Small Engine Performance Limits Turbocharging, Combustion or Design

William Attard

Submitted in total fulfillment of the requirements of the degree of Doctor of Philosophy 2007

Department of Mechanical and Manufacturing Engineering The University of Melbourne Produced on archival quality paper

ii

You learn more from failure than success

iii

iv

Abstract
P O W E R E D B Y W A T T A R D

Growing concerns about interruption to oil supply and oil shortages have led to escalating global oil prices. In addition, increased public acceptance of the global warming problem has prompted car manufacturers to agree to carbon emission targets in many regions including most recently, the Californian standards. Other legislating bodies are sure to follow this lead with increasingly stringent targets. As a result of these issues, spark ignition engines in their current form will need significant improvements to meet future requirements. One technically feasible option is smaller capacity downsized engines with enhanced power that could be used in the near term to reduce both carbon emissions and fuel consumption in passenger vehicles. This research focuses on exploring the performance limits of a 0.43 liter spark ignited engine and defining its operating boundaries. they restrict small engine performance. Limiting factors such as combustion, gas exchange and component design are investigated to determine if The research gives direction to the development of smaller gasoline engines and establishes the extent to which they can contribute to future powertrain fuel consumption reduction whilst maintaining engine power at European intermediate class requirements. As no small OEM production engine could be adapted to evaluate this concept, a four valve, inline two cylinder engine was designed and constructed to withstand the high combustion and inertia forces associated with near two bar boost pressure and engine speeds exceeding 10,000 rev/min. The asymmetric (odd) fire configuration required the uneven flow conditions in the intake and exhaust systems to be accommodated in the engine design to reliably achieve 25 bar brake mean effective pressure. This is believed to be the highest recorded value for a small spark ignition engine operating on pump gasoline. In addition, the test engine achieved a maximum of 37% brake thermal efficiency.

Best engine performance, efficiency and CO2 benefits across normally aspirated, supercharged and turbocharged modes were found to match or exceed the capabilities of typical larger bore engines found in passenger vehicles. The case study performed determined the feasibility of replacing a 1.25 liter normally aspirated engine found in the 2007 Ford Fiesta with the development engine in the turbocharged mode. Results show that the power and hence acceleration Simulated performance of the larger engine could be readily matched with the smaller turbocharged unit, with a 66% reduction in engine capacity. performance over the New European Drive Cycle showed a possible 22% reduction in fuel consumption and CO2 emissions, including a reduction of 62% at idle conditions. These benefits are a consequence of operating the test engine closer to peak efficiency, together with engine and chassis mass reductions. The reduction in CO2 would shift the study vehicle well under the 2012 Euro target of 120 g/km. Across all modes, it was evident that the small test engine could operate with considerably higher compression ratio for a given manifold absolute pressure when compared to larger bore, lower speed engines. This was demonstrated with normally aspirated results showing potential for engine operation at a compression ratio exceeding 13. However, the dominant performance limiting factor was experimentally found to be abnormal combustion, specifically knock in the end-gas region, with the highest knock intensities deduced to occur on the intake side of the pent roof combustion chamber. Thus it is concluded that further efficiency gains are possible with higher octane fuel, as the turbocharged engine was knock limited at a compression ratio of 10. Extrapolation of the efficiency and maximum performance data for the compression ratio range of 9 to 13 could aid a well-to-wheel study defining what the optimum fuel octane number is, assuming that refinery energy requirements are known for different octane fuels.

vi

Declaration
P O W E R E D B Y W A T T A R D

I hereby declare that this thesis comprises only my original work towards the PhD, except where indicated by reference and acknowledgment in the text. I further certify that this thesis is less than 100,000 words, exclusive of tables, maps, bibliographies and appendices.

William Attard

vii

viii

Acknowledgements
P O W E R E D B Y W A T T A R D

After spending nearly a decade enrolled as a student at this university, the end has finally come following many amazing and heartbreaking experiences. Thinking back on how this project initially started six years ago, I would have never imagined it could lead to a PhD. As an undergraduate student who had taken two years study leave, I had aims of designing and constructing a specific Formula SAE engine. It has been a life changing experience that has consumed me and I am forever indebted to many people who contributed to the wellbeing of myself and my work throughout this soul searching journey. To my understanding family, my appreciation cannot be expressed in words to my father, mother and brother who provided me with the opportunity to further my education, supplying endless amounts of encouragement and instilling the belief that I could succeed. They each supported me in different ways throughout this journey. One cant forget my mothers excellent cooking, which always brought me home even after traveling to many parts of the world. Thank you. To my loving partner, Elisa Toulson, a lovely lady who I met in the thermodynamics laboratory who reminded me that life does not revolve around engines. Who would have thought spending so many hours in one place would have paid off so handsomely. You have brought balance to my life and been the most important aspect to come out of my time at university. I am struggling to find words to describe all her efforts, but most importantly thank you for always having time for me. To my supervisor, mentor and friend, Professor Harry Watson, a great man whose wealth of knowledge, many ideas and shear enthusiasm got me into and out of trouble on many occasions. Who would have thought we would have come so far together after that telephone call six years ago. Thanks for the opportunity and your continual support. Truly a great man who cares.

ix

To fellow students Steve Konidaris and Mohammad Ali Khan, thank you for spending many late nights and early mornings in the thermodynamics laboratory testing engines, retrieving broken bits and solving problems, not all to do with the engines. Your persistence and support during the most trying times of the engine development phase has made both of you my two closest friends. I am also grateful to academics, fellow students, friends and technical staff involved in this project. There are many names to mention but I am especially grateful to Dr Ferenc Hamori, Faisal Lodi, Phuong Pham, Mark Gledhill, Terry Karagounis and George Zakis for their generous assistance and continual support throughout various stages of my PhD. Additionally, I would like to thank Don Halpin and Ted Grange for their technical assistance, good humor and on occasion, turning a blind eye. Altogether, a great bunch of people to work with. Lastly, to the numerous sponsors outlined in Appendix A whose generosity made the UniMelb WATTARD engine possible, I offer my sincere thanks for supporting the excellent learning activity at the University of Melbourne.

Contents
P O W E R E D B Y W A T T A R D

Abstract Declaration Acknowledgements Table of Contents List of Figures List of Tables Nomenclature Chapter 1 - Introduction
1.1.1

v vii ix xi xxi xli xlv 1

1.1 Motivation for Research .............................................................................. 2 Initial Formula SAE Objectives ............................................................ 2 1.2 Research Objectives.................................................................................... 3 1.3 Outline of Thesis ........................................................................................ 3

Chapter 2 - Background and Review

2.1 Global Vehicle Trends ................................................................................. 7 2.2 Fuel Consumption....................................................................................... 9 2.3 CO2 Emissions and Global Warming ........................................................... 11 2.4 Spark Ignition versus Diesel Engines.......................................................... 14 2.5 Small Engines........................................................................................... 15 2.6 Turbocharging and Engine Downsizing....................................................... 16 2.7 Combustion.............................................................................................. 18 2.7.1 Abnormal Combustion......................................................................... 20 2.7.2 Knock Reduction Methods ................................................................... 21 2.8 Summary ................................................................................................. 21

Chapter 3 - Engine Design

23

3.1 Introduction ............................................................................................. 23 3.2 Design Brief ............................................................................................. 24

xi

3.3 Theoretical Analysis .................................................................................. 24 3.3.1 Engine Capacity and Configuration ...................................................... 24 3.3.2 CFD Simulation................................................................................... 27 3.4 Achieving Balance in the Design Process .................................................... 30 3.4.1 The Feasibility of OEM Engine Adaption ............................................... 30 3.4.2 OEM Component Adaption .................................................................. 31 3.4.3 New Components ............................................................................... 34 3.5 Rotating and Reciprocating Assembly......................................................... 35 3.5.1 Load Analysis ..................................................................................... 35 3.5.2 Connecting Rod.................................................................................. 39 3.5.3 Piston Assembly ................................................................................. 44 3.5.4 Crankshaft ......................................................................................... 47 3.6 Cylinder Block Design ............................................................................... 50 3.6.1 Gasketless Interface Design ................................................................ 51 3.6.2 Torque Plate Design ........................................................................... 57 3.6.3 Thermal Analysis ................................................................................ 60 3.6.4 Manufactured Design.......................................................................... 67 3.7 Manifold Design........................................................................................ 67 3.8 Final Specifications and Assembly .............................................................. 73 3.9 Summary ................................................................................................. 73

Chapter 4 - Experimental Apparatus and Methodology

77

4.1 Introduction ............................................................................................. 77 4.2 Engine Test Rig ........................................................................................ 78 4.2.1 Dynamometer and Control Unit ........................................................... 81 4.2.2 Dynamometer-Engine Coupling ........................................................... 82 4.2.3 Engine Cradle..................................................................................... 83 4.2.4 Cooling system................................................................................... 83 4.2.5 Electrical System ................................................................................ 84 4.2.6 Exhaust System.................................................................................. 86 4.3 Instrumentation and Data Acquisition ........................................................ 87 4.3.1 Fuel and Air Flow Measurement........................................................... 87 4.3.2 Exhaust Emissions Analyzer ................................................................ 88 4.3.3 Blow-by Measurement ........................................................................ 88

xii

4.3.4 Cylinder Pressure Measurement........................................................... 89 4.3.5 Data Acquisition and Pre-Processing .................................................... 89 4.4 Experiments ............................................................................................. 92 4.4.1 Experimental Objectives...................................................................... 92 4.4.2 Experimental Results Analysis.............................................................. 92 4.5 Test Modes .............................................................................................. 93 4.6 Test Methodology .................................................................................... 98 4.6.1 Test Matrix......................................................................................... 98 4.6.2 Test Sequence ................................................................................... 99 4.6.3 Test Procedure ..................................................................................101 4.6.4 Tuning Strategy.................................................................................102 4.7 Summary ................................................................................................103

Chapter 5 - Engine Development

105

5.1 Introduction ............................................................................................105 5.2 Barrel and Liner Gasketless Interface ........................................................108 5.2.1 Manufacturing and Assembly..............................................................108 5.2.2 Initial Testing ....................................................................................110 5.2.3 Failure Analysis and Rectification ........................................................110 5.2.4 Interface Performance .......................................................................113 5.3 Electrical System .....................................................................................114 5.3.1 Ignition System .................................................................................114 5.3.2 Reference and Synchronization Signals ...............................................115 5.4 Piston Assembly ......................................................................................117 5.4.1 Piston ...............................................................................................117 5.4.2 Piston Ring Pack................................................................................121 5.4.3 CR Variation ......................................................................................123 5.4.4 Piston Oil Cooling ..............................................................................124 5.4.5 Piston to Valve Clearance...................................................................126 5.4.6 Piston Skirt and Liner.........................................................................128 5.5 Inlet Manifold ..........................................................................................129 5.5.1 Reliability ..........................................................................................129 5.5.2 Fuel Injector Location ........................................................................131

xiii

5.5.3 Intake Runner Geometry....................................................................135 5.6 Exhaust Manifold .....................................................................................137 5.7 Engine Balance and Vibration ...................................................................139 5.8 Camshaft and Timing Chain .....................................................................142 5.9 Summary ................................................................................................144

