67



4.4.4 Exhaust Emissions.................................................................................71

4.5 M odified Start-up Strategy with CM CV closed........................................... 73

4.5.1 M odification of Fuel Injection and Spark Timing................................ 73


10

4.5.3 Com bustion Characteristics................................................................... 75

4.5.4 Exhaust Em issions................................................................................ 79

Chapter 5 : Summary, Conclusions, and Recommendations ............................................ 81

5.1 Overview ...................................................................................................... 81

5.2 Steady State Experim ents .............................................................................. 81

5.3 Engine Start-up Experim ents....................................................................... 82

5.4 Overall A ssessm ent and Recom m endations................................................. 85

References ............................................................................................................................ 87

Appendix .............................................................................................................................. 89

11

List of Figures

Figure 2-1 Three different types of charge motion control valves: (a) Base CMCV<CMCV(B)>, (b) Tumble type CMCV <CMCV(T)>, (c) Swirl type CMCV<CMCV(S)>. Flow area was 25% for all three CMCVs....................................... 28

Figure 2-2 Schematic of fast-response FID sampling head from Cambustion. ....................... 29

Figure 2-3 Schematic of fast-response NDIR sampling head from Cambustion...................... 30Figure 3-1 Geometry of charge motion control valve. 25% notch area valve made of 1mm

alu m in u m p late . ......................................................................................................... 3 8Figure 3-2 COV of the Net-IMEP as a function of spark timing with the CMCV open and closed

at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 200 C coolant.................................................................................................................................... 3 9

Figure 3-3 0-10% and 10-90% combustion durations as a function of spark timing with theCMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuelratio w ith the 20 0C coolant. .................................................................................. 40

Figure 3-4 Location of 50% MFB as a function of spark timing with the CMCV open and closedat 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20'C coolant.................................................................................................................................... 4 0

Figure 3-5 Exhaust port gas temperature as a function of spark timing with the CMCV open andclosed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20'Cc o o la n t. ...................................................................................................................... 4 1

Figure 3-6 Exhaust port gas temperature as a function of 50% MFB location with the CMCVopen and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio withthe 20'C coolant................................................................................................... . . 42

Figure 3-7 Exhaust port HC concentration as a function of spark timing with the CMCV openand closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the2 00C co o lan t............................................................................................................... 4 3

Figure 3-8 Exhaust port HC concentration as a function of 50% MFB location with the CMCVopen and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio withth e 20'C coolant..................................................................................................... . 4 3

Figure 3-9 MAP and intake air flow as a function of spark timing with the CMCV open andclosed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20'Cc o o lan t. ...................................................................................................................... 4 4

Figure 3-10 Exhaust port HC mass flow rate as a function of spark timing with the CMCV openand closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the2 00C co o lan t............................................................................................................... 4 5

Figure 3-11 Exhaust port HC mass flow rate as a function of 50% MFB location with the CMCVopen and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio withthe 20'C coolant..................................................................................................... . 46

Figure 4-1 Baseline calibration MAP, RPM, spark timing, and relative air/fuel ratio as a functionof time after crank with the CM CV open. ............................................................. 48

13

Figure 4-2 Exhaust gas emissions measured at the exhaust port exit of cylinder #4 as a functionof time after crank with the CM CV open. ............................................................. 48

Figure 4-3 Three different types of charge motion control valves: (a) Base CMCV<CMCV(B)>, (b) Tumble type CMCV <CMCV(T)>, (c) Swirl type CMCV<CMCV(S)>. Flow area was 25% for all three CMCVs........................................ 50

Figure 4-4 In-cylinder k as a function of cycle number during the first 20 seconds followingen gine crankin g.................................................................................................... . . 5 1

Figure 4-5 Evaporated fuel as a function of cycle number during the transient period for variousC M C V s ...................................................................................................................... 5 2

Figure 4-6 Percentage of evaporated fuel to injected fuel as a function of cycle number duringthe quasi-steady period for various CM CVs.......................................................... 52

Figure 4-7 Net-IMEP as a function of cycle number during the transient period for variousC M C V s ...................................................................................................................... 5 3

Figure 4-8 0-10% and 10-90% combustion durations as a function of cycle number during thetransient period for various CM CV s....................................................................... 55

Figure 4-9 0-10% and 10-90% combustion durations as a function of cycle number during thequasi-steady period for various CM CVs. ............................................................... 55

Figure 4-10 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the transient period for various CMCVs. .......... 57

Figure 4-11 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the quasi-steady period for various CMCVs......... 57

Figure 4-12 Swirl index from flow bench meters as a function of normalized valve lift for variousC M C V s...................................................................................................................... 5 9

Figure 4-13 Tumble index from flow bench meters as a function of normalized valve lift forvarious C M C V s. .................................................................................................... . 59

Figure 4-14 Required mass of injected fuel as a function of cycle number for the same in-cylinderlambda trace with the CMCV open and closed during the first 20 seconds aftercranking com m enced. ............................................................................................ 62

Figure 4-15 Net-IMEP as a function of cycle number during the transient period for the CMCVopen case and the CMCV closed case with reduced fuel injection. ....................... 63

Figure 4-16 0-10% and 10-90% combustion durations as a function of cycle number during thetransient period for the CMCV open case and the CMCV closed case with reducedfu el injectio n .............................................................................................................. 64

Figure 4-17 0-10% and 10-90% combustion durations as a function of cycle number during thequasi-steady period for the CMCV open case and the CMCV closed case withreduced fuel injection. ............................................................................................ 64

Figure 4-18 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the transient period for the CMCV open case andthe CMCV closed case with reduced fuel injection............................................... 66

Figure 4-19 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the quasi-steady period for the CMCV open caseand the CMCV closed case with reduced fuel injection........................................ 66

Figure 4-20 Spark timing as a function of cycle number during the first 20 seconds followingengine cranking for various spark modifications with the CMCV closed............. 67

14

Figure 4-21 Required mass of injected fuel as a function of cycle number for the same in-cylinderlambda trace with various spark modifications during the first 20 seconds aftercranking com m enced. ............................................................................................ 68

Figure 4-22 Net-IMEP as a function of cycle number during the transient period for various sparkm odifications with the CM CV closed. ................................................................... 69

Figure 4-23 Location of 50% mass fraction burned as a function of cycle number during the 20seconds after engine cranking for various spark modifications with the CMCVc lo se d ......................................................................................................................... 7 0

Figure 4-24 0-10% and 10-90% combustion durations as a function of cycle number during thetransient period for various spark modifications with the CMCV closed.............. 70

Figure 4-25 0-10% and 10-90% combustion durations as a function of cycle number during thequasi-steady period for various spark modifications with the CMCV closed. .......... 71

Figure 4-26 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the transient period for various spark modificationsw ith the C M C V closed. ......................................................................................... 72

Figure 4-27 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the quasi-steady period for various sparkm odifications w ith the CM CV closed. ................................................................... 72

Figure 4-28 Injected fuel mass as a function of cycle number with the CMCV open and closedduring the first 20 seconds after cranking. Reduced sequential fuel injection strategywas applied to the CM CV closed case.................................................................. 74

Figure 4-29 Spark timing as a function of cycle number with the CMCV open and closed duringthe first 20 seconds after cranking. Retarded spark timing strategy was applied to theC M C V closed case. ............................................................................................. . . 74

Figure 4-30 In-cylinder k as a function of cycle number during the first 20 seconds followingengine cranking with the CMCV open and closed. Reduced sequential fuel injectionand retarded spark timing strategy was applied to the CMCV closed case........... 75

Figure 4-31 Net-IMEP as a function of cycle number during the transient period with the CMCVopen and closed. Reduced sequential fuel injection and retarded spark timing strategywas applied to the CM CV closed case.................................................................. 76

Figure 4-32 0-10% and 10-90% combustion durations as a function of cycle number during thetransient period with the CMCV open and closed. Reduced sequential fuel injectionand retarded spark timing strategy was applied to the CMCV closed case........... 77

Figure 4-33 Location of 50% mass fraction burned as a function of cycle number during thetransient period with the CMCV open and closed. Reduced sequential fuel injectionand retarded spark timing strategy was applied to the CMCV closed case........... 77

Figure 4-34 0-10% and 10-90% combustion durations as a function of cycle number during thequasi-steady period with the CMCV open and closed. Reduced sequential fuelinjection and retarded spark timing strategy was applied to the CMCV closed case. 78

Figure 4-35 Location of 50% mass fraction burned as a function of cycle number during thequasi-steady period with the CMCV open and closed. Reduced sequential fuelinjection and retarded spark timing strategy was applied to the CMCV closed case. 79

Figure 4-36 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the transient period with the CMCV open andclosed. Reduced sequential fuel injection and retarded spark timing strategy wasapplied to the CM CV closed case........................................................................... 80

15

Figure 4-37 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as afunction of cycle number during the quasi-steady period with the CMCV open andclosed. Reduced sequential fuel injection and retarded spark timing strategy wasapplied to the CM CV closed case......................................................................... 80

16

List of Tables

Table 2-1 Engine specifications.............................................................................................. 25Table 2-2 Unleaded test gasoline (Indolene, UTG-91) properties [14]. ................................ 32

17

Nomenclature

ABDC after bottom dead centerATDC after top dead centerBBDC before bottom dead centerBDC bottom dead centerBTDC before top dead centerCA crank angleCMCV charge motion control valveCOV coefficient of variationDAQ data acquisitionDOHC dual overhead camshaftECM engine control moduleEGR exhaust gas recirculationEVC exhaust valve closingEVO exhaust valve openingFFID fast flame ionization detectorFID flame ionization detectorFTP federal test procedureHC hydrocarbonIAC idle air controlIMEP indicated mean effective pressureIVC intake valve closingIVO intake valve openingMAP manifold air pressureMDS modular development systemMFB mass fraction burnedNDIR non-dispersive infraredPVC positive crankcase ventilationRPM revolutions per minuteSI spark ignitionTDC top dead centerUEGO universal exhaust gas oxygen

19

Chapter 1

Introduction

1.1 Background and Motivation

Meeting future, stringent emissions standards will require precise engine control

strategies that significantly improve the emissions performance under transient conditions.

In particular, greater emphasis is being put upon the cold start-up process where a signifi-

cant fraction of the total hydrocarbon (HC) emissions can be produced. The use of catalytic

converters for aftertreatment is an effective means of reducing tailpipe-out HC emissions

from spark ignition (SI) engines. However, the majority of the total HC emissions during

the Federal Test Procedure (FTP) are emitted during the first few seconds of engine

operation before the catalyst reaches light-off temperature, approximately 2500 C. Due to

the fact that catalytic converters are essentially ineffective until light-off temperatures are

achieved, increased importance is being placed on understanding and improving engine-out

HC emissions during the first few seconds of engine operation after cranking.

There are various HC emissions formation mechanisms for SI engines: the filling of

crevice volumes with unburned mixture, flame quenching at the combustion chamber walls,

oil layer absorption and desorption, incomplete combustion, liquid fuel wetting of in-

cylinder surface, deposit absorption and desorption, and air/fuel ratio abnormality [1,2].

During the engine start-up process, these mechanisms play increased roles in determining

the engine-out HC emissions. Of particular importance is the influence of the mixture

preparation process which can greatly affect these mechanisms. However, it is difficult to

achieve appropriate mixture preparation in the cold start-up process due to the unfavorable

fuel transportation and evaporation environment such as low port air velocity, initially at-

mospheric port pressure, and low port and cylinder wall temperatures.

21

Since only a small portion of injected fuel vaporizes and contributes to the combusti-

ble mixture in a cold engine [3,4], the cold start-up process of port fuel injected SI engines

generally involves significant over-fueling in the intake manifold. A substantial amount of

fuel is injected to overcome unexpected transients as well as poor mixture preparation. This

fuel enrichment provides a sufficiently large amount of vaporized fuel so that smooth and

robust start-up can be assured. A large portion of this excess fuel, however, does not vapor-

ize and enters the cylinder as liquid [5]. This fuel may be stored in various places in the cyl-

inder (e.g. deposits, oil layers on the cylinder liner, piston top, and cylinder head, or com-

bustion chamber crevices), survive the subsequent combustion process unburned, and come

out into the bulk exhaust, where it would significantly contribute to the exhaust HC emis-

sions.

The complex nature of these start-up behaviors requires an improvement in mixture

preparation for robust combustion with low HC emissions. The mixture preparation process

in contemporary port fuel injected SI engines is significantly affected by intake charge mo-

tion, especially during the engine start-up process. It is well known that the in-cylinder

charge motion is one of the major factors that controls the air-fuel mixing and combustion

process. Since the initial in-cylinder flow pattern is set up by the intake process, enhanced

intake charge motion is essential for improving mixture preparation and combustion char-

acteristics, which determine engine-out HC emissions behavior. There are several methods

of enhancing charge motion: intake manifold modification, intake valve or cylinder head

masking, and charge motion control valves (CMCV). The charge motion control valve is

one of the most cost-effective methods for a high level of in-cylinder charge motion.