Chapter 6 - Turbocharging

147

6.1 Introduction ............................................................................................147 6.2 Turbocharger Selection ............................................................................148 6.3 Oil Control in Throttled Compressors.........................................................149 6.4 Turbocharger Development......................................................................151 6.4.1 Oil Control ........................................................................................152 6.4.2 Cooling .............................................................................................158 6.4.3 Intake Boost Regulation.....................................................................159 6.5 Turbocharger Performance Optimization ...................................................161 6.5.1 Exhaust Manifold Geometry ...............................................................162 6.5.2 Inlet Manifold Geometry ....................................................................169 6.5.3 Valve Timing - Camshaft Specification ................................................172 6.5.4 Final Results .....................................................................................178 6.6 Final Turbocharger Matched Operation .....................................................179 6.7 Summary ................................................................................................184

Chapter 7 - Operating Limits and Experimental Results

187

7.1 Introduction ............................................................................................187 7.2 Operating Limits ......................................................................................188 7.3 Contour Plot Generation...........................................................................196 7.4 Performance Contours .............................................................................198 7.4.1 BMEP................................................................................................198 7.4.2 Brake Power .....................................................................................200 7.4.3 Spark Timing.....................................................................................202 7.5 Efficiency Contours ..................................................................................204 7.5.1 Lambda ............................................................................................204 7.5.2 BSFC and Brake Thermal Efficiency ....................................................206

xiv

7.5.3 Volumetric Efficiency..........................................................................210 7.5.4 Mechanical Efficiency .........................................................................212 7.6 Emissions Contours..................................................................................212 7.6.1 BSHC ................................................................................................214 7.6.2 BSNOx ..............................................................................................219 7.6.3 BSCO2 ...............................................................................................222 7.6.4 BSCO................................................................................................226 7.7 Summary ................................................................................................228

Chapter 8 - Performance, Efficiency and Emission Comparisons

229

8.1 Introduction ............................................................................................229 8.2 Carburetion and PFI Fuel Delivery.............................................................230 8.3 Intake Flow Restrictor Effects ...................................................................232 8.4 NA, SC and TC Mode Comparisons............................................................238 8.5 Comparison to FSAE Engines ....................................................................243 8.6 Comparison to Small Engines ...................................................................246 8.7 Comparison to Larger Bore Engines ..........................................................248 8.8 Extension to a Future Application - Feasibility for a 1.25 Liter Replacement .251 8.8.1 Performance .....................................................................................253 8.8.2 Fuel Consumption and CO2 Emissions .................................................255 8.9 Summary ................................................................................................260

Chapter 9 - Combustion Analysis

263

9.1 Introduction ............................................................................................263 9.2 Combustion in Small Engines....................................................................264 9.3 Data Post Processing ...............................................................................264 9.3.1 E-CoBRA ...........................................................................................266 9.4 NA Combustion........................................................................................268 9.5 TC Combustion ........................................................................................276 9.5.1 Ignition Energy..................................................................................284 9.6 Knock .....................................................................................................288 9.6.1 Knock Location ..................................................................................289 9.6.2 Implemented Knock Control Strategies ...............................................293

xv

9.6.3 Observations of Knocking Combustion ................................................293 9.7 Discussion...............................................................................................299 9.7.1 Comparison of Small versus Large Bore Combustion............................299 9.7.2 Further Suggested Knock Control Strategies........................................302 9.8 Summary ................................................................................................305

Chapter 10 - Conclusions

307

10.1 Introduction ..........................................................................................307 10.2 Research Achievements .........................................................................308 10.2.1 Mechanical Design and Development ................................................308 10.2.2 Experiments ....................................................................................310 10.2.3 Formula SAE ...................................................................................312 10.3 Conclusions to the Research...................................................................312 10.4 Recommendations for Future Work .........................................................315 10.4.1 Implementation into a Passenger Vehicle ..........................................317

Awards and Publications References Appendix A - Sponsor Acknowledgments B - Formula SAE

319 321 347 347 349

B.1 Introduction ............................................................................................349 B.2 What is FSAE?.........................................................................................349 B.3 Engine Rules and Regulations ..................................................................351 B.4 Improving Engine Performance ................................................................352 B.4.1 Specific Output .................................................................................352 B.4.2 Packaging .........................................................................................352 B.5 Review of Past Formula SAE Engines ........................................................353 B.6 New Engine Targets for Formula SAE .......................................................356

xvi

C - Fuel Properties and Emission Regulations

357

C.1 Fuel Properties ........................................................................................357 C.2 Emission Regulations ...............................................................................359

D - Restrictor Calculations

361

D.1 Maximum Theoretical Restrictor Mass Flow Rate .......................................361 D.2 Flow Through a Venturi ...........................................................................362 D.3 Compressible Flow Intake Model ..............................................................363 D.4 Pre-Restrictor Fuel Injection.....................................................................363 D.4.1 Restricted Mass Flow.........................................................................363 D.4.2 Manifold Air Temperature ..................................................................365 D.5 Maximum Power .....................................................................................366

E - Turbocharger Selection

367

E.1 Introduction ............................................................................................367 E.2 Aspect ....................................................................................................368 E.2.1 Availability and Cost ..........................................................................368 E.2.2 Matching...........................................................................................368 E.2.3 Implementation.................................................................................373 E.2.4 Control .............................................................................................373 E.2.5 Mass and Packaging ..........................................................................375 E.2.6 Cooling and Lubrication .....................................................................375 E.2.7 Final Selection ...................................................................................375

F - Engine CFD Simulation

377

F.1 Introduction ............................................................................................377 F.2 Ricardo WAVE Models ...........................................................................378 F.2.1 Cylinder Model ..................................................................................378 F.2.2 Pipe Model ........................................................................................379 F.2.3 Complex Pipe Junction Model .............................................................380 F.2.4 Turbocharger Model...........................................................................381 F.2.5 Turbocharger Wastegate Control Model ..............................................381 F.2.6 Muffler Model ....................................................................................383 F.2.7 Friction Model....................................................................................383

xvii

G - Transmission Ratio Optimization

385

G.1 Gear Ratio Optimization...........................................................................385 G.2 Final Matched Transmission .....................................................................386

H - Flow Bench Airflow Testing

389

H.1 Introduction............................................................................................389 H.2 Steady versus Pulsating Flow ...................................................................390 H.3 Super-Flow SF600 Flow Bench .................................................................390 H.4 Flow Test Rig and Experiments ................................................................392 H.5 Flow Test Results ....................................................................................393 H.5.1 Cylinder Head Mass Flow...................................................................394 H.5.2 Cylinder Head Discharge Coefficients .................................................395 H.5.3 Engine Discharge Coefficients ............................................................396 H.5.4 Intake Restrictor Nozzle Mass Flow ....................................................397

I - Engine Subsystem and Transmission Development

399

I.1 Introduction.............................................................................................399 I.2 Clutch Assembly.......................................................................................400 I.3 Transmission ...........................................................................................404 I.4 Lubrication System...................................................................................405 I.5 Water Cooling System ..............................................................................410

J - Test Rig Development and Calibration

413

J.1 Introduction ............................................................................................413 J.2 Brake Torque Measurement and Calibration...............................................414 J.3 Dynamometer-Engine Coupling Development and Calibration .....................415 J.4 Fuel Measurement and Calibration ............................................................419 J.5 Blow-by Measurement and Calibration.......................................................420 J.6 Cooling System Calibration .......................................................................421 J.7 Cylinder Pressure Measurement and Calibration .........................................422 J.8 TDC Alignment ........................................................................................425 J.9 Compression Ratio Calibration ..................................................................426

xviii

K - Piston Repair

429

K.1 Piston Repair Techniques .........................................................................429

L - Engine Run-In

431

L.1 Engine Run-In Procedure .........................................................................431

M - Exhaust Gas Analysis

433

M.1 ADS-9000 Exhaust Gas Analyzer ..............................................................433 M.2 Exhaust Gas Sampling Position ................................................................434 M.3 ADS-9000 Emission Correction .................................................................436 M.3.1 Eliminating Air Leakage .....................................................................436 M.3.2 Correcting for Hydrocarbon Type .......................................................436 M.3.3 Hydrocarbon Sensitivity.....................................................................437 M.3.4 Wet/Dry Analysis and Compensation..................................................437 M.4 AFR Calculation.......................................................................................438 M.4.1 Calculation from Air and Fuel Measurement........................................438 M.4.2 Oxygen Measurement .......................................................................438 M.4.3 Calculation from the Exhaust Products ...............................................439 M.5 Brake Specific Emissions Calculation.........................................................441 M.6 AFR Variation..........................................................................................441 M.6.1 Efficiency Correction to Alternate ....................................................441 M.6.2 Emissions Correction to Alternate ....................................................442

N - Combustion Modeling

447

N.1 Introduction............................................................................................447 N.2 Model Outline..........................................................................................447 N.3 Engine Geometric Relationships ...............................................................449 N.4 Combustion Modeling ..............................................................................450 N.4.1 Residual Gas Mass Fraction................................................................450 N.4.2 Compression and Expansion Process ..................................................450 N.4.3 Combustion Process ..........................................................................452 N.5 Heat Transfer..........................................................................................454 N.6 Chemical Equilibrium ...............................................................................456

xix

N.6.1 Chemical Equilibrium Solver Accuracy.................................................463 N.7 Flame Geometry .....................................................................................463 N.8 Laminar and Turbulent Flames.................................................................465 N.8.1 Laminar Flame Speed........................................................................466 N.8.2 Actual Flame Speed...........................................................................467 N.8.3 Flame Speed Ratio ............................................................................468 N.8.4 Turbulence Intensity .........................................................................468 N.9 Knock Amplitude .....................................................................................469

O - Error Analysis

471

O.1 Measurement and Calculation Error..........................................................471