1.2 Previous Work

Over the past few decades, great efforts have been placed on reducing vehicle emis-

sions. With the development of catalytic converters, the HC emissions in the start-up proc-

ess of SI engines have become a significant part of the total emissions. A number of studies

have been done to investigate the various aspects of the start-up process and reduce the HC

emissions with robust combustion. Mixture preparation during the start-up process has been

22

the focus of previous research. Most of previous studies have concentrated on the effects of

fuel metering and liquid fuel behavior on mixture preparation and HC emissions.

In a previous study performed by Henein et al. [6], the effect of engine transients on

HC emissions was investigated. The initial piston position was found to affect the HC

emissions during the cold start-up process. The first cycle fuel metering effect on HC emis-

sions was examined by Castaing et a]. [7]. The open valve injection was detrimental to the

mixture preparation process during the start-up because it caused mixture non-uniformity,

directly deposited fuel in the cylinder which increased engine-out HC emissions. There

have been some studies investigating the effect of enhanced charge motion on engine per-

formance. In a study conducted by Takahashi et al. [8], the use of the CMCV increased the

bum rate and combustion stability. More improvement in combustion stability was ob-

tained by optimizing the geometry of the CMCV. Similar research by Nishizawa et al. [9]

showed that the cold start-up HC emissions were reduced by an electronically actuated

CMCV with a high-speed starter and a fast light-off catalyst. Jung et a]. [10] showed some

benefits of the CMCV in the fuel consumption and cam retard at light load conditions. At

medium and high load conditions, the CMCV caused an excessive pumping work penalty

due to the flow restriction.

1.3 Research Objectives

There has been little study trying to examine the real engine start-up process with the

CMCV. Most of the previous experimental studies on enhancing the charge motion for the

cold start-up process have been performed under the simulated steady state conditions due

to the difficulty of engine control. Therefore, it would be worth to investigate the effects of

enhanced charge motion on mixture preparation, combustion characteristics, and HC emis-

sions.

The objective of this research is to quantify the effects of various intake charge mo-

tion control approaches on the cold start-up process of a port fuel injected SI engine. Four

main tasks were carried out:

23

* Develop diagnostics to examine the impacts of controlled charge motions on

mixture and combustion.

* Identify desirable flow patterns for cold start-up by examining various charge

motion control valves.

* Assess the effectiveness of these changes in charge motion on mixture prepara-

tion, combustion, and emissions.

* Develop recommendations for improving the charge motion during the engine

start-up process.

24

Chapter 2

Experimental Apparatus and Procedures

2.1 Experimental Setup

2.1.1 Engine and Dynamometer

Experiments were performed on a 2003 model year, 2.2 liter L61 Ecotec (GMX357)

engine to investigate the effects of various intake charge motion control approaches on the

cold start-up process of a port fuel injected SI engine. The engine was coupled to an eddy

current, absorbing only, dynamometer (Froude Consine AE-80). This test setup used the

engine starter motor for cranking operation with full control over the engine control mod-

ules (ECM). Additional engine specifications are shown in Table 2-1.

Table 2-1 Engine specifications

Ecotec 2.2-liter, 4-cylin er Engine SpecificationsType Naturally Aspirated, Line 4Displacement Volume (cc) 2189Clearance Volume (cc) 65Bore (mm) 86Stroke (mm) 94.6Connecting Rod (mm) 146.5Wrist Pin Offset (mm) 0.8Compression Ratio 10.0:1Firing Order 1-3-4-2Valve Configuration 16v DOHC

Valve Timing IVO: 7 BTDC I IVC: 560 ABDCEVO: 49 BBDC I EVC: 19 ATDC

Fuel System Sequential Fuel Injection

25

2.1.2 Engine Subsystems and Instrumentation

Air mass flow rate was measured by a thermal mass flow meter (EPI, Series 8000MP).

Throttle position control was controlled by a stepper motor (Pacific Scientific SinMax 1.8'

motor with 5230 indexer/driver) connected via a cable to the engine throttle body. Incre-

mental movement of the throttle was adjusted by two momentary contact switches with a

selectable stepping rate. The positive crankcase ventilation (PCV) system was utilized dur-

ing the engine operation, but the exhaust gas recirculation (EGR) and secondary air injec-

tion system were not used for these experiments.

Fuel was supplied by external fuel pump with an inline filter, accumulator, and pres-

sure regulator. The regulator outlet was connected to the fuel injector rail. A constant fuel

rail pressure of 52 psig was maintained for a returnless fuel injection system. Injected fuel

mass was calculated from the fuel pulse width duration and pressure differential across the

injector.

The coolant temperature was controlled by an external heat exchanger. The engine

coolant system was a closed looped system driven by an internal water pump that circulated

fluid through the block and cylinder head. The thermostat was removed and the heater core

was modified to eliminate the recirculation of coolant within the cylinder head and block.

Coolant exited the head and flowed to an external water pump installed in the engine's

cooling circuit. The pump outlet was connected to an external heat exchanger and to a cool-

ant reservoir tank before being routed back to the engine. Coolant temperature was varied

via a set-point controller that actuated a valve allowing for plant water to flow through an

external heat exchanger. The engine oil was not externally cooled and its temperature was

measured in the oil pan.

Time-resolved exhaust gas temperatures were obtained from chromel-alumel (type-K)

exposed junction thermocouples (0.8-mm bead diameter) with custom radiation shielding.

Exhaust air/fuel ratio was measured by a universal exhaust gas oxygen (UEGO) sensor lo-

cated at the exhaust manifold collector. A Horiba MEXA-700 analyzer interpreted the sig-

nal and displayed the air/fuel equivalence ratio.

26

2.1.3 Engine Control System

The Delco Modular Development System (MDS) [11] was used to control and modify

the engine control module (ECM). Specific ECM operations were monitored and recorded

by the MDS. This system provided the user control of parameters including the change of

calibration data by read only memory (ROM) emulation. A PC was connected to the MDS

stack that allowed for internal and external data logging. The stack was composed of sev-

eral units. The main instrument unit (MIU) was the core of the MDS and contained two

embedded processors: the main instrument processor (MIP) and the input output processor

(IOP). The computer interface buffer internal logging (CIBIL) provided MDS communica-

tion with a computer. The analog conversion module (ACM) contained 8-BNC outputs and

2 instantaneous switches scaled from 0 to 5 volts. A 1MB GMPX Pod was connected to the

X-pod that interfaced with the ECM and allowed the ECM's EPROM to be flashed for

standalone ECM operation. The shell program allowed the user to have full access to

lookup tables and relevant environmental variables. Real-time monitoring and modification

of several parameters (RPM, MAP, spark timing, air/fuel ratio, and idle air control (IAC)

valve position) was accomplished by using the MDS's external display unit (DU). Full con-

trol of the engine was possible using the Delphi Electronics Instrument Tool Suite (ITS)

running the Saturn Legacy Software. File handling between the computer and MDS unit

was achieved using the ITS software. The program CalTools was used to modify ECM pa-

rameters and lookup table variables.

2.1.4 Charge Motion Control Valve

The charge motion control valve (CMCV) device was installed between the intake

manifold and intake port. The asymmetrical CMCV (base CMCV) was used for the steady

state experiment as shown in Figure 2-1. Three different shapes of charge motion control

valves were investigated for the engine start-up experiment in order to better understand the

effects of the enhanced charge motion on mixture preparation, combustion, and emissions.

All three different CMCVs were made of 1 mm aluminum plate and reduced the port cross

sectional area by 75% (i.e. 25% notch area). Since the CMCV device was located upstream

of the fuel injector, it did not impede the fuel spray.

27

(a) Base CMCV (b) Tumble type CMCV (c) Swirl type CMCV

Figure 2-1 Three different types of charge motion control valves: (a) Base CMCV <CMCV(B)>, (b)Tumble type CMCV <CMCV(T)>, (c) Swirl type CMCV <CMCV(S)>. Flow area was 25%for all three CMCVs.

2.1.5 In-cylinder Pressure measurements

Each cylinder was equipped with a flush-mounted piezoelectric pressure transducer

(Kistler 6125A). The cylinder head passage sleeve incorporated an eight-hole flame arrestor

to minimize the occurrence of thermal shock. The current signal from the transducer was

converted to a voltage signal by a charge amplifier (Kistler 5010B). The transducers were

statically calibrated at several pressure points using a dead weight tester. For pressure refer-

encing, an absolute pressure transducer (OMEGA PX-176) was installed in the intake

manifold runner that provided better transient response compared to the OEM MAP trans-

ducer. In-cylinder pressure was referenced on an individual cycle basis over a 20 CA inter-

val centered around BDC of the intake event, 1700 ATDC to 190' ATDC, using the intake

manifold runner pressure.

2.1.6 Data Acquisition System

The data acquisition (DAQ) system utilized two DAQ boards from National Instru-

ments installed in a personal computer. The first card (PCI-6071E) acquired 32 differential

channels of high-speed signals. The second card (PCI-6024E) was used with a multiplexing

chassis (SCXI-1000) and a 32 channel thermocouple module (SCXI- 1102) to acquire tem-

perature data from type-K thermocouples. The two boards were operated at different

speeds; high-speed data was captured with a frequency of 40000hz (once per 0.025ms),

while temperature data was acquired with a frequency of 4000hz (once per 0.25ms). Both

28

DAQ cards were triggered frOln signals provided by an incremental encoder (BEl Series

H25E) coupled to the engine crankshaft. LabView software was used to create a virtual in

strument for data scaling, processing, and logging.

2.1.7 Fast-Response Emissions Analyzers

• Fast-Response He Analyzer

Time-resolved HC concentrations were measured via a Cambustion HFR400 fast

response flame ionization detector (FID). The FID is the industry standard method of

measuring HC concentration. The fast-response FID consists of a main control unit (MCU)

and two remote sampling heads. The sample gas is introduced into a hydrogen flame inside

the FID. Any hydrocarbons in the sample will produce ions when they are burned. Ions are

detected using a Inetal collector which is biased with a high DC voltage. The current across

this collector is thus proportional to the rate of ionization which in tum depends upon the

concentration of HC in the salnple gas. This fast-response FID has a 10-90% response time

of approximately 1 millisecond. A heated transfer sampling line (TSL-H) with a hole di

ameter of 0.026" was placed at the exhaust port exit, 7 em from the exhaust valve seats.

The fast-response FID was calibrated before and after each experiment using 1500ppmC3

propane gas (span) and nitrogen gas (zero).

Fast Response FID

High voltage __ ion collector

NozzLe ---

SampLe capillary

-----~ Vacuum

Rame chamber

_____ <f-- Air ______ <f--

Fuel gas

Figure 2-2 Schematic of fast-response FID sampling head from Cambustion [12].

29

Fast-Response CO and CO2 Analyzer

Time-resolved CO and CO2 concentrations were measured VIa a Cambustion

NDIR500 fast-response non-dispersive infrared (NDIR) detector. The NDIR is the industry

standard method of measuring engine exhaust CO and CO2 emissions. The instrument is a

two-channel analyzer with the detectors located in remote sample heads. It measures CO

and CO2 simultaneously from one sample head, providing four streams of data per analyzer.

Satnple gas from the engine is conveyed through narrow heated capillaries directly into the

sample chamber. The optical filters are mounted on a chopping wheel; reference and blank

sectors on the chopping wheel correct for minute changes in temperature and IR emitter

performance. The instrument gives a T lO- 90 response time of approximately 8 milliseconds.

The main control unit is housed in a mobile cabinet with the vacuum pump located in the

base. The analyzer is controlled via RS232 or RS485 serial communications from a laptop

computer with the installed operating software. The interface provides full control of the

instrument including auto-calibration, start-up, self-test, and trouble-shooting and help

functions with video assistance. A heated sample probe was located at the exhaust port exit,

7 CIn from the exhaust valve seats. Recommended cleaning process was conducted before

and after each experiment to remove accumulations from optical windows.

Temperature controlled

Temperature Controlled Ext.maIC~ Heatstnk

Pressure controUed dlamber

~

1

AddItional volume can be added here for in-cylinder

sampling.

-----++To vacuum pump

Figure 2-3 Schematic of fast-response NDIR sampling head from Cambustion 113J.

30

2.2 Experimental Conditions and Procedure

2.2.1 Steady State Test Conditions and Procedure

Steady state cold-fluid experiments were performed to understand the basic effects of

the CMCV on mixture, combustion, and emissions. In order to simulate the park idle condi-

tion, the engine was operated at the condition of 1000 rpm, 2.5 bar Net-IMEP (net indicated

mean effective pressure) with the base CMCV open and closed. Spark timing sweeps were

conducted with a fixed coolant temperature of 20'C and a stoichiometric air/fuel ratio. The

coolant temperature was regulated by an external heat exchanger. The exhaust system,

however, was not externally cooled and was allowed to reach a hot stabilized temperature

for steady state experiments. A set of 300 cycle data was recorded for each operating condi-

tion. All experiments were performed with unleaded test gasoline (UTG), commonly

termed "indolene". This is a standardized fuel type typically used for emissions certifica-

tion and laboratory work. UTG-91 from Chevron Phillips Chemical Company LLC was

used for this work to represent the performance of gasoline. The relevant specification of

this fuel is in Table 2-2.

31

Table 2-2 Unleaded test gasoline (Indolene, UTG-91) properties [141.

UTG-91Property Typical Value SpecificationSpecific Gravity at 60160 *F 0.7350 0.7343 - 0.7440API Gravity 61.0 ReportCopper Corrosion, 3 h at 50 'C 1 1 maxExistent Gum (washed), mgi100 mL 2 5 maxSulfur, ppm 130 1000 maxReid Vapor Pressure, psia 9.0 8.8- 92Lead, gtgal 0.0010 0.0050 maxPhosphorus, gfgal 0.001 0.002 maxHydrogen, wt % 13.7 ReportCaibon, wt % 86.3 ReportCabon Density, g/gal ReportDistilation Range at 760 mmHg, *F

Initial Boling Point 88 75-9510% 122 120-13550% 212 200-23090% 321 300-325End Point 399 415 max

Oxidation Stabdity, min > 1440 1440 minHeat of Combustion, Net, Btuflb 18500 ReportComposition, vol %

Aromatics 24.0 35.0 maxOlefins 6.0 10- maxSaturates 70.0 Report

Research Octane Number 90.8 90.3- 91.7Motor Octane Number 82.8 ReportAnti-Knock Index, (R+MY2 86-8 87.0 maxSensitivity 7.8 7.5 minOxygenates, vol % 0.0 maxBenzene, vol % Report

Test MethodASTM D 4052ASTM D 1250ASTM D 130ASTM D 381ASTM D 5453ASTM D 5191ICPIOESICPIOESASTM D 5291ASTM D 5291CalculatedASTM D 86

ASTM D 525ASTM D 240ASTM D 1319

ASTM D 2699ASTM D 2700CalculatedCalculatedChromatographyChromatography

2.2.2 Start-up Test Conditions and Procedure

Engine start-up experiments were conducted with the three different geometric charge

motion control valves. Mixture preparation, combustion characteristics, and HC emissions

were evaluated with the various fuel injection and spark timing strategies during the first 20

seconds following engine cranking. General Motors provided the target speed-load point

during the park idle period of the FTP (0-15 seconds) for this engine. The baseline calibra-

tion produced a Gross-IMEP of 3.0 bar and 950 RPM with 0.40 bar MAP and 60BTDC

spark timing. Therefore, the engine was calibrated using the MDS hardware and CalTools

software to achieve the target operation condition for the park idle period, 2.5 bar Net-

IMEP and 950 RPM. All engine start-up tests were conducted with UTG-91 (Table 2-2) at

approximately 20C environmental conditions. The coolant temperature was initially fixed

at 20C by the external coolant control system.

32

After each start-up test, the engine was operated and warmed up to a coolant tempera-

ture of 70 0C to avoid residual fuel build-up in the intake port [15]. After shutdown, the en-

gine was purged by cranking with the injector off for 1 minute. Metal temperatures

throughout the engine and exhaust system, then, were force-cooled to ambient temperatures

before another experiment was conducted.

Several previous studies showed that the engine start-up behavior is affected by the

initial phasing of the piston positions [6,7]. For a 4-cylinder 4-stroke engine, the engine

tends to stop with one of the pistons at mid stroke (~90' from the BDC) of compression.

Therefore, the piston starting point was fixed at 90 'ATDC of the expansion stroke of cylin-

der #4 to avoid excessive complexity in data analysis.

Engine speed and load were not regulated by the dynamometer controller (Digalog

Series 1022A) due to the unstable transient control of the dyno-engine system during start-

up RPM flare. The required engine idle load was achieved by utilizing engine accessories

with the dynamometer coupled but not absorbing power from the engine. A hydraulic

power steering pump and alternator were added to the engine setup to simulate park idle

load conditions observed during the first 15 seconds of the FTP. The power steering pump

was throttled to 800 psig by the use of a needle valve. The low pressure line was routed

through a heat exchanger before returning oil to the reservoir tank. The regulated 14.7 volt

output from the alternator was isolated from the 12 volt battery bus by a zener diode and

connected to bank of power watt resistors, which had total 0.2 Ohm (three 0.6 Ohm resis-

tors in parallel).

33

2.3 Analysis Method

2.3.1 Residual Gas Fraction

To calculate in-cylinder air/fuel ratio and engine-out HC mass emissions, it is crucial

to be able to estimate the cycle-by-cycle composition of the combustion chamber charge.

This charge consists of fresh air, fuel, and residual gas from the previous cycle. The con-

ventional SI combustion process is largely affected by the residual gas fraction. Residual

gas in SI engines has a profound effect on emissions, performance, and combustion stabil-

ity. Residual gas affects the combustion process through its influence on charge mass, dilu-

tion, temperature, and flame speed. Residual gas influences combustion mainly by acting as

a diluent which decreases the flame speed and temperature of the resulting charge. These

effects are especially important for engine idle stability and for HC emissions. Also,

interpreting experimental pressure data through chemical energy release and exhaust mass

flow analysis requires that the residual gas mass fraction be known. Fox model [16] was

used to estimate the residual gas fraction during the engine start-up process since the model

is applicable for low to medium engine speeds such that the cylinder pressure does not

substantially differ from the exhaust port pressure at IVO. In the model, the contribution of

the back flow from the exhaust port to the cylinder during the valve overlap period is ac-

counted for explicitly in terms of a valve overlap factor. For more detailed information re-

garding the residual gas fraction refer to the Section A.2 in the Appendix.

OF P --0.87 4

x, =1.266 -P - + 32_.-N P r

2.3.2 Air/Fuel Ratio Determination

Exhaust gas composition depends on the relative proportions of fuel and air fed to the

engine, fuel composition, and completeness of combustion. These relationships can be used

to determine the operating air/fuel equivalence ratio of an engine from a knowledge of its

exhaust gas composition. A general formula for the composition of fuel can be represented

as CnHmOr. For conventional petroleum-based fuels, oxygen will be absent. Since Indolene

34

(UTG-91) also has no oxygen in its composition, CHy (c.f. y = m/n) was used for the fol-

lowing combustion reactions. The overall combustion reaction can be written explicitly as:

CH +n, (0, +3.773- N2)

+ n,, -{i' q C02 + ' o -CO+ Xc's - CH, + X' o -H2 0+ X' o - NO+ X' H + X' -N2 + X' -02}

> n { Xc - C2 + cCO HCH+X -CH,+ H 2 O+ NO . NO+XH, - H+ - N -}~nPxC@~±xOCOxCHCH~H 2 0II2+ XNNO± 2H H2 +N, 2 + @ 2

no : moles of 02 per mole CHy

nr moles of residual gas per mole CHy

nP moles of exhaust products per mole CHy

' mole fraction of residual gas component

X : mole fraction of exhaust gas component

From the exhaust concentrations of HC, CO, and C0 2, a carbon balance air/fuel ratio

was derived. Fox model was combined with this combustion equation in order to provide

the mole fraction of residual gas. Throughout the calculation, it has been assumed that the

unburned hydrocarbons have the same H/C ratio as the fuel. Since NOx concentrations are

usually sufficiently low during the cold start-up process, NO and NO2 were negligible for

this air/fuel ratio evaluation.

A Mar [W +C'W- CW' y]- I-+1 - n

F Mfuel C, CT 2

CT XC + K0O + XCH

W =- ic+ -XH20 +1CH, W= 'o + I' + 1's2 22 n2 2 H 2 )

K = XCOXHO CO XXC H XH 2

Mair, M fuel: molecular weight

35

2.3.3 Bum Rate Analysis

The burn rate analysis was performed using the in-cylinder pressure data acquired

from cylinder #4. A single-zone thermodynamic model [18] incorporating the effects of

residual fraction, heat transfer, and crevices was used for the burn rate analysis in order to

quantify combustion characteristics with the CMCV open and closed. Fox model was used

for the residual gas fraction in this model, heat transfer was handled by the Woschni corre-

lation, and total crevice volume was estimated to be equal to 2% of the clearance volume.

QCH - Y -1 V + 1 VP + CR HT

ao Y ao Y -I ao ao ao

where:

QcH Fuel chemical energy

0 Crank angle

y Ratio of specific heat

p Cylinder pressure

V Cylinder volume

QCR Energy in crevices

QHT Heat transfer

36

Chapter 3

Steady State

Experimental Results

3.1 Charge Motion OverviewGas motion into and within the engine cylinder is one of the major factors that con-

trols the air-fuel mixing and combustion process. It also has a significant impact on heat

transfer. Since the initial in-cylinder flow pattern is set up by the intake process, the intake

charge motion plays a critical role in enhancing both the air-fuel mixing and combustion

process. During the engine start-up process, it is difficult to achieve an appropriate in-

cylinder mixture for robust combustion due to the unfavorable mixture preparation envi-

ronment such as cold port wall temperatures, initially atmospheric port pressure, and low

air velocity. Inappropriate mixture preparation can also cause a high level of HC emissions

and a slow combustion process, which decreases combustion stability. Enhanced in-

cylinder charge motion, therefore, is essential for mixture preparation, combustion charac-

teristics, and HC emissions during the cold start-up.

Steady state cold-fluid testing was performed to understand the basic effects of the

CMCV on mixture, combustion, and emissions. Various parameters were measured at the

condition of 1000 rpm, 2.5 bar Net-IMEP. This speed-load point represents the park idle

operating condition for this engine.

An asymmetrical charge motion control valve (CMCV), shown in Figure 3-1, was

used to promote more rapid air-fuel mixing and also speed up the combustion process in

this steady state experiment. The charge motion control valve was installed between the

intake port and the intake manifold. This CMCV was made of 1 mm aluminum plate and

reduced the port cross sectional area by 75%.

37

(a) Base CMCV

Figure 3-1 Geometry of charge motion control valve. 25% notch area valve made of 1mm aluminumplate.

3.2 Combustion Characteristics

3.2.1 Combustion Stability

Combustion stability was quantified and investigated in terms of the coefficient of

variation (COV) in net indicated mean effective pressure (Net-IMEP). Cycle-by-cycle

combustion variations in the combustion process are caused by several factors: variation in

gas motion in the cylinder during combustion, variation in the amounts of fuel, air, and re-

sidual gas supplied to a given cylinder each cycle, and variation in mixture composition

within the cylinder each cycle [1]. Net-IMEP and COV were calculated from in-cylinder

pressure data acquired from cylinder #4. A set of 300 cycle data was analyzed for each op-

erating condition. Figure 3-2 shows the COV of the Net-IMEP as a function of spark timing

with the CMCV open and closed. With increased spark retardation, combustion stability

was found to worsen as combustion occurred in a rapidly expanding cylinder volume.

However, significant improvements in combustion stability were achieved with the CMCV

closed at late spark timings.

38

9

8

-. 6 ~

[;; 5 ::;; Z

> 4 o U

3

2

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

35 30 25 20 15 10 o -5 -10 -15

Spark Timing (OBTDC)

Figure 3-2 COY of the Net-IMEP as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

3.2.2 Bum Rate Analysis

A single-zone thermodynamic bum rate analysis incorporating the effects of heat

transfer, residuals, and crevices was used to quantify combustion characteristics with the

CMCV open and closed case. Increased charge motion with the CMCV closed was found

to shorten the 0-100/0 and 10-90% combustion duration as shown in Figure 3-3. Figure 3-4

shows that the reduced combustion duration causes 50% mass fraction burned to occur up

to 5° CA earlier for the same spark timing.

39

38

......... 36 0) Q)

"'0 34 « S 32 c 0 .~ 30

~ 0 28 c 0

26 ~ ::l .0 24 E 0 U 22 cF. 0 20 r-

6 18

2.5 bar NIMEP, 1000 RPM, 20°C, 1..=1.0

0 open010 .. ciosed010 38 o open1090

.. closed1090 0) 36 r-------.

0-10%

0 0 0 o A-A- 0

A- 0 0 o A-A A

A A

0

0 A-

A-

Q)

"'0 34 o « u A ~ 32

o .~ 30

~ o 28 c 0 .~ 26 ::l .0 E 24 0 U

22 cF. 0 en 20 6

18

353025 20 15 10 5 0 -5 -10-15

Spark Timing (oBTDC)

10-90%

o

o

o A

o A-

0

A o A-

0 0 A A A

35 30 25 20 15 10 5 0 -5 -10-15

Spark Timing (BTDC)

Figure 3-3 0-10% and 10-90% combustion durations as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

2.5 bar NIMEP, 1000 RPM, 20°C, 1..=1.0

70

60

U 0 l- SO « 'L.. 0) 40 LL ~

"* 30 0 I.()

'0 c 20 .9 co u 10 0 -l

0

35 30 25 20 15 10 5 o -5 -10 -15

Spark Timing (BTDC)

Figure 3-4 Location of 50% MFB as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

40

3.3 Exhaust Emissions

3.3.1 Exhaust Gas Temperature

The cylinder #4 port exit exhaust gas temperatures were measured with the CMCV

open and closed at various spark timings as shown in Figure 3-5. The CMCV decreased the

exhaust gas temperature between 5-30°C at a given spark timing, which was attributed to

increased in-cylinder heat transfer during combustion. The exhaust gas temperatures were

observed to increase with the spark retardation for both the CMCV open and closed case

due to the fact that the burned gas is not as fully expanded and does not perform as much

work on the piston.