P - Drawing Registry

473

xx

List of Figures
P O W E R E D B Y W A T T A R D

Figure 2.1: Current and projected trends in global motor vehicle registration, Year: 1945-2025 [1, 71, 237, 239]. ..................................................................... 7 Figure 2.2: World motor vehicle population per 1000 persons (1994 and 2005) [231, 240]................................................................................................. 8 Figure 2.3: Chinese motor vehicle population showing the rapid growth since 1990. Vehicles exclude the ~50 million motorcycles and ~20 million agricultural vehicles [237]. ......................................................................... 9 Figure 2.4: Crude oil prices and consumption heavily influenced by world events [194]. ..................................................................................................... 10 Figure 2.5: US motor vehicle fuel economy, heavily influenced by oil prices [194, 233]. ...................................................................................................... 10 Figure 2.6: (Left): Share of greenhouse warming due to different greenhouse gases [104]. (Right): Share of worldwide CO2 emissions from the combustion of fuel, by sector [106]. ........................................................................... 11 Figure 2.7: Global mean temperatures (land and ocean) [165, 168]. ................ 12 Figure 2.8: Vehicle CO2 emissions. (Black): Fleet target averages over the NEDC [46, 130]. (Blue): Alternative technologies. (Red): Californian regulations for the FTP 75 cycle [61]............................................................................... 12 Figure 2.9: Large vehicle class technologies and the effects on cost and CO2 emissions [43]. (A6) six speed automatic transmission, (DCP) dual camshaft phasing, (DCT) dual clutch six speed transmission, (DE-ACT) cylinder deactivation, (DVVL) discrete variable valve lift, (EACC) electric accessories, (EPS) electric power steering, (GDI-S) gasoline direct injection at stoichiometric, (G-HCCI) gasoline homogenous charge compression ignition, (HEV) hybrid electric vehicle, (HSDI) high speed diesel, (IA) improved alternator, (ICP) inlet camshaft phasing, (ISG) integrated starter generator, (TC) turbocharged. .................................................................................. 13 Figure 2.10: Cost benefit analysis for CO2 improvements over the NEDC using various technologies in SI and diesel engines [130, 174]............................ 15

xxi

Figure 2.11: The effects on CR and fuel consumption over the NEDC for NA and various capacity TC downsized engines with equal power output in the same test vehicle [174]. ................................................................................... 17 Figure 2.12: Normal and abnormal (heavy spark knock) combustion in the test engine. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11, 12 BTDC spark timing, peak knock amplitude = 3 MPa. .................................................... 19 Figure 3.1: Predicted engine air consumption needed to maintain choked flow operating conditions for varying swept capacities and operating conditions with model validation from previous experimental results........................... 26 Figure 3.2: Simulation results for the proposed two cylinder TC design and the Melbourne University FSAE teams previous engine (NA Suzuki GSX-R600), with Suzuki experimental results used for model validation. Both engines flow restricted. ............................................................................................... 28 Figure 3.3: Ricardo WAVE model block diagrams used to predict FSAE engine performance for the flow restricted condition. (Upper): NA Suzuki GSX-R600 model used for initial software validation. (Lower): Proposed highly TC engine model created to explore the expected performance gains from engine downsizing. ............................................................................................. 29 Figure 3.4: Rocker cover patterns and magnesium alloy casting....................... 35 Figure 3.5: Sectional view of the final engine design highlighting the rotating and reciprocating components. ....................................................................... 36 Figure 3.6: Crankshaft crankpin loads caused by inertia (reciprocating component movement) and gas pressure (combustion) forces at maximum operating conditions. (Upper): Combustion dominated force at maximum BMEP. (Lower): Inertia dominated force at maximum speed................................. 38 Figure 3.7: Evaluation of H and I section beam connecting rods. (Upper): Equal mass CAD models with beam cross section detail. (Lower): FEM model analysis comparing von Mises stress distribution with x 100 distorted profile shown..................................................................................................... 40 Figure 3.8: FEM model analysis comparing von Mises stress distribution for both connecting rod designs under bending loads (misalignment or abnormal pressure forces) with x 10 distorted profile shown. .................................... 41

xxii

Figure 3.9: Various combustion chamber geometries highlighting highest loading locations due to abnormal pressure forces (spark knock) thus favoring various connecting rod designs to minimize deflection and ensure bearing reliability. .............................................................................................................. 41 Figure 3.10: (Left): I section connecting rods during the manufacturing process. (Right): Manufactured items to specified designs and drawings. ................. 42 Figure 3.11: FEM model analysis comparing von Mises stress distribution for various connecting rod little-end oil hole locations (30 kN tensile load). (Left): Single central hole in high loaded region and resultant stress concentration. (Right): Two smaller holes. ...................................................................... 43 Figure 3.12: The design and manufacture of forged aluminum custom pistons were required due to the high boost levels employed. (Left): Off-the-shelf forged piston blank prior to machining. (Right): Manufactured piston design with single compression ring. ................................................................... 44 Figure 3.13: FEM model piston analysis comparing von Mises stress distribution for various gudgeon pin designs. .............................................................. 46 Figure 3.14: FEM crankshaft evaluation with bending and torsional loads. Note the higher stresses in the second cylinders crankpin due to the transmission drive path. .............................................................................................. 48 Figure 3.15: Manufactured crankshaft version. (Left): Original design. (Right): Altered crankshaft design as detailed in Chapter 5.7. ................................. 50 Figure 3.16: Section view highlighting the cylinder block assembly (barrel and liner) featuring the gasketless sealing arrangement with the cylinder head not installed. ................................................................................................. 54 Figure 3.17: Assembled section view highlighting the cylinder block assembly (barrel and liner) sealing arrangement. The cylinder head causes the barrel to elastically distort, thus creating a robust seal......................................... 56 Figure 3.18: FEM model analysis simulating cylinder head loading on the barrel and liner assembly. (Left): Von Mises stress distribution. (Right): Resultant displacement. .......................................................................................... 57 Figure 3.19: Cylinder head stiffness measurement, needed for torque plate design. (Left): Schematic highlighting loaded locations (F) and deflections (D). (Right): Physical testing performed. ................................................... 58

xxiii

Figure 3.20: Results from load deflection tests performed on the cylinder head to determine the relative cylinder head stiffness needed for the torque plate design. ................................................................................................... 59 Figure 3.21: FEM model analysis simulating torque plate effects to determine the thickness needed to represent the cylinder head. Figure 3.22: Figure 3.23: (Left): Vertical displacement. (Right): Vertical displacement along section A-A. ................. 60 and CAD models used in the thermal FEM analysis, also Temperature distribution on the combustion surface for the highlighting the division of the cylinder bore into five thermal regions. ....... 61 model, indicating that the adiabatic assumption through the intake exhaust plane can be assumed for the model. ................................................... 64 Figure 3.24: Cylinder head surface temperatures at peak combustion pressure The area highlighted along the symmetry plane line for the model.

indicates the reduced temperature due to the gasketless interface. ............ 65 Figure 3.25: Liner surface temperatures at peak combustion pressure along the symmetry plane line on the exhaust side for the model. ........................ 65 Figure 3.26: Cylinder head surface temperatures for peak combustion pressure and exhaust blowdown models along the planes defining the model. ..... 66 Figure 3.27: (Left): Manufactured aluminum cylinder barrel. (Right): Barrel with shrink fitted cast iron liners ready for engine installation. ........................... 67 Figure 3.28: (Left): Conventional plenum design. (Right): Watsons KEC log style rolling flow design where the kinetic energy of the flow is conserved in a vortex about the axis of the plenum. ........................................................ 68 Figure 3.29: Experimental data from a NA racing engine showing the effect of engine speed on the dynamic pressure waves in the inlet tract @ WOT [102]. .............................................................................................................. 69 Figure 3.30: Effective inlet tract lengths created from a wave propagation theoretical model highlighting the tuned engine speed for peak primary and secondary resonance at varying ambient conditions................................... 69 Figure 3.31: (Left): Measured, inspected and cleaned components prior to assembly. (Right): First complete engine assembly prior to dynamometer installation. ............................................................................................. 72 Figure 4.1: Final version of the developed experimental rig. (Upper): Control panel. (Lower): Experimental rig testing facility. ........................................ 79

xxiv

Figure 4.2:

Experimental setup illustrating the basic schematic layout of the

engine, including controllers, sensors and data acquisition systems. ........... 80 Figure 4.3: Heenan and Froude Dynamatic MK-1 eddy current dynamometer and OZY-DYN type MISD109-01 control system. .............................................. 81 Figure 4.4: Engine cradle used to mount and locate the test engine to the dynamometer. (Left): CAD image. (Right): Fabricated cradle frame, housing the test engine. ....................................................................................... 83 Figure 4.5: Schematic layout of the fuel mass measurement system, calibrated to provide fuel consumption rates................................................................. 87 Figure 4.6: Motoring raw cylinder pressure with crankshaft angle encoding from WaveView. Time versus voltage, NA - PFI, 6000 rev/min, WOT, CR = 10.... 91 Figure 4.7: CAD and actual NA modes with fuel delivery variation. (Upper): Mode A - NA with Carburetion (single carburetor shown in CAE model). (Lower): Mode B - NA with PFI............................................................................... 94 Figure 4.8: Externally driven Roots type supercharger assembly. Mode C - SC with PFI. ................................................................................................. 96 Figure 4.9: Mode D - TC with PFI, CAD model and final developed version. ....... 97 Figure 5.1: Test engine fitted into Melbourne University FSAE vehicles for further engine development, testing and validation. (Upper): 2003 vehicle. (Lower): 2004 vehicle...........................................................................................106 Figure 5.2: FEM model analysis simulating the H7-p6 shrink fit between the barrel and liner. (Left): Von Mises stress distribution. (Right): Resultant displacement. .........................................................................................109 Figure 5.3: FEM model analysis simulating both the shrink fit and torque plate effects. (Left): Von Mises stress distribution. (Right): Resultant displacement. .........................................................................................109 Figure 5.4: Failed gasketless interface with clear signs of leakage between the cylinder head and liner mating surfaces. High piston ring leakage is also evident. .................................................................................................110 Figure 5.5: FEM model analysis simulating the gasketless interface effects on the cylinder head. (Upper): Von Mises stress distribution. (Lower): Vertical displacement. .........................................................................................112

xxv

Figure 5.6: Crankshaft synchronization trigger edge (slot) development needed for high speed operation due to the higher torsional and cyclic velocity fluctuations associated with an odd firing twin configuration. ....................116 Figure 5.7: Effect of varying the number of compression rings (heat flux and oil control effects) and fuel quality on combustion. TC - PFI, 7000 rev/min, 200 kPa MAP, CR = 10, 12 BTDC spark timing. .............................................120 Figure 5.8: CR variations ranging from 13-9.6 through piston crown modifications. .............................................................................................................123 Figure 5.9: The effects due to a significant increase in underside oil cooling on the single compression ring piston design (Type A). Disassembly after ~5 hours operation at similar conditions. (Left): Dry crown. (Right): Chamber contamination with oil pooling and deposits. ............................................124 Figure 5.10: The effects of insufficient piston to valve clearances. (Upper): Poppet valve piston contact in individual cylinders. (Lower): Subsequent engine damage as a result of valve failure in Cylinder 2. ......................................127 Figure 5.11: Piston skirt and liner damage during engine development.............129 Figure 5.12: Fatigue crack repairs and continued failures in the lightweight aluminum alloy manifold discovered during development. .........................130 Figure 5.13: Initial polymer and final aluminum alloy inlet manifold versions....131 Figure 5.14: Intake manifold section highlighting the two differing injector setup positions used during development..........................................................132 Figure 5.15: Fuel mixtures at associated MAP, TP and engine speed during Note the fuel mixture excursion from lean to rich on transient vehicle testing for the high injector configuration (Setup 2) depicted in Figure 5.14. accelerator tip in, caused by fuel wall wetting due to the long mixing length. .............................................................................................................133 Figure 5.16: Varying trumpet geometry highlighting the actual and effective length, governed by the defined trumpet parameters (parameter A is a constant). ..............................................................................................135 Figure 5.17: (Upper): Constructed short and long runner intake manifolds. (Lower): Short and long runner engine air consumption comparison over theoretical, CFD and experimental testing. TC - PFI, 4000 rev/min, WOT. .136

xxvi

Figure 5.18: Exhaust manifold development due to the higher thermal and cyclic mechanical stresses at the outlet junction due to turbocharger installation. (Upper): Reliability concerns due to manifold cracking. (Lower): Thermal loads from experiments and FEM analysis highlighting plenum chamber stress concentrations due to vibration. ..............................................................138 Figure 5.19: (Left): Final constructed exhaust plenum design. (Right): FEM analysis resulting in a fourfold reduction in mechanical stresses when compared to the previous design in Figure 5.18 - Lower. ..........................139 Figure 5.20: Varying counterweight geometry designs used to alter the rotating and reciprocation balance proportions for differing crankshafts used in experiments. ..........................................................................................140 Figure 5.21: Failed crankshaft (Left): Heavy metal insert movement. (Right): FEM failure analysis (von Mises stress distribution) simulating the interference fit and the resulting plastic deformation in the alloy steel crank cheek. ......141 Figure 5.22: Engine damage caused by camshaft thrust bearing failure. ...........142 Figure 5.23: Consequences of negligence during the hard anodizing process with the eroding of both cast iron cylinder liners with close-up detail. ...............143 Figure 6.1: Test engine turbocharger layout and resulting upstream compressor throttling due to the mandatory throttle location and fitment of the intake restrictor. ...............................................................................................149 Figure 6.2: Sectional view of the Garrett supplied GT-12 turbocharger with the compressor and turbine housings removed. Figure 6.3: (Note): Water cooling jacket detail altered and oil feed relocated for illustrative purposes only. .............151 Schematic displaying the failed oil control strategy (Table 6.2: Strategy involves Strategy 3) to combat compressor throttled problems.

creating a pressure balance across the sealing system. .............................153 Figure 6.4: (Upper): Sectional view of the Garrett supplied GT-12 turbocharger. (Lower): Sectional view of the redesigned turbocharger to overcome oil consumption problems under throttled conditions. Note: Water cooling jacket detail removed and oil feed relocated for illustrative purposes only............155 Figure 6.5: Strategy 10 - Components used to overcome the oil control issue. (Upper): Modified centre bearing housing with silver soldered insert and orifice vent. (Lower): Rotating assembly with new compressor seal assembly. .............................................................................................................157

xxvii

Figure 6.6:

Boost control schematic highlighting how the raw MAP signal was

manipulated by the ECU. This enabled the wastegate movement to be ECU controlled to achieve the desired boost levels...........................................160 Figure 6.7: Simulated results displaying the performance discrepancies between constant pressure and pulse type turbocharging for the flow restricted test engine fitted with various exhaust manifold geometries. [ ] = Ease of manufacture / 5. ....................................................................................163 Figure 6.8: Simulated pressure pulsations in the turbine housing when applying pulse type turbocharging (Model A) to the odd firing test engine. TC - PFI, 6000 rev/min, single engine cycle............................................................164 Figure 6.9: Simulated brake power results comparing firing interval (odd and even) for constant pressure and pulse type turbocharging. .......................166 Figure 6.10: Varying exhaust manifold geometry. (Upper): Two into one pulse system (Model A). (Lower): Watsons KEC constant pressure plenum system (Model E). ..............................................................................................167 Figure 6.11: Experimental versus simulation BMEP results for constant pressure and pulse type turbocharging. TC - PFI, WOT, CR = 10. ..........................168 Figure 6.12: Experimental BMEP and spark timing results versus MAP for constant pressure and pulse type turbocharging. TC - PFI, 5000 rev/min, CR = 10. 168 Figure 6.13: Simulated results showing cylinder airflow variations for TC operation with varying intake volumes due to varying plenum chamber sizes. Airflow variations associated with charge robbing effects caused by the flow restriction and odd firing interval. ............................................................169 Figure 6.14: Experimental individual cylinder VOL and cylinder variations versus simulated results. Variations due to the charge robbing effects associated with the flow restriction and odd firing interval. TC - PFI, CR = 10, 6 L intake volume. .................................................................................................170 Figure 6.15: Experimental TC intake manifold air temperatures for various engine speeds and MAP. Results highlight the possibility of not intercooling the boosted charge for small engines if matched correctly. .............................171 Figure 6.16: Test engine simulated brake power results (WOT) with varying valve overlap for constant pressure type turbocharging. (Upper): 0.7 L exhaust volume. (Lower): 1.5 L exhaust volume. .................................................174

xxviii

Figure 6.17: Test engine simulated brake power results (WOT) with varying IVC timing for constant pressure type turbocharging. 2 L exhaust volume. ......175 Figure 6.18: Test engine simulated brake power results (WOT) with varying EVO timing for constant pressure type turbocharging. 2 L exhaust volume. ......176 Figure 6.19: Comparison of experimental and predicted engine performance for the test engine operating in the TC mode. ...............................................178 Figure 6.20: Garrett GT-12 compressor map with engine operating points overlaid for varying engine speeds and MAP. ........................................................180 Figure 6.21: Garrett GT-12 turbine map with engine operating points overlaid for varying engine speeds and MAP. .............................................................181 Figure 6.22: Intake and exhaust pressures and the effects on engine pumping work for varying engine speeds. (Upper): PMEP. (Middle): Exhaust pressure. (Lower): Log pressure-volume, highlighting the PMEP reduction for rising MAP at 6000 rev/min......................................................................................182 Figure 7.1: NA mode knock limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard to compensate for knock. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): NA - Carburetion. (Lower): NA - PFI. ...........................190 Figure 7.2: Boosted mode knock and airflow limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard and/or fuel enrichment to compensate for knock. Shaded areas indicate non-operation due to knock levels above the DL or limited airflow. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): SC PFI. (Lower): TC - PFI............................................................................192 Figure 7.3: TC - PFI mode (boost controlled) knock and airflow limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard and/or fuel enrichment to compensate for knock. Shaded areas indicate non-operation due to knock levels above the DL or limited airflow. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): 200 kPa MAP limited. (Lower): 150 kPa MAP limited.....194

xxix

Figure 7.4:

BMEP trends versus engine speed, MAP and CR.

Shaded areas

indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................199 Figure 7.5: Brake Power trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................201 Figure 7.6: Spark timing trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................203 Figure 7.7: fueling requirements versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Lower Left): SC - PFI. (Upper): NA - Carburetion. (Middle): NA - PFI.

(Lower Right): TC - PFI...........................................................................205 Figure 7.8: BSFC trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................207 Figure 7.9: Corrected BSFC trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI...........................................................................209 Figure 7.10:

VOL trends versus engine speed, MAP and CR.

Shaded areas

indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................211 Figure 7.11: MECH and exhaust pressure versus engine speed and MAP. Shaded areas indicate airflow limited regions. PL is the performance limit line. TC PFI, CR = 10. .........................................................................................212 Figure 7.12: Engine-out raw emission concentrations for varying and power outputs, achieved with engine speed and MAP variation. TC - PFI, CR = 10. .............................................................................................................213

xxx

Figure 7.13: BSHC trends versus engine speed, MAP and CR.

Shaded areas

indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................215 Figure 7.14: Corrected BSHC trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI...........................................................................217 Figure 7.15: BSNOX trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Lower Left): SC - PFI. (Lower Right): TC - PFI. .................219 Figure 7.16: Engine-out NOX emission concentrations for varying and power outputs, achieved with engine speed, MAP and CR variation. (Upper): NA Carburetion: MAP = 100 kPa. (Middle): NA - Carburetion: CR = 13. (Lower): SC - PFI: CR = 11...................................................................................221 Figure 7.17: BSCO2 trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................223 Figure 7.18: Corrected BSCO2 trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI...........................................................................224 Figure 7.19: Corrected BSCO2 trends ( = 0.9) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI...........................................................................225 Figure 7.20: Mole fraction CO trends for varied and brake power (engine speed and MAP variations). Shaded areas indicate unexplored regions. (Left): SC PFI. (Right): TC - PFI..............................................................................226 Figure 7.21: BSCO trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC PFI. .......................................................................................................227

xxxi

Figure 8.1: Performance, efficiency and emission effects for alternative CRs and fuel delivery (carburetion and air restricted PFI). NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): BMEP. (Middle): BSFC. (Lower): Engine-out HC mole concentrations. ..............................................................................231 Figure 8.2: Performance effects for odd (0.43 L, inline twin, CR = 13) and even (0.6 L, inline four, CR = ~12) fire configurations in restricted (NA - PFI) and unrestricted (NA - CARBS) modes. WOT, MBT spark timing......................233 Figure 8.3: Air consumption effects for restricted (NA - PFI) and unrestricted (NA CARBS) modes. WOT, CR = 13, MBT spark timing. (Upper): Individual cylinder VOL. (Middle): Engine air consumption and limitations. (Lower): Cylinder and mode airflow variations. ......................................................235 Figure 8.4: Predicted air consumption effects versus time at a given upstream condition (x) for odd (0.43 L, inline twin) and even (0.6 L, inline four) fire configurations in restricted and unrestricted modes. .................................236 Figure 8.5: Maximum MAP values achieved at the PL (WOT) across all test modes at the experimental test CR closest to the HUCR. .....................................238 Figure 8.6: Performance comparisons at the PL (WOT) for all modes at the experimental test CR closest to the HUCR. (Upper): BMEP. (Middle): Brake Power. (Lower): Spark Timing. ................................................................240 Figure 8.7: Efficiency comparisons at the PL (WOT) for all modes at the (Upper): VOL. (Middle): BSFC. experimental test CR closest to the HUCR.

(Lower): BSFC and TH corrected to = 1................................................242 Figure 8.8: WOT performance comparisons for a variety of engines tested for FSAE application. Tests conducted on the same dynamometer with engines fitted with a 20 mm intake flow restriction. ............................................245 Figure 8.9: WOT performance comparisons for a variety of small engines tested and compared to the test engine across NA and TC modes. Engine specifications outlined in Table 8.3. .........................................................247 Figure 8.10: WOT performance comparison between the Ford Fiesta 1.25 L NA engine and the smaller TC test engine used in experiments. (Upper): MAP. (Middle): BMEP. (Lower): Brake Power. Two performance curves are shown for the smaller test engine; (1) Limited MAP by wastegate (170 kPa), (2) Limited airflow by a restriction (20 mm). ...............................................254