800

U 750 'L.

~ 700 co Ci3

E 650 Cl)

I-Vl co C)

~ 0...

tl :J co ..c x

llJ

600

550

500

450

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

35 30 25 20 15 10 5 0 -5 -10 -15


Figure 3-5 Exhaust port gas temperature as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

41

Since the location of 50% mass fraction burned gives a better indication of when

combustion is occurring, it can be used as the indicator of combustion phasing. Figure 3-6

shows that exhaust gas temperatures increased linearly with combustion phasing. At an

equivalent phasing, there is a close agreement between the CMCV open and closed exhaust

gas temperatures.

800

E 750

Q)

:s 700 ro Cii E 650 Q)

I-(f) C'Q

<...?

0 CL t=) :::J C'Q ..c x

l!J

600

550

500

450

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

----0--- CMCV open --A- CMCV closed

-5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75

Location of 50% MFB (OATDC)

Figure 3-6 Exhaust port gas temperature as a function of 50% MFB location with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

3.3.2 Exhaust He Emissions

Figure 3-7 shows the exhaust HC concentrations with the CMCV open and closed for

various spark timings. HC concentrations decreased at retarded spark timings mainly due to

the reduced crevice volume fraction at peak pressure and increased burned gas temperature

at the end of the expansion and during exhaust, which increased oxidation. The exhaust HC

concentrations were also investigated as a function of combustion phasing (location of 500/0

MFB). Figure 3-8 shows that the He concentrations of the CMCV open and closed cases

are almost coincident at a given combustion phasing.