xxxii

Figure 8.11: NEDC operating points for the Ford Fiesta chassis, used to compare fuel consumption and CO2 emissions for the OEM 1.25 L engine and the downsized 0.43 L test engine used in experiments. (Upper): Combined Urban and Extra Urban drive cycle forming the NEDC [61]. (Lower): Generated time frequency distribution for the NEDC [121]. ...............................................257 Figure 9.1: Crankshaft effects at high engine speeds throughout one cycle. NA Carburetion, 10000 rev/min, WOT, CR = 10. (Upper): Velocity fluctuations. (Middle): Torque fluctuations. (Lower): FEM analysis crankshaft elastic deformation. ..........................................................................................265 Figure 9.2: The effects of re-sampling the raw pressure trace to 0.5 CA increments, required for the E-CoBRA software input. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11............................................................................267 Figure 9.3: Pressure-volume diagram for 40 consecutive cycles highlighting the low combustion variability. NA - Carburetion, 10500 rev/min, WOT, CR = 13. .............................................................................................................268 Figure 9.4: Combustion burn duration effects for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): 0-10% MFB. (Middle): 10-90% MFB. (Lower): 0-90% MFB. Solid lines associated with CA axis, dashed lines with time axis......................................................................269 Figure 9.5: Combustion burn effects for alternative CRs and fuel delivery. NA, 10000 rev/min, WOT, = 0.9 0.02, MBT spark timing [ ]. (Upper): MFB versus CA. (Middle): MFB versus MFBR. (Lower): MFBR versus CA...........271 Figure 9.6: Estimated turbulence intensity u for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing. ...........................................272 Figure 9.7: Combustion flame speed effects for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): Actual flame speed @ 50% MFB. (Middle): Actual flame speed ratio @ 50% MFB. (Lower): Actual peak flame speed. ..................................................................................273 Figure 9.8: Combustion burn effects for varying engine speeds. NA - Carburetion, WOT, CR = 10, = 0.9 0.02, MBT spark timing [ ]. (Upper): MFB versus CA after Spark. (Middle): MFB versus MFBR. (Lower): MFBR versus CA.........275

xxxiii

Figure 9.9: Combustion variability and burn effects for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT......................................................................................................277 Figure 9.10: Combustion burn effects for varying MAP. TC - PFI, 7000 rev/min, CR = 10, spark timing [ ]. (Upper): MFB versus CA after Spark. (Middle): MFB versus MFBR (solid lines - left y axis) and MBR (dashed lines - right y axis). (Lower): MFBR versus CA. ............................................................279 Figure 9.11: Combustion pressure, temperature and flame speed effects for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.................................................................281 Figure 9.12: Combustion knock amplitude effects for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT......................................................................................................282 Figure 9.13: Combustion effects @50% MFB for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.283 Figure 9.14: Effect of varying ignition energies on boosted combustion over consecutive tests. TC - PFI, 6000 rev/min, 160 kPa MAP, CR = 10, = 0.85 +/- 0.02.................................................................................................286 Figure 9.15: Effect of varying ignition energies on flame development over consecutive tests. TC - PFI, 6000 rev/min, 160 kPa MAP, CR = 10, 17 - 18 BTDC spark timing..................................................................................287 Figure 9.16: Heavy knocking in the test engine well above the DL. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11, 12 BTDC spark timing. Firing, motoring and high pass filtered pressure traces......................................................289 Figure 9.17: Adverse effects of knock on engine components, with highest knock intensities identified to occur in the end-gas on the intake side of the pent roof combustion chamber. (Upper): Inlet side piston land failures. (Lower): Cylinder head deck and piston erosion near the bore periphery on the inlet side. ......................................................................................................291

xxxiv

Figure 9.18: Cylinder head and piston crown surfaces after engine operation, highlighting that the exhaust side is relatively hotter when compared to the intake in four valve, pent roof combustion chambers. ...............................292 Figure 9.19: Simulation results from Teraji et al. [222] validating experimental findings of the intake side being the most prone to high knock intensities. Integral value is the knock predictor. .......................................................292 Figure 9.20: The effects of knock on in-cylinder pressures across 10 consecutive cycles. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Raw pressure. (Middle): Knock amplitude. (Lower): Smoothed log pressure-volume.....................................................................................294 Figure 9.21: Varying knock intensities extracted from Figure 9.20. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Raw pressure. (Middle): Knock amplitude - high pass filtered pressure. (Lower): Smoothed pressure.................................................................................................295 Figure 9.22: Combustion burn effects for varying knock intensity cycles. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): MFB versus CA. (Middle): MFB versus MFBR. (Lower): MFBR versus CA. ....................296 Figure 9.23: Combustion burn effects for varying intensity knocking cycles. TC PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Actual flame speed versus CA. (Lower): Actual flame speed ratio versus CA. .......298 Figure 9.24: Proposed pent roof design to reduce the intake side knock propensity found from experiments. (Upper): Offset spark plug towards the intake side. (Lower): Effect of ignition point on combustion burning profiles and suggested knock location. .......................................................................303 Figure 10.1: The UniMelb WATTARD engine (TC - PFI mode) on display in the foyer of Building 170, Mechanical Engineering at the University of Melbourne. .............................................................................................................308 Figure B.1: International and local entries at the 2004 Australasian FSAE competition. ...........................................................................................350 Figure B.2: Two extremes in FSAE engine selection. (Left): OEM Yamaha YZF 0.45 L single cylinder motocross engine. (Right): Western Washingtons prototype 0.55 L V8. ...............................................................................355

xxxv

Figure D.1:

Subsonic, near sonic and supersonic flow for air ( = 1.4) in the

divergent section of an intake venturi. (Upper): Pressure ratio (P/PO). (Middle): Mach number (Ma). (Lower): Throat area ratio (A/A5) [245]. ......362 Figure E.1: Theoretical and assumed test engine air consumption needed to cause choked flow through the intake restriction for varying engine speeds. .......369 Figure E.2: Turbocharger limits and corresponding compressor pressure ratio and engine air consumption required to maintain choked flow over the desired operating speed range. ...........................................................................370 Figure E.3: Garrett GT-15 compressor map with WOT engine operating points over the choked operating speed range overlaid. .............................................371 Figure E.4: Garrett GT-15 turbine maps (VNT). ...............................................372 Figure E.5: Turbocharger exhaust turbine fitted with VNT technology highlighting the vane position movement. ..................................................................373 Figure F.1: Garrett GT-12 maps generated using Ricardos TC-map. (Upper): Compressor. (Lower): Turbine.................................................................382 Figure G.1: Engine vehicle matching through transmission design using the engine performance data from Chapter 3.3.2. (Upper): 6-speed transmission for the Suzuki GSX-R600 adapted to FSAE. (Lower): Optimized 3-speed transmission for the test engine..................................................................................387 Figure H.1: SF600 flow bench schematic [213]. ..............................................391 Figure H.2: Orifice plate measurement detail for AS 2360.1.1-1993 and BS 1042 standards...............................................................................................391 Figure H.3: Experimental flow rig setup. ........................................................392 Figure H.4: Port experimental mass flow values at various test pressures. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Inlet. (Lower): Exhaust. ........394 Figure H.5: Port discharge coefficients at various test pressures. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Inlet. (Lower): Exhaust. .......................................395 Figure H.6: Engine discharge coefficients versus CA. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Test pressure = 6 kPa. (Lower): Test pressure = 10 kPa.....396

xxxvi

Figure H.7: Intake restrictor experimental flow bench airflow data geometry outlined in Table H.2. Figure I.1: (Upper): Mass flow rate.

[78, 245]. (Lower):

Device A and B correspond to various Dall restrictor nozzle and diffuser Discharge coefficient...............................................................................397 FEM model analysis simulating varying spring loads applied to various pressure plate designs after clutch slippage due to turbocharging. Both pressure plate designs have similar mass and rotational inertia. (Left): Von Mises stress distribution. (Right): Resultant displacement..................401 Figure I.2: FEM model analysis simulating the clutch basket at peak torque due to turbocharging. displacement. (Left): Von Mises stress distribution. (Right): Resultant (Top): Original basket. (Bottom): Developed version with

circumferential stiffening ring. .................................................................403 Figure I.3: Developed clutch housing enabling a twofold increase in torque transmission in a confined space which was needed for boosted operation.404 Figure I.4: Transmission drive dog development needed for the high torque output associated with intake boosting. ...................................................405 Figure I.5: Irreparable crankshaft and connecting rod damage as a result of oil surge under high lateral vehicle acceleration, leading to big-end bearing failure. ...................................................................................................406 Figure I.6: Figure I.7: Dry sump oil path schematic, featuring internal oil and scavenge Brake performance effects for both wet and dry sump lubrication pumps with an external oil pressure relief valve........................................408 systems. ................................................................................................409 Figure I.8: Initial inadequate turbocharger water cooling circuit, resulting in the majority of flow bypassing the engine. .....................................................411 Figure J.1: Dynamometer torque calibration. (Upper): Methodology and calculation. (Lower): Actual load cell calibration........................................414 Figure J.2: Brake torque calibration, beam calibration versus dynamometer displayed. ..............................................................................................415 Figure J.3: Dynamometer coupling development. (Upper Left): Direct drive with torsional damping. (Upper Right): 1st iteration - Counter levered shaft with chain drive. (Lower Left): 2nd iteration - Chain drive with torsional dampening. (Lower Right): 2nd iteration with chain tensioner.......................................417

xxxvii

Figure J.4: Brake torque measurement correction to compensate for the increased losses associated with the chain drive when compared to the direct drive system...................................................................................................417 Figure J.5: Dynamometer coupling development (Upper): Direct coupling with torsional damping. (Middle): Chain drive. (Lower): Chain drive with torsional dampening.............................................................................................418 Figure J.6: Injector calibration using gasoline, Bosch 0-280-156-124. (Upper): Injector pulse width. (Lower): Duty cycle. ................................................420 Figure J.7: Volumetric calibration of the modified Fisher and Porter FP-1-20-G-8/78 rotameter used to determine blow-by rates (T = 20C, P = 101.3 kPa)......421 Figure J.8: Volumetric calibration of the modified Fisher and Porter FP-1-35-G10/83 rotameter used to determine cooling system flow rates (T = 20C, P = 101.3 kPa). ............................................................................................422 Figure J.9: Pressure transducer calibration (Kistler 601-B1, SN: C65704). .........423 Figure J.10: Mounting comparisons for cylinder pressure transducers in combustion chambers and resulting pressure data quality effects. (Upper): Non-flush mounting with connecting bore. (Lower): Flush mounting [129]. .............................................................................................................423 Figure J.11: Experiment in-cylinder pressure oscillation effects due to the nonflush mounted transducer. Single cycle oscillations shown to be independent of fuel quality and transducer type. SC - PFI, 6000 rev/min, 150 kPa MAP, CR = 11, 15 BTDC spark timing. .................................................................424 Figure J.12: Pressure transducer setup and mounting. (Left): Combustion chamber and pressure transducer location. (Right): Pressure transducer and associated components allowing direct flush mounting into the chamber. ..425 Figure J.13: Motoring trace, gathered with a first cylinder WOT ignition cut. NA PFI, 6000 rev/min, 100 kPa MAP, CR = 10, drift compensated, single cycle. .............................................................................................................426 Figure K.1: Piston repairs needed due to failures and long lead times. (Left): Welded piston. (Right): Final repaired product with no failures recorded...430 Figure L.1: Engine run-in procedure. NA - Carburetion. .................................431 Figure L.2: Blow-by effects upon returning to idle conditions during the engine run-in procedure shown in Figure L.1. New pistons and rings fitted. NA Carburetion, fast idle speed (2300 rev/min), 50 kPa MAP, CR = 10............432

xxxviii

Figure M.1: ADS-9000 and HEGO sensor sampling position. (Left): Dynamometer test rig setup. (Right): Vehicle setup........................................................435 Figure M.2: BMEP effects for varying . NA - PFI, 6000 rev/min, WOT, CR = 10. .............................................................................................................442 Figure M.3: Block diagram displaying the carbon balance methodology used to correct BSCO2 emissions to constant values. Methodology also applied to BSCO emissions......................................................................................444 Figure M.4: Individual cylinder HC emission concentrations (ppm C6) for gasoline with and power variation (engine speed and load) within a given data set. TC - PFI, CR = 10. ..................................................................................445 Figure N.1: Solution procedure flow chat for E-CoBRA. ....................................448 Figure N.2: Basic geometry of the reciprocating IC engine. ..............................449 Figure N.3: Open system boundary for single-zone combustion chamber. .........451 Figure N.4: Open system boundary for a two-zone combustion chamber. .........452 Figure N.5: Specific heat at constant pressure cp/R, as a function of temperature for species CO2, H2O, O2, N2, H2, and CO [164].........................................460 Figure N.6: Fuel Isooctane, equilibrium burned gases as a function of equivalence ratio at T = 1750 K, 2250 K, and 2750 K, at 30 bar. (Upper): Molecular weight. (Middle): Specific heat. (Lower): Ratio of specific heats. Results from E-CoBRA.............................................................................461 Figure N.7: Fuel Mole fraction of equilibrium combustion products of isooctaneair mixture as a function of equivalence ratio. (Upper): T = 1750 K, (Middle):

T = 2250 K, (Lower): T = 2750 K. Results from E-CoBRA. .......................462

Figure N.8: Spherical flame geometry in a combustion chamber.......................464 Figure N.9: Flame geometry at varying crank angle (BTDC) for the test engine. Pent roof chamber, CR = 10. (Upper): Burned volume vs. flame radius. (Lower): Flame area vs. flame radius. ......................................................465 Figure N.10: (Upper): Schematic of wrinkled turbulent flame structure [99]. (Lower): Turbulent structure of jet during intake [98, 99]. ........................466 Figure O.1: In-cylinder pressure traces for 40 consecutive cycles with average pressure and standard deviation highlighted. NA - Carburetion, 10500 rev/min, WOT, CR = 13. .........................................................................472

xxxix

xl

List of Tables
P O W E R E D B Y W A T T A R D

Table 2.1: The effect of various parameters on the ONR of a SI engine [90, 154, 169, 195, 217, 220, 226, 264]. ................................................................. 22 Table 3.1: Table 3.2: H7-p6 locational shrink fit upper and lower limits at ambient Boundary conditions used in all steady state thermal FEM analysis temperature and worse case conditions. ................................................... 55 (Upper): Combustion. (Lower): Exhaust blow-down. Note the higher liner temperatures at the bottom of the bore due to the absence of water cooling and exposure to crankcase oil. ................................................................. 63 Table 3.3: Test engine specifications. .............................................................. 72 Table 4.1: Electrical system fuel and ignition modules together with ECU sensor inputs required for engine operation. ........................................................ 86 Table 4.2: PC-30D analog to digital converter specifications.............................. 91 Table 4.3: Fuel delivery specifications for the carbureted and PFI systems........ 93 Table 4.4: Intake manifold geometry details for all test modes to suit the varying fuel delivery systems. .............................................................................. 95 Table 4.5: Defining the test modes desired maximum operating conditions....... 98 Table 4.6: Engine speed versus MAP test matrix at the CR closest to the HUCR. 98 Table 4.7: Engine speed versus CR test matrix at the MAP defined in Table 4.6. 98 Table 4.8: Knock and damage limits previously defined by Rothe et al. [192]...100 Table 4.9: Tuning strategy employed over the varying load conditions. ...........103 Table 5.1: Engine reliability development sequence in chronological order associated with NA and boosted engine testing. .......................................107 Table 5.2: Varying piston designs used during development, differentiated by the number of compression sealing rings. ......................................................117 Table 5.3: Different piston ring packs tried throughout development across NA and boosted modes. W: Width, P: Profile, M: Material..............................122 Table 5.4: Varying crankshaft design specifications used in experiments..........140 Table 6.1: Decision table used in turbocharger selection. Details in Appendix E.2. .............................................................................................................148

xli

Table 6.2: Methods and resulting affects for controlling turbocharger oil consumption under compressor throttled conditions. TC - PFI, 2300 rev/min (fast idle speed), 50 kPa MAP, 3 bar turbocharger oil feed pressure...........152 Table 6.3: Summary of test engine TC simulation results with NA and TC camshafts. Valve timing changes and associated improvements. ..............177 Table 7.1: Chapter 7 contour and cross plot data expressed over multi-domains with parameter variation. ........................................................................188 Table 7.2: Estimated HUCR and the experimental test CR which was closest to the estimated value. ...............................................................................193 Table 8.1: Specifications for odd and even fire engines compared in Figure 8.2. .............................................................................................................232 Table 8.2: Engine specifications and performance data for a range of in-house developed FSAE engines across NA and boosted modes............................244 Table 8.3: Engine specifications and performance data for a range of small engines across NA and boosted modes. ...................................................246 Table 8.4: Comparing the best performance, efficiency and emissions of the test engine with typical larger bore engines found in passenger vehicles. .........249 Table 8.5: Specifications for the Ford Duratec engine fitted to the 2007 Fiesta Mark VII series, which is compared against the TC - PFI test engine..........253 Table 8.6: Possibilities of adapting the smaller test engine into the Fiesta chassis and the effects on vehicle performance, fuel consumption and CO2 emissions against the standard OEM vehicle. ...........................................................256 Table 8.7: Vehicle comparisons for several configurations involving the Fiesta chassis fitted with both OEM and smaller engines. The effects of vehicle mass on the power required to drive the vehicle at equal speeds over the NEDC and the resulting effects on fuel consumption and CO2 emissions.....................259 Table 8.8: Summary of fuel consumption and CO2 emission benefits arising from replacing the larger 1.25 L NA engine found in the Ford Fiesta with the smaller 0.43 L TC test engine. Compiled from Tables 8.6 and 8.7. ............260 Table 9.1: Implemented knock control strategies and their affects. .................293 Table 9.2: Combustion comparison between small bore (test engine) and large bore (Ford Falcon [25, 120]) engines at stoichiometric conditions..............301 Table B.1: FSAE event points breakdown [197]...............................................350

xlii

Table B.2: Engine power to vehicle mass comparisons for two extreme cases in FSAE engine selection. ............................................................................353 Table B.3: Advantages and disadvantages of differing FSAE engine packages. ..353 Table C.1: Shell Optimax fuel specifications [203]. ..........................................357 Table C.2: Properties of various fuels used in SI engines [228]. .......................358 Table C.3: Documented and measured RON numbers for fuels used during development and experiments.................................................................358 Table C.4: Current and future US federal emissions regulations and Californian standards for light-duty vehicles [56, 61]. ................................................359 Table C.5: Future Californian fleet average GHG emission standards [61]. ........359 Table C.6: EU emission standards for passenger cars (Category M1*), g/km [61]. .............................................................................................................360 Table G.1: FSAE trim gear reduction ratios for the Suzuki GSX-R600 and the test engine. ..................................................................................................387 Table H.1: Test conditions and configurations trialed in airflow experiments. ....393 Table H.2: Dall restrictor geometry outlined in Figure H.7. ...............................398 Table M.1: ADS-9000 measurement ranges and accuracy [20]........................434 Table M.2: ASD-9000 sensitivity to various fuels [62, 266]..............................437 Table M.3: Raw engine-out emission data from the first cylinders ADS-9000 emissions analyzer. TC - PFI, 8000 rev/min, 70 kPa MAP, CR = 10. ..........440 Table M.4: Actual emission concentrations after Table M.3 correction..............440 Table N.1 Constant coefficients amn. ...............................................................458 Table N.2 Constant coefficients fmn and g0n. ....................................................458 Table O.1: Independent measurements, test ranges and associated errors. .....471 Table O.2: Error propagation from Table O.1. ................................................472

xliii

xliv

Nomenclature
P O W E R E D B Y W A T T A R D

Abbreviations
ABDC ACEA ADC AFR ASTM ATDC BBDC BDC BMEP BSCO BSCO2 BSFC BSHC BSNOX BTDC CA CAD CAE CAFE CAP CARB CDI CFD CFR CI CNC CO CO2 COG COP CoV CR CVT after bottom dead center European Automobile Manufacturers' Association analogue to digital converter air-fuel ratio American Society for Testing Materials after top dead center before bottom dead center bottom dead center brake mean effective pressure brake specific carbon monoxide brake specific carbon dioxide brake specific fuel consumption brake specific hydrocarbons brake specific oxides of nitrogen before top dead center crank angle computer aided design computer aided engineering Corporate Average Fuel Economy crankcase absolute pressure California Air Resource Board capacitive discharge ignition computational fluid dynamics Cooperative Fuel Research compression ignition computer numerically controlled carbon monoxide carbon dioxide center of gravity coil on plug coefficient of variation (standard deviation/mean) compression ratio continuously varying transmission

xlv

DAQ DI DISI DL DOHC DSG E-CoBRA ECU EGR EMI EU EVC EVO FEM FID FMEP FSAE FSR FTP GHG GM HC HCCI HEGO H2 H 2O HSDI HT HUCR ID IMEP IPCC IVC IVO K KA KEC KL L

data acquisition direct injection direct injection spark ignition damage limit double over head camshafts direct shift gearbox experimental combustion burn rate analysis engine control unit exhaust gas recirculation electromagnetic interference European Union exhaust valve closed exhaust valve open finite element method flame ionization detector friction mean effective pressure Formula SAE flame speed ratio (ST,a /SL) Federal Test Procedure greenhouse gas General Motors hydrocarbon homogeneous charge compression ignition heated exhaust gas oxygen hydrogen water high speed diesel high tension highest useful compression ratio inside diameter indicated mean effective pressure Intergovernmental Panel on Climate Change inlet valve closed inlet vale open Kei knock amplitude kinetic energy conservation knock limit liter

xlvi

LFL LHV MAP MAT MBT MBR MFB MFBR MPS NA NDIR NEDC NOX NVH O2 OD OEM ONR OPEC PCV PFI PL PMEP POT ppm RFI RON RTV SC SG SI SLS ST TC TDC TIG TP TWC UBA

lower flammable limit lower heating value manifold absolute pressure manifold air temperature maximum brake torque mass burned rate mass fraction burned mass fraction burned rate mean piston speed normally aspirated non-dispersive infra-red New European Drive Cycle oxides of nitrogen noise, vibration and harshness oxygen outside diameter original equipment manufacturer octane number requirement Organization of the Petroleum Exporting Countries positive crankcase ventilation port fuel injection performance limit pumping mean effective pressure part open throttle parts per million radio frequency interference research octane number room temperature vulcanizing supercharged spherical graphite spark ignition selective laser sintering spark timing turbocharged top dead center tungsten inert gas throttle position three-way catalyst German Federal Environmental Agency

xlvii

UFL UK US VNT VTG VVT WOT

upper flammable limit United Kingdom United States variable nozzle turbine variable turbine geometry variable valve timing wide open throttle