42

3500

3000

U~

E 2500 a.. S c 0 2000 .~

C Q) u c 1500 0 U U 1000 I

500

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

~~~ ~~~g~io~~d o~

\)~

\':\ \:~

'\o~ " ...

"0

35 30 25 20 15 10 5 0 -5 -10 -15


Figure 3-7 Exhaust port HC concentration as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20GC coolant.

u~

E a.. S c 0 .~

C Q) u c 0 U U I

3500

3000

2500

2000

1500

1000

500

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

'0

-5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75

Location of 50% MFB (OATDC)

Figure 3-8 Exhaust port HC concentration as a function of 50% MFB location with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20GC coolant.

43

To achieve the same engine load (Net-IMEP) with late spark timings, the engine's

mass flow rate was increased to offset the reduction in work extracted from the burned gas.

Therefore, the meaningful representation of the HC emissions is the HC mass flow rate in

the steady state. Changes in exhaust mass flow rates were addressed to better interpret the

results from the HC concentration data as shown in Figure 3-9.

0.65

0.60

"2 0.55 CIJ e 0.50

~ 0.45

~ 0.40

0.35

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

35 30 25 20 15 10 5 o ·5 -10 -15

11 ,--1 ______ ---.

Vl 109 r-I ---B--- CMCVopen I :9 ~ CMCV closed S o

u::: 8 f-

6

35 30 25 20 15 10 5 o -5 -10 -15

Spark Timing (OBTOC)

Figure 3-9 MAP and intake air flow as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric airlfuel ratio with the 20°C coolant.

44

Figure 3-10 shows the exhaust HC mass flow rate with the CMCV open and closed

for various spark timings. As combustion was phased earlier due to faster burning with the

enhanced CMCV generated charge motion at a given spark timing, higher in-cylinder peak

combustion pressures occurred, which would increase the mass fraction of the HC trapped

in crevice volumes that escaped the oxidation during flame propagation. The important

crevice processes occurring during the engine cycle are the following. As the cylinder pres

sure rises during compression, air-fuel mixture is forced into each crevice region. Since

these volumes are thin, they have a large surface/volume ratio; the gas flowing into the

crevice cools by heat transfer to close to the wall temperature. During combustion while the

pressure continues to rise, air-fuel mixture, depending on engine type, continues to flow

into these crevice volumes. After flame arrival at the crevice entrance, burned gases will

flow into each crevice until the cylinder pressure starts to decrease. Once the crevice gas

pressure is higher than the cylinder pressure, gas flows back from each crevice into the cyl

inder.

u Q) (f)

10

0, 8 5 (f)

c o .~ 6

'E w u I

2.5 bar NIMEP. 1000 RPM. 20°C. 1..=1.0

35 30 25 20 15 10

I---B--- CMCV open I . ............... CMCV closed I

o ·5 ·10 ·15

Spark Timing tSTDC)

Figure 3-10 Exhaust port HC mass flow rate as a function of spark timing with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric air/fuel ratio with the 20°C coolant.

45

The exhaust HC mass flow rates were re-plotted in Figure 3-11 at equivalent levels of

combustion phasing (location of 50% MFB). The line for the CMCV closed case was al

most coincident with the line for the CMCV open case. This indicates that the CMCV gives

no noticeable advantage on the exhaust HC emissions at an equivalent combustion phasing

in the steady state park idle operation.

U ell V1

10

rn 8 5

V1 c o "~ 6

"E L.1J

u I

2.5 bar NIMEP, 1000 RPM, 20°C, A=1.0

---B--- CMCV open --A- CMCV closed

'0

2~~~~~~~~~~~~~~~~~

-5 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75

Location of 50% MFB (OAT DC)

Figure 3-11 Exhaust port HC mass flow rate as a function of 50% MFB location with the CMCV open and closed at 2.5 bar NIMEP, 1000 RPM, and stoichiometric airlfuel ratio with the 20°C coolant.

46

Chapter 4

Engine Start-up

Experimental Results

4.1 Engine Start-up Overview

Ambient engine start-up is a complicated process that requires delivery of an adequate

relative air/fuel ratio around the spark plug gap for a robust combustion event during engine

cranking. The engine undergoes severe transients before it reaches the idle speed. Upon the

first cylinder firing, the engine accelerates to a speed up to about 1600 RPM, resulting in a

rapid decrease in the intake manifold pressure. During the first part of this transient period,

spark timing is advanced as engine speed and in-cylinder exhaust gas residuals increase.

The engine then decelerates to normal idle speeds during the second part of this transient

period; therefore, the intake manifold pressure and spark timing also transition to the idle

calibration state. In order to better understand the typical engine start-up behavior, several

operating parameters and exhaust emissions were acquired during the first 20 seconds after

cranking commenced as shown in Figure 4-1 and Figure 4-2. A substantial amount of fuel

was injected into cylinder #4 for the first few cycles to prepare the appropriate in-cylinder

mixture for robust combustion. Significantly rich in-cylinder mixture conditions and high

transient HC concentrations were also observed during the first few seconds.

The critical processes will be discussed in the following sequence: mixture prepara-

tion, combustion, and HC emissions. This is because the intended use of the CMCV is to

reduce HC emissions by improving mixture preparation and combustion characteristics.

However, the mixture preparation analysis used the engine-out exhaust emissions data as

shown in Section 2.3.2.

47

1.00.90.80.70.6

Figure 4-1 Baseline calibration MAP, RPM, spark timing, and relative air/fueltime after crank with the CMCV open.

20*C Engine Start-up Behavior, Baseline Calibration

C- 2CC

- HC

HC

0 2 4 6 8 10 12 14 16 18

ratio as a function of

120100

80

60

40

20

0

40000

3500030000

2000015001000050000

4)E0.

C)I

20

Time after crank (sec)

Figure 4-2 Exhaust gas emissions measured at the exhaust port exit of cylinder #4 as a function oftime after crank with the CMCV open.

48

a-(U

.0

fl 0.52 0.4

0.30.2

4035

i- 25201511050

-10

2

120

100

80

60

40

20

0

14

12108

4

20

00

U-

060

0

N

0

20*C Engine Start-up Behavior, Baseline CalibrationII1 2000

1800MAP 1800-- 6D--- RPM i 14w

1200

8 00

-; 800400200

2 4 6 8 10 12 14 16 18 20

_____1.3

1.1

0.90.8070.8

T i - 05S 2 4 6 8 10 12 14 16 18 20

Time after crank (sec)

During a typical 20C start-up process, the baseline calibration resulted in an idle

speed of 950 RPM and 60 BTDC spark timing at approximately 3 seconds after the onset of

engine cranking. The average Net-IMEP of 2.45 bar with 5% COV was obtained during the

quasi-steady idle period of 3-20 seconds.

Engine operation in the first 20 seconds showed two clearly distinct periods, which

were separated by engine operating conditions around 3 seconds following engine cranking.

Thus, in subsequent discussions, the earlier period and the later period are denoted as the

transient period and the quasi-steady period, respectively.

4.2 Comparison Results of Various CMCVs

4.2.1 CMCV Overview

In-cylinder charge motion is a transient phenomenon and is developed during the in-

take process. Gas motion within the cylinder is one of the major factors that controls the

air-fuel mixing and combustion process. It is difficult to achieve robust combustion in the

start-up process due to the unfavorable mixture preparation environment such as cold port

wall temperatures, initially atmospheric port pressure, and low air velocity; each of these

conditions is detrimental to fuel vaporization. Therefore, during the start-up process, an in-

creased level of in-cylinder charge motion is helpful for improving fuel vaporization, air-

fuel mixing, and tolerance to high levels of residual dilution. Three different shapes of

charge motion control valves were investigated to better understand the effects of the en-

hanced charge motion on mixture preparation, combustion, and emissions. Figure 4-3

shows the three different charge motion control valves named as base CMCV, tumble type

CMCV, and swirl type CMCV. All three different CMCVs were made of 1 mm aluminum

plate and reduced the port cross sectional area by 75% (i.e. 25% notch area).

49

(a) Base CMCV (b) Tumble type CMCV (c) Swirl type CMCV

Figure 4-3 Three different types of charge motion control valves: (a) Base CMCV <CMCV(B)>, (b)Tumble type CMCV <CMCV(T)>, (c) Swirl type CMCV <CMCV(S)>. Flow area was 25%for all three CMCVs.

4.2.2 Mixture Preparation

The task of the engine induction and fuel systems is to prepare an air-fuel mixture that

satisfies the requirements of the engine over its operating regime. Mixture preparation is

usually discussed in terms of the relative air/fuel ratio. In order to explore this topic here,

mixture preparation is defined by the in-cylinder mixture condition just before combustion

occurs. The in-cylinder mixture condition just before the start of the combustion was quan-

tified by the calculated in-cylinder relative air/fuel ratio, resulting from fuel vaporization in

the port and cylinder produced by the charge motion. In-cylinder relative air/fuel ratio was

derived from the carbon based lambda evaluation, which was developed from time-resolved

emission measurements, as described in Section 2.3.2.

With the CMCV closed, consistently richer in-cylinder lambda levels were observed

from numerous experiments with different amounts of fuel injections. To avoid unnecessar-

ily rich in-cylinder mixture conditions with the CMCV closed, a 4% reduction in fuel injec-

tion, compared to the baseline CMCV open calibration, was applied to all the CMCV open

and CMCV closed cases for this comparison experiment. The spark timing trajectories,

however, remained the same as the baseline CMCV open calibration for all cases. In addi-

tion, with the CMCV closed, the idle air control (IAC) stepper positions were varied in or-

der to maintain the same air flow rate as the CMCV open case. Consequently, the same in-

jected fuel and same inlet air mass were delivered to the cylinder, cycle by cycle, for all

four different experiments.

50

45 I-

40 l-~

Q)

~ 35 ~ AO u 0, 8 5 30 ~ (i) X :::J

lJ.. 25 ~ "'0 Q)

ro 20 0

Cl.. ro >

l.J.J 15

10

o

Evaporated Fuel of Transient Period I

X ~

()~ 0

o~

0

A o~

5

I I -'-0 CMCVopen A CMCV closed [] CMCV closed(t) X CMCV closed(s)

-

-

* o ~ ,- _

li~!~~~~~.ai~~~~§~t~i

10 15

Cycle No.

20 25 30

Figure 4-5 Evaporated fuel as a function of cycle number during the transient period for various CMCVs.

Percent Evaporated Fuel to Injected Fuel 110

.& .&

.& .&

~ 105 .& .&

.& Q) 0

:::J lJ..

100 0 0 "'0 Q)

ti Q) .C'

95 B Q) :::J

lJ.. 90 "'0 Q) 0 CMCVopen ro 0

~ co 0 ~

.& CMCV closed(B) X Cl.. 85 X

0 CMCV closed(T) ro CMCV closed(S) > X

l.J.J - •• Linear Fit of CMCV open -- Linear Fit of CMCV closed(B)

80 30 45 60 75 90 105 120 135 150 165

Cycle No.

Figure 4-6 Percentage of evaporated fuel to injected fuel as a function of cycle number during the quasi-steady period for various CMCVs.

52

4.2.3 Combustion Characteristics

Combustion stability was investigated qualitatively during the transient period, and

was quantified by the COV of the Net-IMEP during a quasi-steady period. Net-IMEP was

calculated from in-cylinder pressure data acquired from cylinder #4. More stable Net-IMEP

was observed in all three CMCV closed cases compared with the CMCV open case during

the transient period as shown in Figure 4-7. During the quasi-steady period, the COV of

Net-IMEP was also decreased with the base CMCV and the swirl type CMCV; a COV of

5.2% and of 5.5% were obtained, respectively, while the CMCV open case yielded a COV

of 5.9%. The combustion stability, however, slightly worsened with the tumble type

CMCV up to a COV of 6.2%.

10 CMCV open 1 CMCV closed(B)

80M 6- co -

2 - z

0- 0 -. U

li 0~ liuiL~lihI ..llluuu1Illff0 5 10 15 20 25 30 0 5 10 15 20 25 30

10-CMCV closed(T) 10 CMCV closed(S)

6 --

.a.

LJ 4-

z

|11:41 1|111 2 111112 01 111 10 5 10 15 20 25 30 0 5 10 15 20 25 30

Cycle No. Cycle No.

Figure 4-7 Net-IMEP as a function of cycle number during the transient period for various CMCVs.

53

Burn rate analysis was conducted on the 0 to 20 seconds after cranking. The flame-

development angle (0-10% energy-release fraction) and rapid-burning angle (10-90% en-

ergy-release fraction) were investigated with the CMCV open and with the various CMCVs

closed. It is known that the flame development process is primarily influenced by the mix-

ture state, composition, and motion in the vicinity of the spark plug. The rapid burning

stage, during which the major portion of the charge bums as the flame propagates across

the chamber, is obviously influenced by conditions throughout the combustion chamber.

The 0-10% and 10-90% combustion durations were shortened with all three different

CMCVs as shown in Figure 4-8 and Figure 4-9. In particular, the 0-10% and 10-90% com-

bustion durations were significantly reduced with the CMCV closed during the transient

period. There are two plausible mechanisms for these faster burning processes. First, differ-

ences in in-cylinder mixture composition, resulting from the enhanced mixture preparation,

could affect the combustion duration. On the other hand, differences in gas motion in the

vicinity of the spark plug, resulting in differences in motion of the flame kernel during the

early stages of development, could also affect the combustion duration.

Least square fit lines for the CMCV open and the base CMCV closed case were added

to Figure 4-9 for a more clear illustration. Significant reductions were also achieved with

the CMCV closed during the quasi-steady period, due to the combined effect of richer in-

cylinder mixture and enhanced gas motion within the cylinder, while the injected fuel mass

and spark timing trajectories remained the same for all cases.

54

Combustion Duration of Transient Region

o CMCV openA CMCV closed(B)

- 0 CMCV closed(T)X CMCV closed(S)

00

0

0 P 0

A

0-10%-

o 5 10 15 20 25 30

Cycle No.

50

45

40

35

300

25

E0 20C-,

o 15

10

0

- -

-x

0 0 A-

A A~~0

0 d,A>A&

-210-90%

0 5 10 15 20 25 30

Cycle No.

Figure 4-8 0-10% and 10-90% combustion durations as a function of cycle number during the tran-sient period for various CMCVs.

Combustion Duration of Quasi Steady Period

37

35

33

< 310C 2901 27

O 25

2 23

.o 21Eo 19

' 170

615

13

42

40

a 38

< 36U

c 34

O 32

30S0C0= 28

Ct)

-O 26E0o 24

_ 22

620

18

30 45 60 75 90 105120135150165

Cycle No.

Figure 4-9 0-10% and 10-90% combustion durations assteady period for various CMCVs.

0

0 0(0 0

0 0 0-0 00 0 00 0

00 0 EVO0 *oQ

cxy 00 0tm qj M-Ac 0-' 15 ' ' ' ' 'W M

- O -- < CuAO AmOnci ;xa MMxCU= cx M GOP A

-0) ctnoc Aor) >*DM%0 M4"

xc-x So *E3<XEs A

.m x A e oxx*-" hf A A-

06~"AA*>A

AAo0iCA A0 0 10-90%/C

30 45 60 75 90 105120135150165

Cycle No.

a function of cycle number during the quasi-

55

50

45

40

35

30

25

20

15

=3

0OOR-o

10

0 CMCV open_ A CMCV closed(B)

0 CMCV closed(T)- X CMCV closed(S)

--- Linear Fit of CMCV open- Linear Fit of CMCV closed(B)

00 0 0

-X 0 o o -(A& 0

-COOO cx<)00< 006 octl A 0-.) 0 XO OX MAD ~ X* (a xUa-x M> arflM&Wma

ARA 00 AO cAoA- 4OAA AA00 0 0

0-10%, 1 1 , , , , , , , , , , , 7, ,

4.2.4 Exhaust Emissions

Time-resolved cylinder #4 exhaust port exit HC, CO and CO 2 concentration data were

taken for the first 20 seconds of engine operation. During the transient period, high HC and

CO emissions were observed due to the rich in-cylinder mixture condition and the in-

creased charge dilution. As the engine accelerated, the intake manifold pressure decreased

to a low value of about 0.26 bar, which increased the residual gas fraction due to the in-

creased backflow during the valve overlap period. The highest HC mole fraction occurred

around cycle #13 where the intake manifold pressure dipped to its minimum value. As the

intake manifold pressure started to increase to the idle calibration condition, there was a

progressive drop in HC mole fraction.

Using the same fuel injection and spark timing trajectories, higher levels of engine-

out HC emissions relative to the CMCV open case were observed with all three different

CMCVs. There were several plausible reasons for the trends shown in Figure 4-10 and Fig-

ure 4-11. Richer in-cylinder mixture conditions with the CMCV closed cases mainly pro-

duced higher engine-out HC emissions. As combustion was phased earlier due to faster

burning with the enhanced charge motion, higher in-cylinder peak combustion pressures

occurred and also increased the mass fraction of the HC trapped in crevice volumes that

escaped the oxidation during flame propagation.

56

Engine-out HC Emissions of Transient Period

7 X 0 CMCV openA CMCV closed(B)0 CMCV closed(T)

> 6 -X CMCV closed(S)

.3 A

E 5-C,

o X

Ew 3 -

LO AI ~ 0 X

A2 --

0 X

_ 0 10 0 0

0 1

0 5 10 15 20 25 30

Cycle No.

Figure 4-10 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a func-tion of cycle number during the transient period for various CMCVs.

0.7

S0.6

E20.5

C,).V)

ELU 0.4

S0.3

LLJ

0.2

Engine-out HC Emissions of Quasi-steady Period

0 CMCV openx A CMCV closed(B)

x 11 CMCV closed(T)- O X CMCV closed(S)

A*

. x x ~~ xxX oc 6

i bXS,%~~ A Ax) X

0 A a A A-

30 45 60 75 90 105 120 135 150 165

Cycle No.

Figure 4-11 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a func-tion of cycle number during the quasi-steady period for various CMCVs.

57

4.2.5 Flow Testing Results

In order to quantify the in-cylinder charge motion with the CMCVs, flow bench tests

were conducted at GM Powertrain Warren Engineering Center. The raw in-cylinder flow

motion information, either the swirl or tumble, was indicated as a torque. However, the

torques, by themselves, are of little value and must be corrected for flow rate and test con-

ditions. By normalizing torque output values with engine geometries (bore and stroke) and

plotting them versus valve lift, the swirl and tumble indexes were produced. Physically, the

index represents a measure of the rate at which the charge in the cylinder is swirling or

tumbling around its axis per engine revolution if all the inducted charge entered the cylin-

der at the given valve lift. More information about the swirl and tumble index calculation is

described in Appendix A.3.

Figure 4-12 and 4-13 show the swirl and tumble index from the raw torque values that

come from flow bench meters as a function of normalized valve lift. The strongest swirl

flow was observed with the swirl type CMCV and the moderate increase of swirl flow was

also shown with the base CMCV, while little impact on the swirl motion was identified

with the tumble type CMCV. All three designs, however, increased the tumble index rela-

tive to the CMCV open case, and the base CMCV was found to give slightly higher tumble

flow than the others.

The differences of these swirl and tumble motions are likely to affect the liquid fuel

transport, distribution, and evaporation; therefore, those flow testing results can be corre-

lated with the previous results of mixture preparation, combustion characteristics, and HC

emissions.

58

6

5

4

3

2

0

-1

Swirl Index

0.00 0.05 0.10 0.15 0.20 0.25 0.30

L /DV V

Figure 4-12 Swirl index from flow bench meters as a function of normalized valve lift for variousCMCVs.

2.0

1.5

1.0

0.5

0.0

-0.5

-1.0

-1.5

-2.00.00

Tumble Index

AA

-0-o

'0 ----

0 --- 0---0------- ------

--0--- CMCV open-- CMCV closed(B)

E CMCV closed(T)X-CMCV closed(S)

0.05 0.10 0.15 0.20 0.25 0.30

L /DV V

Figure 4-13 Tumble index from flow bench meters as a function of normalized valve lift for variousCMCVs.

59

0)

- --- CMCV open-- CMCV closed(B)

E3 CMCV closed(T)x- CMCV closed(S)

- x

0-)-----0------- --- ~ O .. ~

E

I I I I I I

4.2.6 Overall Assessment of Different CMCVs

Different geometric charge motion control valves were found to yield differences in

fuel evaporation, combustion rate, and engine-out HC emissions. With the same fuel injec-

tion and spark timing profile, the highest fuel evaporation was achieved with the base

CMCV, which has the highest tumbling and also significant swirling flow motion. Lower

fuel evaporation and higher HC emissions were observed with the swirl type CMCV, which

has the strongest swirl motion and moderate tumble motion, than the other CMCVs. The

swirl type CMCV showed slower 0-10% combustion but faster 10-90% combustion than

the tumble type CMCV that enhances only the tumble motion.

A plausible mechanism can be developed by correlating the in-cylinder flow motion

characteristics with the other results. With the swirl type CMCV, the bulk flow was forced

to rotate strongly through the cylinder liner side; therefore, more of fuel droplets were pre-

sumably distributed near the cylinder wall than the centrally located spark plug gap. This

was likely to cause the lack of fuel in the vicinity of spark plug gap and the excess of fuel

near the cylinder wall. This fuel distribution can explain relatively slower flame initiation

(0-10% combustion duration) in the earlier stage of combustion and higher flame propaga-

tion (10-90% combustion duration) in the later stage of combustion with the swirl type

CMCV. In addition, by this excess swirl motion through the cylinder wall, more liquid fuel

could wet the cylinder wall, resulting in increased amounts of fuel deposited in cylinder and

also engine-out HC emissions. In contrast, with the tumble motion, strong flow patterns can

be expected near the piston top and cylinder head surface. Thus, higher flow motion can be

achieved in the vicinity of spark plug gap.

The overall results indicate that in-cylinder swirl and tumble motion should be appro-

priately combined by optimizing the shape of the CMCV. The base CMCV, which moder-

ately enhanced the swirl and tumble motion, showed the greatest potential for mixture

preparation, combustion characteristics, and HC emissions.

60

4.3 Effects of Reduced Fuel Injection

4.3.1 Modification of Fuel Injection

Since relatively richer in-cylinder mixture conditions were observed with the CMCV

closed, reduced sequential fuel injection was then used with the CMCV closed case. In or-

der to achieve the same in-cylinder lambda trace with the baseline CMCV open calibration,

reduced amounts of fuel were injected for the CMCV closed case. Spark timings, however,

were maintained at the same levels of the baseline CMCV open calibration. The impacts of

spark timing on combustion and emissions were thus decoupled from the effects of reduced

fuel injection.

The comparison of the various charge motion control valves indicates that the base

CMCV has the greatest potential for improving mixture preparation and combustion char-

acteristics even though it produces higher HC emissions (due mainly to the higher in-

cylinder lambda conditions). Thus, the base CMCV was used for further comparison ex-

periments.


The effect of reduced fuel injection with the CMCV closed on mixture preparation

was evaluated while using the same spark timing trace as the base CMCV open calibration.

Figure 4-14 shows the injected fuel mass as a function of cycle number during the first 20

seconds following engine cranking, in order to obtain the same in-cylinder lambda trace as

the baseline CMCV open calibration. The amount of cycle-by-cycle injected fuel mass was

reduced approximately by 7.5% during the transient period, including a 12% reduction for

the first two cycles. During the quasi-steady period, overall 3.5% less fuel was injected

with the CMCV closed. Meanwhile, the same baseline spark timings were applied to both

the CMCV open and closed case.

61

Injected Fuel Mass125 k I I ' * * I

120 - 0 CMCV open115 A CMCV closed

110 &-> 105

045E A

40

CO 35

30

L 25

20 -

0 A

w15- 10

5

00 20 40 60 80 100 120 140 160

Cycle No.

Figure 4-14 Required mass of injected fuel as a function of cycle number for the same in-cylinderlambda trace with the CMCV open and closed during the first 20 seconds after crankingcommenced.


Combustion characteristics were investigated in terms of the variation of Net-IMEP

and combustion duration in order to assess the effectiveness of the CMCV with reduced

fuel injection. A smaller variation of Net-IMEP, shown in Figure 4-15, was obtained with

the CMCV closed (relative to the CMCV open case) during the transient period, while the

same in-cylinder lambda and spark timing trace were maintained. During the quasi-steady

period, Net-IMEP averaged 2.45 bar with 5% COV for both the cases. Test-to-test

variations in average Net-IMEP and COV were ±0.1 bar and ±0. 3%, respectively.

62

10

.0

6

4

8

6

4

2

0

CMCV open

0 5 10 15 20 25 30

Cycle No.

Figure 4-15 Net-IMEP as a function of cycle number during the transient period for the CMCV opencase and the CMCV closed case with reduced fuel injection.

Burn rate analysis was performed on the first 20 seconds following engine cranking to

determine the 0-10% and 10-90% combustion duration. Figure 4-16 and Figure 4-17 show

clear differences between the CMCV open and closed case. Noticeable reductions in both

the 0-10% and 10-90% combustion duration were found with the CMCV closed, especially

during the transient period. Reduced combustion durations of the CMCV closed case could

allow additional spark retard while injecting less fuel overall.

63

i CMCV closed

-UmmCL

025 300 5 10 15 20

Cycle No.

10

Combustion Duration of Transient Region60 , 1 6 . . . . . .. .

0-10% 10-90%55 55 0

0) 050 - 0-0 0 0 0 0 0

045 - 45 o0

-0 40 0 .2 40 -C 00400 0 0o 0 .0 35 0 35 - 0C - 0 0 0 -0 00. 0 0 . 0 ODo

30 - 0 O 30 -A 0 AAA 0 0 0 0

~ 5 A A 0 A A E 25 A A A"AE 0 2 - A 20 - 0 A

0 0 oMA Ab~20 -A A - 20

15 - 0 CMCV open 0 15

IA CMCV closed A

10 10 I I I0 5 10 15 20 25 30 0 5 10 15 20 25 30

Cycle No. Cycle No.

Figure 4-16 0-10% and 10-90% combustion durations as a function of cycle number during the tran-sient period for the CMCV open case and the CMCV closed case with reduced fuel injec-tion.

Combustion Duration of Quasi Steady Region

0 CMCV open 035 - A CMCV closed - 40 0 -

- -Linear Fit of CMCV open A- Linear Fit of CMCV closed W 0 0 0

"0 0 AcLo A AAoo

030 00 - U 35 A 0 0 A 00 o AC0040 r_ 0 0 A

.o C o amo o A At 02 0 o OOA ,36 O(AS o0oAL 4 ao GD Ao o*& A 0 OAOO vo XOOft* dMd

O cD cxDO Od A $2110 0 cA oc 0 o AoeoM) 25 -a Ci A 30 o 0 Ala tCooC oaaiA ~ A A o0 A * A.O 0s sa Me A Ao& oA

. 0 A AA A A - L eOAAAL AAAE -A& AM&!,a! <k AAA ' J E 2 AAA AAn".Gr A A oA AITL AAAUA&

o AA

6-1iS 10-90%I I~ 1 i e . I Ii .1 , i 1 i I I 1. , , ,

25 40 55 70 85 100115130145160 25 40 55 70 85 100115130145160

CyCle No. Cycle No.

Figure 4-17 0-10% and 10-90% combustion durations as a function of cycle number during the quasi-steady period for the CMCV open case and the CMCV closed case with reduced fuel injec-tion.

64

Gas motion effects on air-fuel mixture in the port and cylinder can be divided into two

areas: the effect on mixture preparation which is represented by in-cylinder air/fuel ratio,

and the effect on flame-development which is represented by the 0-10% energy-release

fraction angle. Since the same in-cylinder lambda trajectory was maintained with the same

spark timing for both the cases, the difference of the 0-10% combustion duration can be

considered as the direct charge motion impact on flame-development. The CMCV thus de-

creased the 0-10% combustion duration, 80 CA on average during the transient period and

2' CA during the quasi-steady period, as a result of enhanced charge motion impacts on

combustion initiation.


With the CMCV open, the high engine-out HC emissions region, shown in Figure 4-

18, was observed near where the lowest intake manifold pressure occurred. The low in-

cylinder pressure presumably assisted the vaporization of liquid fuel deposited throughout

the engine. However, with the CMCV closed, high engine-out HC emissions were found in

the earlier stages of engine start-up. This implies that the CMCV enhanced fuel evaporation

rate right from the beginning; therefore, less liquid fuel should be deposited in the cylinder

during the transient period. Figure 4-19 shows that the engine-out HC emissions during the

quasi-steady period were reduced with the CMCV closed by 6.5%, while the same in-

cylinder lambda and spark timing profile was maintained.

65

7

6

5

E

C 4

ELU

3

U

2

LU

0

Engine-out HC Emissions of Transient Period

A 0 CMCV openA CMCV closed

AA

A A o00

0 0 A A 00 0

4A AAA 00 AA AAA&&f464*AA

0L I 15 20 25 I0 5 10 15 20 25 30

Cycle No.

Figure 4-18 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a func-tion of cycle number during the transient period for the CMCV open case and the CMCVclosed case with reduced fuel injection.

0.65

0.60

S0.55

C)E 0.50

o 0.45

E 0.40LUJ

I 0.35

06 0.30

0.25LUI

0.20 12


0 CMCV open0 A CMCV closed

A- - Linear Fit of CMCV open-- Linear Fit of CMCV closed

m - O

A 0 0 0A OA 0

- A 0 A

-~ A -

-

40 55 70 85 100

Cycle No.

15 I ,I 160115 130 145 160

Figure 4-19 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a func-tion of cycle number during the quasi-steady period for the CMCV open case and theCMCV closed case with reduced fuel injection.

66

. I . I . I . I . I .

3

4.4 Effects of Retarded Spark Timing

4.4.1 Modification of Spark Timing

Since considerable shortening in combustion duration and improvement in combus

tion stability were achieved with the use of the charge motion control valve, additional

spark retard was feasible with the CMCV closed. Three different spark modifications with

the CMCV closed were investigated during the first 20 seconds following engine cranking.

Spark timings, shown in Figure 4-20, were retarded (~esp = _5° and -10°) with respect to

the baseline spark timing (~esp = 0°), while the same in-cylinder lambda trace was main

tained for all three different spark modifications.

40

35

30

25

G 20 0

t-eo

15 'L-0) c 10 ·E f= 5 .::,t:.

m 0 Cl. if)

-5

-10

-15 0

Spark Modification with the CMCV closed

20 40 60

--~El =0° sp

- - - ~e = _5° sp

- - . - . ~e = -1 0° sp

80 100 120 140 160

Cycle No.

Figure 4-20 Spark timing as a function of cycle number during the first 20 seconds following engine cranking for various spark modifications with the CMCV closed.

67


Figure 4-21 shows the required Inass of injected fuel for the same in-cylinder lalnbda

trace as the baseline spark tuning (~esp = 0°) case. The steady-state experilnental results

showed that the air and fuel mass flow rates through the engine were increased with an in

creased level of spark retardation to maintain the same mean Net-IMEP. Similar trends

were observed in the start-up experiments. As spark timing was retarded, more fuel was