Symbols
A b c cp cv C CD Cf D g hc hfg L M Ma n N P Q R r s SL ST,a T u
area bore diameter speed of sound specific heat at constant pressure specific heat at constant volume heat transfer coefficient discharge coefficient friction coefficient diameter acceleration due to gravity heat transfer coefficient enthalpy of the phase change from liquid to vapor length mass flow rate molecular weight Mach number number of moles engine rotational speed pressure heat transfer gas constant radius stroke length laminar flame speed actual turbulent flame speed temperature turbulence intensity

xlviii

U V VC VCR VS v W COMP MECH TH VOL

internal energy volume clearance volume vapour compensation ratio swept volume voltage work transfer angle ratio of specific heats (cp /cv) compressor efficiency mechanical efficiency brake thermal efficiency volumetric efficiency crank angle relative air fuel ratio dynamic viscosity kinematic viscosity density equivalence ratio magnetic flux combustion chamber average gas velocity

xlix

CHAPTER 1
P O W E R E D B Y W A T T A R D

Introduction

In recent times, research into spark ignition (SI) engine downsizing has grown in popularity [130, 136, 174] as governments begin to limit carbon dioxide (CO2) emissions and consumers strive for cost savings due to rising oil prices. Thus, manufacturers are trying to improve performance and efficiency while meeting legislative pollutant emissions standards. Downsizing larger engines appears to be a major way forward in satisfying consumer and manufacturer requirements. For downsized engines to be comparable to their larger counterparts, the specific output performance must be increased by a ratio equal to the reduction in engine size. This high specific output can only be achieved with the help of increased engine speeds and/or intake boosting. This increases the induced amount of air and fuel, thus enabling the performance of the downsized engine to be improved to match its larger counterpart. Turbocharging seems to be the most acceptable solution to meeting the requirements, with high pressure ratios achievable and well documented improvements in efficiency [9, 42, 123, 174, 191, 248]. Engines found in compact sized passenger vehicles in todays automobile marketplace are considered small, but are usually more than twice the capacity of the engine designed, developed and tested for this research. Furthermore, the technology needed to replace these larger engines with smaller capacity downsized units already exists in normally aspirated (NA) versions, which are commonly mass produced for motorcycle and other applications. Applying turbocharger technology to smaller engines provides an economically viable solution in the near term prior to the implementation of alternative powertrain technology (fuel cell, electric or hybrid).

This research parallels steps already present in the market place for diesel engines and is essential in giving direction to the development of smaller gasoline engines [65, 118], which will be needed to meet both the worlds future powertrain and environmental requirements. This work differs from the Japanese Kei (K) class of passenger vehicles, which feature similar capacity and reduced performance engines fitted to microcars, as found in the Smart vehicle range [205]. The research presented focuses on replacing engines found in compact sized regular passenger vehicles.

1.1 Motivation for Research

The original intent of this development program was to achieve success in Formula SAE (FSAE) competition (Appendix B) by using a far superior engine package when compared to conventional original equipment manufacturer (OEM) motorcycle units. However, from the research and development process, results from this small engine operating in a variety of modes have significance in defining the performance limits for small capacity downsized SI engines. Results may give some insight to the extent by which engines can be downsized and the specific areas where future research should be directed. This has significant relevance to manufacturers, who continue to strive for swept capacity reductions, while maintaining performance, improving efficiency and meeting emission regulations.

1.1.1 Initial Formula SAE Objectives

Initially, the project surrounded designing, constructing and developing a specific downsized powertrain for FSAE (Appendix B.6), for fitment into Melbourne University vehicles for competition. The specific FSAE objectives when compared with the typical four-cylinder 0.6 liter (L) reference engine were to: Maintain maximum power over a wide operating engine speed range Increase maximum power Improve engine packaging and vehicle dynamic characteristics Reduce fuel consumption

1.2 Research Objectives

The intent of this project is to determine the performance limiting factors of small (~0.5 L) spark ignition internal combustion engines for passenger vehicles, as engine downsizing grows in popularity. Specifically, can the theoretical performance (work per cycle) and thermal efficiency reductions associated with the smaller bore size and cylinder capacity be overcome to be comparable to or surpass that of typical larger bore engines found in current passenger vehicles? This requires answers to questions about the effects and limitations of engine component design, turbocharging, intake and exhaust systems and combustion on small engine performance. The aims of the research may be summarized with the following questions: 1. Can the mechanical and thermal aspects of the component design be overcome so that they do not compromise engine performance limits? 2. Does the engine configuration, specifically the reduction in the number of cylinders and associated torque pulsations and vibrations restrict engine performance? 3. Can the gas flow abnormalities (intake and exhaust) associated with the reduced number of cylinders and unequal firing spacing be overcome so as not to be detrimental to performance? 4. Is the engine performance limited by the turbocharger; particularly can the compressor deliver enough boost, what impact does rotational inertia (transient response) have and is the turbocharger mechanical design adequate? 5. Are there any combustion limitations at high engine speed, manifold absolute pressure (MAP) or compression ratio (CR)? 6. How do emissions, especially CO2, compare to modern larger passenger vehicle engines with similar power output?

1.3 Outline of Thesis

The present chapter introduces the motivations behind the research. This includes how the project initiated from producing a specific FSAE engine and how the project evolved into more complete PhD research on small engine performance limits. In addition, the aim and research questions are outlined.

Chapter 2 provides background information on increasing global vehicle trends and the resulting effects on fuel consumption and global warming. order to mitigate growing world concerns. The third chapter details some aspects of the engine design, with focus placed on key features and novel aspects which were vital in ensuring reliability for the highly turbocharged (TC) package. Chapter 4 describes the experimental apparatus, including a description of the engine test rig, instrumentation and data acquisition systems used to complete experiments. An outline of the experimental objectives and results analysis is also presented. Furthermore, the test modes and methodology used to complete the research are described, with the test modes covering NA, supercharged (SC) and TC engine operation. Engine development needed to commission the engine design is presented in the fifth chapter. This includes dynamometer and in-vehicle development completed throughout the experimental testing phase. Chapter 6 presents turbocharger implementation, including the selection and development needed to successfully achieve boosted operation with the throttled compressor intake layout. Performance optimization, considerably aided by Turbocharger simulation tools and experimental testing is also described. together with the valve events is also presented. The seventh chapter presents the operating limits and experimental results over varying engine speed, CR, MAP and fuel delivery domains. Contour plots are used to display performance, efficiency and emission experimental results over these domains. Chapter 8 compares the performance, efficiency and emissions across test engine modes. In addition, comparisons are made to other small and larger engines An in-depth comparative found in FSAE and passenger vehicles respectively. vehicle with the TC test engine is also presented. Engine downsizing and turbocharging are also compared to many other technologies in

implementation which involved the optimization of the intake and exhaust systems

analysis of hypothetically replacing a 1.25 L NA engine found in a Ford Fiesta

Discussion of combustion in small engines is presented in Chapter 9 over varying engine speed, MAP, CR and fuel delivery domains. Differences and similarities to larger bore engines are also highlighted. The effects of ignition energy and spark knock on boosted combustion are presented, with further in-depth analysis on knock in the ultra high brake mean effective pressure (BMEP) condition. Chapter 10 summarizes the research achievements and conclusions. Furthermore, recommendations for future work are presented, including the requirements for replacing a modern production engine with a smaller downsized option in the compact sized, regular passenger vehicle class.

CHAPTER 2
P O W E R E D B Y W A T T A R D

Background and Review

2.1 Global Vehicle Trends

Global vehicle numbers have been increasing rapidly ever since the early part of the 20th century, when mass production of the automobile was made possible with the introduction of the moving assembly line by Henry Ford. In 2003, it was reported that there are almost 800 million motor vehicles (excluding motorcycles) driving on the worlds roads [240], a 15 fold increase since World War II (Figure 2.1).
1200 World Motor Vehicles (Millions) 1000 800 600 400 200 0 1945 Motorcycles Commercial Cars Projected (Cars + Commercial)

1955

1965

1975

1985

1995

2005

2015

2025

Figure 2.1: Current and projected trends in global motor vehicle registration, Year: 19452025 [1, 71, 237, 239].

Since 1970, the global fleet has been steadily expanding at approximately 16 million vehicles per year [257]. By 2030, it is anticipated that this linear growth will result in 1.2 billion cars and commercial vehicles driving on the worlds roads [71]. Over the past century, motor vehicle growth has been influenced by a growing population, but more importantly as a result of the car becoming a more affordable and convenient form of transport. Furthermore, the status that has been associated with owning and driving a motor vehicle in the past has increased sales and is continuing to increase sales, especially in developing countries. Figure 2.2 displays per capita car ownership in different regions of the world in both 1994 and 2005 [231, 240]. Car ownership is highest in developed countries, with the United States (US) leading the way with over 800 motor vehicles registered per 1,000 people in 2005. However, in the past 10 years vehicle growth in developed countries has slowed and even declined in some countries such as Canada. In contrast to the slowing vehicle growth in wealthy nations, motor vehicle registrations have been exploding in the rapidly developing economies of Asia. However, vehicle population in Asian countries is still quite low, with China having less than 4% of the vehicles per 1,000 people when compared to the US in 2005. Figure 2.3 displays Chinas 11.6% annual motor vehicle growth rate since 1990. This trend is projected to continue, with a large contributing factor being the 23% annual growth rate in private vehicles as a result of a growing portion of the Chinese population able to afford a motor vehicle. The growth in motor vehicle numbers has placed increased strain on the worlds natural resources and environment.
United States Canada Pacific Europe, W est Europe, East South America Asia, Middle East Asia, Far East Africa China 0 100 200 Motor Vehicles per 1000 People 300 400 500 600 700 800 900 2005 1994

Figure 2.2: World motor vehicle population per 1000 persons (1994 and 2005) [231, 240].

20 Motor Vehciles (Millions) Total Vehicles 15 Private Vehicles

11.6% Annual Growth Rate

10

23% Annual Growth Rate

0 1990

1992

1994

1996

1998

2000

2002

Figure 2.3: Chinese motor vehicle population showing the rapid growth since 1990. Vehicles exclude the ~50 million motorcycles and ~20 million agricultural vehicles [237].

2.2 Fuel Consumption

The increasing vehicle population has been accompanied by a similar increase in fuel consumption and hence production [257]. Figure 2.4 displays oil consumption, production and cost trends since 1940. The world demand for oil has been steadily increasing, with currently over 80 million barrels of oil consumed each day. The increasing demand for oil has resulted in conflict in some of the worlds oil producing regions, which has led to price fluctuations evident since the 1970s, when the Organization of the Petroleum Exporting Countries (OPEC) oil shocks initiated price instability. Several major fluctuations are highlighted in Figure 2.4, including that caused by the OPEC oil shocks in the 1970s, the Gulf war and most recently a combination of Asian growth coupled with the recent situation in Iraq. As a result, oil prices are currently over 60 $US per barrel and future prices are uncertain. Although the worlds demand for oil has been increasing, there is only a finite amount of oil beneath the earths crust. Therefore, it is certain that if oil consumption continues at this pace, prices will continue to rise as oil becomes scarcer. The Hubbert peak oil theory [105] is based on this observation and the fact that extraction roughly follows the discovery curve after a time lag, typically about 35 years for development. Hubberts theory was proven with the US reaching peak oil in the early 1970s and has since been shown to apply to many countries with dwindling oil resources [85].