injected and also more air was inducted for the same in-cylinder latnbda trajectory.

Injected Fuel Mass

20~,~~,~~,~~,~~,~~,~;~,=~==,====,==~~,

Q3 14 ~ u.. "0 Q)

ti .~ 12 r-

A ~o = 00

sp * ~o = _50 sp

~o = _100

sp

10 I I I I I I I I I I

-

30 45 60 75 90 105 120 135 150 165

Cycle No.

Figure 4-21 Required mass of injected fuel as a function of cycle number for the same in-cylinder lambda trace with various spark modifications during the first 20 seconds after cranking commenced.


As combustion was phased later in the cycle, spark timing retardation was limited by

combustion stability. Therefore, at each of the three different spark modifications, combus

tion stability was assessed during the first 20 seconds following engine cranking. Figure 4-

22 shows the Net-IMEP during the transient period. As the level of spark retardation was

increased, there was a noticeable increase of fluctuation in the Net-IMEP during the

transient period. During the quasi-steady period, the COY of Net-IMEP was also increased;

68

sient period. During the quasi-steady period, the COY of Net-IMEP was also increased; the

COY of5 .2%, 6.0%, and 6.9% were observed for the ~esp = 0°, ~esp = _5 °, and ~esp = -10°

spark modification, respectively.

10

8

6 ro e

e... w ~ 4 Z

2

o

10 10 l18 = 0° _ l18 = _5° I ""t*4Jl18 = -10°

sp sp sp

ro e e... w ~

o 5 10 15 20 25 30

Cycle No,

8

6

4

ro e e... w ~

o 5 10 15 20 25 30

Cycle No.

8

6

4

2

o o 5 10 15 20 25 30

Cycle No.

Figure 4-22 Net-IMEP as a function of cycle number during the transient period for various spark modifications with the CMCV closed.

The location of 50% mass fraction burned, 0-100/0 combustion duration and 10-90%

combustion duration, shown in Figure 4-23 , Figure 4-24 and Figure 4-25 , were determined

by bum rate analysis. As the spark timings were retarded, the locations of 500/0 mass frac

tion burned were also retarded during both the transient and quasi-steady period. With the

most aggressive spark retardation (~esp = -10°), relatively larger variations in the location

of 50% mass fraction burned and the 10-90% combustion duration were observed as com

bustion occurred in a rapidly expanding cylinder volume. However, relatively less varia

tions were observed in the 0-100/0 combustion duration, which is influenced primarily by

the in-cylinder mixture preparation and in-cylinder gas motion. This is likely because the

in-cylinder lambda trace was fixed and gas motions in the vicinity of the spark plug were

fairly silnilar due to the use of the CMCV for all three spark modifications.

69

Location of 50% Mass Fraction Burned

60 o

o 20 40 60 80

Cycle No.

100

... ~6 = 0° sp * ~6 = _5° sp

o ~6 = _10° sp

120 140 160

Figure 4-23 Location of 50% mass fraction burned as a function of cycle number during the 20 seconds after engine cranking for various spark modifications with the CMCV closed.

Combustion Duration of Transient Region

45 t--

~ 40 t--'0

« ~ 35 t--c o

o o

o

& ~e = 0° sp * ~e = _5° sp

o ~e = _10° sp

o

.~ 30 t-- & 0 0 :; .. A 0 0 f:!J. &

§ 25 Co ;*~;frJob .~ ~;~& ~ ~ 20 r* T & & o ,. *& &

U M g 15 r ~ * ,......

6 10 0-10%

I I I 1 11 ~

o 5 10 15 20 25 30

Cycle No.

45

cr, ~ 40

« U '-"'35 c o .~ 30 ::J o C .g 25 (f)

::J .0 E 20 o U

g 15 en 6 ,...... 10

-

* 0 -

0

& *'IE -

0 & ~& 0

* *

10-90% ~ _I I 1 I 1 1

o 5 10 1 5 20 25 30

Cycle No.

Figure 4-24 0-10% and 10-90% combustion durations as a function of cycle number during the transient period for various spark modifications with the CMCV closed.

70

0, Q) "0

<t:

35

S:?. 30 c o .~

~ o 25 c o

'f7; ::J

.D E 20 o U

"* o r-

6 15

Combustion Duration of Quasi-steady Region

~ 69 = 00

sp

* 69 = _50 sp

0 69 = _lOa sp

W1!!Il(

~ ~ o~

OIl ~o~d\ ~o ~o ~ *-~

0.0 ____

~DO _~~ __ BAlIlib'.

~~ ~&:J_~~~ ... ~ ... l-..:ro CI&l JlIiiIit. .... .. ~,..

tl~ ~* ~~*~;tJ* 0 -~~ ;eJ *

c- '* 0-10% _L I _I L I

45~~~~~~~~~~

"* 020 (j)

6

*

o

o

* * o * **

00 0 * ~0[JlK:

~ ~. o ITJ",*_

o 0 *_ * GIE*J i3 .-:J ~o ** '* DID*: ~

10-90% 15~~~~~~~~~~

30 45 60 75 90105120135150165 30 45 60 75 90105120135150165

Cycle No. Cycle No.

Figure 4-25 0-10% and 10-90% combustion durations as a function of cycle number during the quasisteady period for various spark modifications with the CMCV closed.


Time-resolved cylinder #4 exhaust port exit HC, CO and CO2 concentration meas

urements were taken for the first 20 seconds of engine operation. In order to investigate the

effect of retarded spark timing, the engine-out HC mass emissions were evaluated for three

different spark modifications. Figure 4-26 and Figure 4-27 show that the engine-out HC

emissions were reduced with the moderate spark retardation (~esp = _5°) by 220/0 during the

transient period and by 140/0 during the quasi-steady period. However, with the more ag

gressive spark retardation (~esp = -10°), the engine-out HC emissions were increased by

200/0 during the transient period and by 14% during the quasi-steady period. As described in

Section 4.4.2, the air and fuel mass flow rates through the engine were increased with in

creased levels of spark retardation to maintain the same mean Net-IMEP. Thus, a decrease

in HC concentrations with the spark retardation was more than offset by an increase in

mass flow rate.

71

..--..6 CI)

U >. u 0,5 S (f)

c o 'Vi (f)

~ 3 U I

"5 2 o a) c '0, 1 c

l.LJ

o

III I

'- \ r

;-

r

r-

o

Engine-out HC Emissions of Transient Period I I I I I I

---A--- ~e = 00

sp

- -)f- - ~e sp

= _50

···D···~e = _100

sp

~\ -

-. .' \ .'" $ \ b\ .~

)iii '. 0 '. -~: b '0: '.

~ & .... *,0. "'=~?-~~~} '0 p ... -* ··*-*t~i,;j"iH= .. ~~i

10 15 20 25 30

Cycle No.

Figure 4-26 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a function of cycle number during the transient period for various spark modifications with the CMCV closed.

Q) u

0.8

~ 0.7 0, S (f)

§ 0.6 'Vi (f)

'E W 0.5 U I

"5 o 0.4 a) c '0, c W 0.3

30


o

45 60 75

t>9 = 00

sp * t>9 = _50 sp

o t>9 = _100

sp

-- Linear Fit of t>9 = 00

sp

- .• - Linear Fit of t>9 = _50 sp

- - • Linear Fit of t>9 = _100

sp

90 105 120 135 150 165

Cycle No.

Figure 4-27 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a function of cycle number during the quasi-steady period for various spark modifications with the CMCV closed.

72

4.5 Modified Start-up Strategy with CMCV closed

4.5.1 Modification of Fuel Injection and Spark Timing

As discussed in the previous sections, mixture preparation and combustion character-

istics were improved with the CMCV closed by enhancing fuel vaporization and increasing

burning rate. Greater fuel vaporization allowed reduced fuel injection and the faster burning

rate enabled additional spark retard. In addition, the extended retardation of spark timing

decreased HC emissions. Therefore, an appropriate combination of reduced fuel injection

and retarded spark timing with the CMCV closed could be a way to achieve more robust

combustion and lower HC emissions during the engine start-up process.

In order to achieve the same in-cylinder lambda profile as the baseline CMCV open

calibration, reduced sequential fuel injection strategy was applied to the CMCV closed case

as shown in Figure 4-28. The amount of reductions in fuel injection was determined by the

results of Section 4.3. At the same time, retarded spark timing strategy was also utilized to

the CMCV closed case for more reductions in HC emissions as shown in Figure 4-29.

Spark timing was modified using the results of Section 4.3 and Section 4.4.

During the transient period (0 to 3 seconds), a total of 7.5% reduced fuel, including a

12% reduction for the first two cycles, and 80 CA retarded spark timing were applied to the

CMCV closed case. During the quasi-steady period (3 to 20 seconds), overall 3.5% reduced

fuel and 50 CA retarded spark timing were applied to the CMCV closed case with respect to

the baseline CMCV open case.

73

0)

EC-)

CO

LL_0C-)C'

125

120

115

110

105

45

40

35

30

25

20

15

10

5

0

Injected Fuel Mass

-0o CMCV open--A CMCV closed

0

A

A-

- -0

0 20 40 60 80 100 120 140 160

Cycle No.

Figure 4-28 Injected fuel mass as a function of cycle number with the CMCV open and closed duringthe first 20 seconds after cranking. Reduced sequential fuel injection strategy was appliedto the CMCV closed case.

Spark Timing

U

C

CO

LI)

35

30

25

20

15

10

5

0

-5

-10

CMCV open--- \---- CMCV closed -

IV'

\ -

- --

I I I I . I . I .

0 20 40 60 80 100 120 140 160

Cycle No.

Figure 4-29 Spark timing as a function of cycle number with the CMCV open and closed during thefirst 20 seconds after cranking. Retarded spark timing strategy was applied to the CMCVclosed case.

74


In-cylinder lambda was calculated from the exhaust HC, CO, and CO2 emissions data

to investigate mixture preparation before combustion. Figure 4-30 shows in-cylinder

lambda traces as a function of cycle number for the cases described in the previous section.

Cycle-by-cycle in-cylinder lambda points were well matched for both cases within ±2%.

Figure 4-30 suggests that both cases had similar mixture preparation conditions before

combustion occurred.

In-cylinder A.

1.8 I I I I

I 0 CMCVopen ~ 1.6 f-.A. • CMCV closed

0 1.4 I- -

1.2 I- -

0.4 L.-LI----'----'I'-'--"--I...&.-.-........ I---'-----'I----'~l'__"__...L.._ 1-'--....L1----,---,-,i

o 20 40 60 80 100 120 140 160

Cycle No.

Figure 4-30 In-cylinder J... as a function of cycle number during the first 20 seconds following engine cranking with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.


Net-IMEP and combustion durations were calculated from in-cylinder pressure data

acquired from cylinder #4 in order to investigate combined effects of reduced fuel and re

tarded spark timing with enhanced charge motion on combustion characteristics. Qualita

tive observation of Net-IMEP shown in Figure 4-31 indicates that combustion stability was

improved with the CMCV closed during the transient period. However, the COY of Net-

75

IMEP during the quasi-steady period was increased by about 1 %; the COY of 5% and 6%

were observed for the CMCV open and closed case, respectively.

ro e 0... ll.J

8

6

~ 4 z

2

o o

_CMCVopen

10 15 20 25 30

Cycle No.

'-ro e

0... ll.J

8

6

~ 4 Z

2

o

_ CMCV closed

o 5 10 15 20 25 30

Cycle No.

Figure 4-31 Net-IMEP as a function of cycle number during the transient period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case

Bum rate analysis was conducted on the first 20 seconds following engine cranking.

Figure 4-32 and 4-33 show a comparison of the 0-100/0 and 10-90% combustion duration

and the location of 50% mass fraction burned during the transient period. Significantly re

duced combustion duration was achieved with CMCV closed even using 7.5% less fuel and

8° CA spark retardation. The location of 500/0 mass fraction burned shown in Figure 4-33

was also advanced especially between cycle number 5-15 where intake manifold pressure

has severely low values. Since the location of 500/0 mass fraction burned gives a good indi

cation of when combustion is occurring, the results shown in Figure 4-33 suggests possible

additional spark retardation with the CMCV closed during the transient period.

76

Combustion Duration of Transient Period 60 I I I I I I I 60 I I I I I I

0-10% 10-90% 55 - 55 r. 0

0> 0> 0 Q) 50 -

Q) 50 0 "0 0 "0 0

« 0 « ~ 45 - ~ 45 00

c 0 c

0 0 • • .~ 40 0 - .~ 40 -

:s 0 0 :s o .~O

0 0 35 0 ~

- 35 c- o • • -0 0 c 0 00. c 0

• CD~~ 0 0 o 0 ~ .~ .~ 30 30 -

::::l M,cAc1. ::::l o ~ /.~~ .0 .0

E 25 • E 25 0 -0 o ~ i1 0 0 CD U 0 • Ji. U • cf2. 20 • .6M cf2. 20 ~. ~. • 0 ~.

0 ..-- en ~

6 15 I

0 CMCVopen t 6 15 f- •• I 0 CMCVopen t • CMCV closed • CMCV closed

10 I I L 10 0 5 10 15 20 25 30 0 5 10 15 20 25 30

Cycle No. Cycle No.

Figure 4-32 0-10% and 10-90% combustion durations as a function of cycle number during the transient period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.

65 f-

60 f-

55 f-

U- 50 f-0 f-- 45 ~ 40

r-f-en

l.J.... 35 ~

• f-

cf2. 30 f- 0 0 l!') 25 f-

a c 20 f-0

15 .~ f-

u 10 0 --l

5

0 -'-o

Location of 50% MFB of Transient Period

0 0

()

0 0 o. 0 •• • • 0 • 0 •

o~· ••

1 I

5 10

0

0

• 0

.~. • 0

• •

I I

15

Cycle No.

0°

0

•

I I

• 0 ••

0

• 0 o. • • 0

• 0

0

CMCVopen CMCV closed

20 25

I-

---

---

-

---

I: 30

Figure 4-33 Location of 500/0 mass fraction burned as a function of cycle number during the transient period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.

77

Figure 4-34 and 4-35 show a comparison of the 0-10% and 10-900/0 combustion dura

tion and the location of 50% mass fraction burned during the quasi-steady period. Rela

tively large variations in the 10-90% combustion duration and the location of 50% mass

fraction burned were observed with CMCV closed. This increased variation was responsi

ble for slightly decreased combustion stability (COV) during the quasi-steady period.

Variations in the location of 50% mass fraction burned are known as the primary cause of

cyclic variations in the Net-IMEP. Since the expansion ratio decreases rapidly during com

bustion for retarded spark timing, the combustion phasing is one of the major factors which

determines individual cycle Net-IMEP. Both the CMCV open and closed case were found

to experience almost the same combustion phasing on average since two linear fit lines of

the location of 50% mass fraction burned almost coincided, while using reduced sequential

fuel injection and retarded spark timing strategy for the CMCV closed case.

Combustion Duration of Quasi Steady Period 40 45~~~~~~~~~~

o CMCVopen .. CMCV closed

- •• Linear Fit of CMCV open - Linear Fit of CMCV closed

o

eft. o 20 O'l 6

0-10% 10-90%

o A o

10 15~~~~~~~~~~

25 40 55 70 85100115130145160 25 40 55 70 85 100115130145160

Cycle No. Cycle No.

Figure 4-34 0-10% and 10-90% combustion durations as a function of cycle number during the quasisteady period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.

78

G 0 f-

~ co l.L

~

eft. 0 L{)

a c 0 .~

u 0

-.J

55

50

45

40

35

30

25

Location of 50% MFB of Quasi-steady Period

o CMCVopen • CMCV closed

- • - Linear Fit of CMCV open --Linear Fit of CMCV closed o • .. () 0

00 A

20~~~~~~~~~~~--~~~~~~~~

25 40 55 70 85 100 115 130 145 160

Cycle No.

Figure 4-35 Location of 50% mass fraction burned as a function of cycle number during the quasisteady period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.


HC, co and CO2 concentrations were measured at the exhaust port exit of cylinder #4

during the first 20 seconds after engine cranking commenced. Figure 4-36 and Figure 4-37

show the mass of engine-out HC emissions per cycle during the transient period and during

the quasi-steady state, respectively. There was a noticeable decrease in engine-out HC

emissions around the region in which the intake manifold pressure was lowest. The engine

out HC emissions were reduced by 18% during the transient period and by 7% during the

quasi-steady period. A total reduction of 12% in HC emissions was achieved by the charge

motion control valve with the reduced fuel injection and retarded spark timing.

79

I

6 ... Cil U » 0 u 0,5 S l/1 C 0 ·Vi l/1 ·E

3 l.LJ

U I

:5 2 0 cG c ·61 1 c

l.LJ

0 0

Engine-out HC Emissions of Transient Period I I

10

I I -'

I 0 CMCVopen ~ • CMCV closed

'"-----

15

Cycle No.

20

Transient region

I

25 30

Figure 4-36 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a function of cycle number during the transient period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.


0.60

Cil ~) ~ 0.55 u 0, o S 0.50 l/1 C 0 ·Vi

0.45 l/1

·E 0.40 l.LJ

U I 0.35 :5 0 cG 0.30

Quasi steady region c ·61 c 0.25

l.LJ

o CMCVopen • CMCV closed

- • - Linear Fit of CMCV open -- Linear Fit of CMCV closed

o. 20 L..-~~--'---'----I.---''---~''''''''''--'---'----'----'L..-~''''''''''--'---'----I.--' 25 40 55 70 85 100 115 130 145 160

Cycle No.

Figure 4-37 Engine-out HC mass emissions measured at the exhaust port exit of cylinder #4 as a function of cycle number during the quasi-steady period with the CMCV open and closed. Reduced sequential fuel injection and retarded spark timing strategy was applied to the CMCV closed case.

80

Chapter 5

Summary, Conclusions, and

Recommendations

5.1 Overview

The purpose of this research was to investigate the effects of various intake charge

motion control approaches on the cold start-up process of a port fuel injected SI engine.

Experiments were performed to assess the effectiveness of enhanced charge motion on mix-

ture preparation, combustion, and HC emissions. In this chapter, the findings of the re-

search will be summarized, conclusions will be drawn, and implications for future research

will be discussed.

5.2 Steady State Experiments

Steady state experiments were conducted to provide basic insight into combustion

characteristics and HC emissions behavior with the charge motion control valve (CMCV).

The combustion process and exhaust gas state were examined under the fixed engine opera-

tion (2.5 bar Net-IMEP, 1000 rpm) and cold (20'C) fluid condition. The following conclu-

sions were obtained from the steady state experimental results.

* Increased charge motion with the CMCV closed was found to shorten the 0-10%

and 10-90% combustion duration at a given spark timing. This reduced combustion

duration caused the 50% mass fraction burned to occur up to 5' CA earlier for the

same spark timing.

81

* With increased spark retardation from optimum timing, combustion stability was

found to worsen as more of the combustion process occurred in a rapidly expanding

cylinder volume. However, significant improvements in combustion stability (0.5-

1.5% reductions in COV) were achieved with the CMCV closed as spark timing

was retarded. This is due to the increase in burn rate with the CMCV closed, result-

ing from the higher in-cylinder velocity and turbulence.

* The CMCV decreased the exhaust gas temperature between 5-30'C at a given spark

timing, which was attributed to the increased in-cylinder heat transfer during com-

bustion due to the effectively less retarded combustion phasing (location of 50%

mass fraction burned) with the CMCV closed. However, at equivalent levels of

combustion phasing, there was close agreement between the CMCV open and

closed exhaust gas temperatures.

* No noticeable advantage on the exhaust HC emissions was found with the CMCV

closed at an equivalent combustion phasing in steady state idle operation, since the

exhaust HC mass flow rate with the CMCV closed was almost coincident with the

CMCV open case. This is mainly due to faster combustion and slightly higher peak

pressure, which have offsetting effects on the various sources of HC within the cyl-

inder.

* The use of the CMCV improved fuel consumption at retarded spark timings. Im-

provements in ISFC were obtained with the CMCV closed due to enhanced air-fuel

mixing and relatively faster burning, which caused the earlier end of combustion.

The ISFC at a given spark timing was reduced up to 8% relative to the CMCV open

case, as spark timing was retarded.

5.3 Engine Start-up Experiments

Engine start-up experiments were conducted with the three different geometric charge

motion control valves. Mixture preparation, combustion characteristics, and HC emissions

were evaluated with the various fuel injection and spark timing strategies during the first 20

82

seconds following engine cranking. The engine operation was divided into two distinct pe-

riods (transient period and quasi-steady period) at about 3 seconds, after which the engine

entered into the quasi-steady idle operation. The following conclusions were based on the

results of those engine start-up experiments.

Effects of the CMCV with the same fuel injection and spark profile

* The use of the CMCV was found to improve mixture preparation during the engine

start-up process. With the same fuel injection and spark timing trajectory, the

CMCV increased the amount of fuel evaporated relative to the CMCV open case.

This is mainly due to higher mixture velocities through the upper part of valve,

which increases atomization of liquid fuel at the valve seat, resulting in finer fuel

droplets and more rapid distribution and mixing with hot residual gas.

" Significant differences were observed between the different geometric charge mo-

tion control valves. The highest evaporation rate was obtained from the swirl-

tumble type CMCV (base type CMCV as shown in Figure 4-3). The liquid fuel

transport, distribution, and evaporation were affected by the combined swirl and

tumble flow motion, and correlated with combustion characteristics and HC emis-

sions. Moderately enhanced swirl and tumble motion, which was produced by the

base CMCV, showed the greatest potential for improving mixture preparation,

combustion characteristics, and HC emissions.

Effects of the CMCV with reduced fuel injection under the same spark profile

* The same in-cylinder air/fuel ratio profile as the CMCV open case was achieved

with the CMCV closed by injecting 7.5% less fuel during the transient period (0 to

3 seconds) and 3.5% less fuel during the quasi-steady period (3 to 20 seconds) due

to improved fuel transport and evaporation, while the same spark timing trajectory

was maintained.

83

* Charge motion impact on initial flame development was assessed by comparing the

burn rate analysis results with the same in-cylinder air/fuel ratio and spark timing

profile. The 0-10% combustion duration was reduced by 8' CA on average during

the transient period and 20 CA during the quasi-steady period due to enhanced in-

cylinder charge motion resulting from the use of the CMCV. Significant reductions

in the combustion duration were found with the CMCV closed, especially during

the transient period.

* Using the CMCV with the same in-cylinder lambda and spark timing profile, peak

engine-out HC emissions (excluding the first cycle) occurred in the earlier stages of

engine start-up process relative to the CMCV open case. Combining this with the

fact that the same burned fuel mass trajectory was maintained indicates that the

CMCV enhanced liquid fuel transport and evaporation rates, resulting in less liquid

fuel being deposited in the cylinder during the transient period. The engine-out HC

emissions during the quasi-steady period were reduced with the CMCV closed by

6.5%.

Effects of the CMCV with retarded spark under the same fuel injection profile

* Reduced combustion duration with the CMCV closed enabled additional spark re-

tard while injecting less fuel overall. Combustion stability worsened as combustion

was phased later. Moderate spark retard (AOsp = -5') was found to reduce the en-

gine-out HC emissions by 22% during the transient period and by 14% during the

quasi-steady period relative to the baseline spark timing (A0,= 0'). HC emissions

decreased at retarded spark timings due to decreased crevice volume fraction at

peak pressure and increased burned gas temperatures at the end of the expansion

and during exhaust, which increased post combustion oxidation.

Effects of the CMCV with reduced fuel injection and retarded spark profile

* The same in-cylinder air/fuel ratio profile as the baseline CMCV open case was

achieved with the CMCV closed by applying 7.5% reduced fuel and 8' CA spark re-

84

tard for the transient period and 3.5% reduced fuel and 5' CA spark retard for the

quasi-steady period due to increased fuel vaporization and burning rates.

* By the combination of reduced fuel injection and retarded spark timing strategy

with the CMCV closed, the combustion stability was increased during the transient

period. However, the COV of Net-IMEP during the quasi-steady period was wors-

ened from about 5% to 6%.

* A total reduction of 12% in HC emissions during the first 20 seconds of operation

was achieved by the charge motion control valve with the reduced fuel injection and

retarded spark timing. The engine-out HC emissions were reduced by 18% during

the transient period and by 7% during the quasi-steady period. This improvement

was mainly influenced by enhanced liquid fuel transport into the cylinder and in-

creased fuel vaporization, which allowed reduced fuel injection and prevented liq-

uid fuel deposited within the cylinder.

5.4 Overall Assessment and Recommendations

* The CMCV was found to improve the engine's cold start-up characteristics. In par-

ticular, greater improvement in mixture preparation, combustion stability, and HC

emissions were observed during the initial 3 seconds (transient period) before the

engine reached quasi-steady idle operation.

" Considering that a large fraction of the total HC emissions during the Federal Test

Procedure (FTP) is emitted during the first 20 seconds, and also that approximately

40% of the first 20 second HC emissions are emitted during the first 3 seconds, it

would be beneficial to use the charge motion control valve even for the first 3 sec-

onds of engine cold start-up.

* Different shape charge motion control valves, which produced different in-cylinder

flows, were found to yield the differences in fuel evaporation, combustion rate, and

engine-out HC emissions due to the combined effects of the swirl and tumble mo-

85

tion. Flow testing results indicate that better performance was obtained with the

CMCV geometries that produced moderate swirl and tumble intensities rather than

high swirl or tumble alone

* Greater fuel vaporization and faster burning rate with the CMCV closed allowed re-

duced fuel injection and additional spark retard, resulting in a significant reduction

in HC emissions during the engine cold start-up process.

* Since less liquid fuel deposits within the cylinder in the early stage of engine start-

up with the CMCV closed, more precise cycle-by-cycle fuel metering, which is re-

quired for the low emissions engine start-up, can be achieved for the first hundred

cycles by the use of the CMCV.

86

References

References

1. Heywood, J.B., Internal Combustion Engine Fundamentals, McGraw-Hill, New York,

1988.

2. Cheng, W.K., Hamrin D., Heywood, J.B., Hochgreb, S., Min, K., and Norris, M., "An

overview of hydrocarbon emissions mechanisms in spark-ignition engines", SAE Paper

932708, 1993.

3. Takeda, K., Yaegashi, T., Sekiguchi, K., Saito, K., and Imatake, N., "Mixture prepara-

tion and HC emissions of a 4-valve engine during cold starting and warm-up", SAE Pa-

per 950074, 1995.

4. Cheng, W.K. and Santoso, H., "Mixture preparation and hydrocarbon emissions behav-

iors in the first cycle of SI engine cranking", SAE Paper 2002-01-2805, 2002.

5. Meyer, R. and Heywood, J.B., "Evaporation of in-cylinder liquid fuel droplets in an SI

engine: A diagnostic-based modeling study", SAE Paper 1999-01-0567, 1999.

6. Henein, N.A., Tagomori, M.K., Yassine, M.K., Asmus, T.W., Thomas, C.P., and Hart-

man, P.G., "Cycle-by-cycle analysis of HC emissions during cold start of gasoline en-

gines", SAE Paper 952402, 1995.

7. Castaing, B.M., Coward, J.S., and Cheng, W.K., "Fuel metering effects on hydrocarbon

emissions and engine stability during cranking and start-up in a port fuel injected spark

ignition engine", SAE Paper 2000-01-2836, 2000.

8. Takahashi, H., Ishizuka, Y., Tomita, M., and Nishizawa, K., "Engine-out and tail-pipe

emission reduction technologies of V-6 LEVs", SAE Paper 980674, 1998.

9. Nishizawa, K., Momoshima, S., Koga, M., and Tsuchida, H., "Development of new

technologies targeting zero emissions for gasoline engines", SAE Paper 2000-01-0890,

2000.

10. Jung, H.H., Stein, R.A., and Leone, T.G., "Comparison of dual retard VCT to continu-

ously variable event valvetrain", SAE Paper 2004-01-1268, 2004.

11. Development Instrumentation Group, Delco Modular Development System Manual,

Delco Electronics Corporation, 1997.

87

12. Cambustion HFR400, April 2005

<http://www.cambustion.co.uk/instruments/hfr400>.

13. Cambustion NDIR500, April 2005

<http://www.cambustion.co.uk/instruments/ndir500/ndirprinciples.html>.

14. UTG-91 Fuel Properties, Chevron Phillips Chemical Company LLC, October 2004

<http://www.cpchem.com/enu/tdsunsecured/UTG-9 1.pdf>.

15. Klein, D. and Cheng, W.K., "Spark ignition engine hydrocarbon emissions behaviors in

stopping and restarting", SAE Paper, 2002-01-2804, 2002.

16. Fox, J.W., Cheng, W.K., and Heywood, J.B., "A model for predicting residual gas frac-

tion in spark-ignition engines", SAE Paper 931025, 1993.

17. Cheung, H.M. and Heywood, J.B., "Evaluation of a one-zone bum-rate analysis proce-

dure using production SI engine pressure data", SAE Paper 932749, 1993.

88

Appendix A

A. 1 Energy Release Equation

QCHa80

y-1

7

avp-+

80

where:

QCH fuel chemical energy

0 crank angle

y ratio of specific heat

p

V

cylinder pressure

cylinder volume

QCR energy in crevices

Q HT heat transfer

A.2 Residual Gas Fraction Calculation

Fox Model (modified for a 4-valve engine)

x, =1.266.OF( P 087

N P

145 2 LV-D,~OF= . (107+7.8.AO+A02). vnax

B B2

89

1 V2P +QCR +QHTy-1 ao ao ao

(j/ P)-07| P I +0.6 2 -#-

A.3 Swirl and Tumble Index

Swirl Index (SI) :

This is an estimate of the swirl number if the cylinder was filled at this valve lift and there

were no effects of compressibility.

= 2zSp"I g 2

i&d 8

Note: The term (nt2/8) is to correct for "mass weighting" of the mean piston speed when as-

signing an engine speed to match the flow rate of the steady flow bench.

SI swirl index

Ts swirl torque on honeycomb element (N-mm)

S engine stroke (mm)

Pcy = in-cylinder density at test conditions (kg/m 3)

Calculate pcyl from the ideal gas relation:

p = 1000 PD (kg288.17 TD

Note: This equation assumes the gas pressure and temperature at the valve-exit-plane are

the same as those within the cylinder.

PD and TD are pressure and temperature in the cylinder at each test point.

Tumble Index (TI) :

This is an estimate of the tumble number if the cylinder was filled at this valve lift and there

were no effects of compressibility.

90

2rcorSpy .C 2TI = -

12 8

Note: The term (n 2/8) is to correct for "mass weighting" of the mean piston speed when as-

signing an engine speed to match the flow rate of the steady flow bench.

TI = tumble index

S = engine stroke (mm)

Tcor corrected torque (N-mm)

PcyI in-cylinder density at test conditions (kg/M 3)

Calculate Tcor:

2cor = r meas 2 CAL

3 9B

Second term in correction equation is to convert from the moment of inertia for a cylinder

about its axis into the moment of inertia for a cylinder rotated about a central diameter.

S = engine stroke (mm)

B = engine bore (mm)

Tmeas = measured torque (N-mm)

Tcor = corrected torque (N-mm)

CAL = calibration factor to account for fraction of tumble angular momentum intercepted

and measured by sensing element. Current best estimate of calibration to use for WEC fix-

ture is 1.0.

91

effects of charge motion control during cold start of si

Documents