100 $ 2006 80 $US / Barrel 60 40 20 0 1940 $ Money of the day Oil Consumption
2nd OPEC oil price shock Iraq War, Asian Growth

90 80 Million Barrels / Day 70 60

Gulf War

50 40 30 20 2010

1st OPEC oil price shock

1950

1960

1970

1980

1990

2000

Figure 2.4: Crude oil prices and consumption heavily influenced by world events [194].

The effects of rising oil prices on vehicle fuel consumption in the US are shown in Figure 2.5. In the 1970s when oil prices rapidly increased, both car and truck fuel consumption dramatically decreased due to technology improvements and vehicle size reduction, responding to the Corporate Average Fuel Economy (CAFE) regulation. Although the technology continued to improve engine efficiencies, improvements in vehicle fuel consumption stagnated and even began to increase during the 1980s and 1990s, where US motor vehicle manufacturers resisted further tightening of the CAFE standards. This was caused by factors such as power and vehicle size becoming more important to the consumer. Currently, due to increasing oil prices and concerns over global warming, many new fuel efficient technologies have been introduced and there has been a resurgence in the popularity of smaller vehicles. Furthermore, imminent Californian legislation (Appendix C.2) for CO2 emissions is leading to improvements in fuel consumption.
30

Average miles/gallon

25

Cars Both

20
Trucks

15

10 1970

1975

1980

1985

1990

1995

2000

2005

2010

Figure 2.5: US motor vehicle fuel economy, heavily influenced by oil prices [194, 233].

10

2.3 CO2 Emissions and Global Warming

The increased energy demand and fuel use is accompanied by a similar increase in CO2 emissions. CO2 emissions from human activity, such as burning fossil fuels, contributes a 64% share of the planets greenhouse warming gases as shown in left diagram of Figure 2.6 [104]. 2.7). Furthermore, there is a strong correlation between atmospheric CO2 concentration levels and global temperatures (Figure Over the last 30 years, the average global temperature has increased by approximately 0.5C which has coincided with a 20% increase in CO2 concentrations. Motor vehicles produce approximately 900 million tones of CO2 each year [104], and the transportation sector as a whole produces 26% of the total anthropogenic global output (Figure 2.6 - Right). The Intergovernmental Panel on Climate Change (IPCC) estimates that by 2100, the expected average global temperature rise from 1990 will be between 1.4 - 5.8C. This rise in temperature is predicted to increase the frequency of extreme rainfalls and trigger an increase in sea level due to melting of the polar icecaps.

Methane 19% Nitrous Oxide 6%

Production of Energy 41%

Commercial and Other 6% Residential 8%

CFC-12 6%

Transport 26% Other Halocarbons 5% Carbon Dioxide 64% Manufacturing and Construction 19%

Figure 2.6: (Left): Share of greenhouse warming due to different greenhouse gases [104]. (Right): Share of worldwide CO2 emissions from the combustion of fuel, by sector [106].

11

390 CO2 Concentrations (ppmv) 370 350 330 310 290 270 1880 Global Average Temperature (C) CO2 CO2 (Ice Cores) CO2 CO2 (Mauna Loa) Global Temperatures 14.5 14.3 14.1 13.9 13.7 13.5 1900 1920 1940 1960 1980 2000

Figure 2.7: Global mean temperatures (land and ocean) [165, 168].

When using oil based hydrocarbon (HC) fuels, CO2 output and fuel consumption is primarily a function of engine brake thermal efficiency (TH). Vehicle manufacturers are thus currently pressured by government regulators and consumers to improve TH. From the consumer side, the global drive for more efficient engines is primarily motivated by the recent increasing cost of oil. From an environmental perspective, the growing emphasis on reducing greenhouse gas (GHG) CO2 has forced European institutions and legislative bodies to instigate CO2 targets. More recently, California has instituted GHG emission standards, effective from 2009 (Appendix C.2). These CO2 targets and standards (Figure 2.8) can only be achieved by reducing fuel consumption or by increasing the hydrogen to carbon ratio of the fuel.
Vehicle CO2 Emissions (g/km) (g/km) New Fleet Average CO 2 Emissions

200 180 160 140 120 100 80 60 40 20

Gasoline vehicles All fuels Diesel vehicles Californian GHG Regulation EU Ministry for Traffic (D) Foresight Vehicle (UK) UBA (D)

ACEA

Toyota Hybrid (Prius) Best New Engine of 2004

Hydrogen vehicles (zero g/km) 0 1995 2000 2005

2010

2015

2020

2025

Figure 2.8: Vehicle CO2 emissions. (Black): Fleet target averages over the NEDC [46, 130]. (Blue): Alternative technologies. (Red): Californian regulations for the FTP 75 cycle [61].

12

The California Air Research Board (CARB) has conducted extensive studies into the cost and CO2 emission effects of implementing new technologies to the vehicle fleet [43]. Analysis was conducted for near and long timeframes, with some In the near term, results for the large vehicle class displayed in Figure 2.9.

applying existing technologies such as turbocharging, variable valve timing (VVT), cylinder deactivation, improved transmissions and electric accessories to motor vehicles provides the best option to reducing CO2 emissions. With this current technology, a 22.1% reduction in CO2 is achievable in large vehicles, with a small decrease ($US 58) in annualized whole of life cost [43]. From Figure 2.9, downsized TC engines are shown to be the best option for reducing CO2 in the near and medium terms with little impact on costs, when compared to more complex and expensive technologies such as Homogeneous Charge Compression Ignition (HCCI). Coupling this technology to moderate and advanced hybrid vehicles also provides the potential for further long term reductions. In the long term, it is envisaged that the coupling of high speed diesel engines to hybrid technology will also be able to offer a 50% reduction in CO2 over the 2009 baseline, however, the increased complexity significantly increases cost [43].
5500 4500 Incremental Cost ($) 3500 2500 1500 500 -500 0
DCP, A6

Black = Near Term (2009-2013) Blue = Mid Term (2013-2015) Green = Long Term (2015 - )

Break Even

HSDI, Adv HEV Advanced HEV

DE-ACT, DVVL, DCP, A6, ISG, EPS, EACC

G-HCCI, DVVL, ICP, DCT, ISG, EPS, EACC Moderate HEV

GDI-S, DE-ACT, DCP, DCT, IA

GDI-S, DCP, TC, DCT, EPS, IA

GDI-S, TC, DCP, A6, ISG, EPS, EACC

10

15

20

25

30

35

40

45

50

CO 2 Reduction from 2009 Baseline (%)

Figure 2.9: Large vehicle class technologies and the effects on cost and CO2 emissions [43]. (A6) six speed automatic transmission, (DCP) dual camshaft phasing, (DCT) dual clutch six speed transmission, (DE-ACT) cylinder deactivation, (DVVL) discrete variable valve lift, (EACC) electric accessories, (EPS) electric power steering, (GDI-S) gasoline direct injection at stoichiometric, (G-HCCI) gasoline homogenous charge compression ignition, (HEV) hybrid electric vehicle, (HSDI) high speed diesel, (IA) improved alternator, (ICP) inlet camshaft phasing, (ISG) integrated starter generator, (TC) turbocharged.

13

2.4 Spark Ignition versus Diesel Engines

Due primarily to their lower fuel consumption, diesel engines have the potential to eliminate SI engines in the passenger car market segment. In contrast, SI engines have the potential to do the same if they can be made more fuel efficient (comparable to diesel), whilst maintaining their low urban emissions and cost advantage. The head to head competition between SI and diesel is clear, and this is pushing future engine technology towards a direction that shares useful concepts from both technologies (HCCI). Based on the apparent technological convergence between diesel and SI engines, trends suggest future engines will be much smaller in capacity with the addition of boost and high CRs [130, 202]. More than 50% of all newly registered passenger cars in Europe are turbo-diesels [202]. While turbocharging is widely acknowledged as essential technology to This is largely due to the fact that However, downsized TC SI diesel engines, its role in improving the performance and fuel economy of SI engines has yet to be fully appreciated. turbochargers installed on SI engines have for a long time been associated with motor racing and high performance sports cars. engines installed into passenger vehicles can deliver up to a 20% improvement when designed for fuel economy rather than performance [174] (Section 2.6). Figure 2.10 displays a cost benefit analysis for reducing vehicle CO2 emissions and hence fuel consumption over the New European Drive Cycle (NEDC) with various current SI and diesel technologies [130, 174]. The data shows that a TC four valve high speed diesel engine (HSDI) can reduce CO2 emissions by approximately 22% when compared to the baseline gasoline engine running at stoichiometric conditions. Unfortunately, this benefit comes with a significant increase in cost, with the engine and emissions systems 80% more expensive then current port fuel injection (PFI) technology. Furthermore, recently developed SI engines using direct injection (DI) coupled with lean boosting can match the CO2 levels of the diesel engine, at a unit cost which is 34% less than HSDI technology. Moreover, the largest CO2 to cost benefit over conventional PFI engines comes when considering much simpler technologies, such as applying turbochargers to current PFI engines. This negates the need for complex and expensive high pressure DI systems, relying on conventional emission control methods such as a three-way catalyst (TWC) and exhaust gas recirculation (EGR) to satisfy regulations.

14

40 Reduction in NEDC CO2 (%)

30
NA DI (stratified) TC DI (lean) TC DI (EGR) TC 4V HSDI High Speed Diesel TC DI (stratified)

20
TC PFI (stoich)

10

0 0

NA PFI (stoich)

NA DI (stoich)

TC DI (stoich)

20

40

60

80

100

Increase in Engine and Emission Control Unit Cost (%)

Figure 2.10: Cost benefit analysis for CO2 improvements over the NEDC using various technologies in SI and diesel engines [130, 174].

2.5 Small Engines

Over the last century, vehicle trends have shown increasing power density [232] with the evolution and development of the internal combustion engine. Current motor vehicles continue to have increased power to weight ratios of yesteryear; however they have been achieved with smaller capacity engines which have higher mechanical and thermal efficiencies [232]. Goddard [81] believes that todays trends will continue as concerns escalate over dwindling oil resources and global warming, forcing OEM manufactures to further reduce engine capacities, thus producing small engines for passenger vehicles with swept volumes of less than 1 L. This size of engine is commonly found in motorcycles and other applications. It was noted that the worlds motorcycle Trend analyses in the population exceeded 200 million in 2003 (Figure 2.1).

literature support this concept in larger engines, with a concise review of decreasing capacity through the addition of intake boosting presented in Section 2.6. Hence, applying turbocharger technology to small engines provides an However, a thorough economically viable solution to growing world concerns.

literature review found little work in this area specific to small engines. Further research may provide some insight into future engine trends while exploring the limits of small downsized engines.

15

2.6 Turbocharging and Engine Downsizing

Reducing the engines size (swept capacity and/or cylinder number reduction) and retaining performance by intake boosting is referred to as engine downsizing. Engine downsizing through turbocharging improves the engines TH and reduces the vehicles fuel consumption through a number of mechanisms, including: Reducing engine pumping losses. Less throttling required for a given road power even when the turbocharger is not used (over most of the urban driving cycle) as the smaller engine operates at relatively higher MAP. Hence, the engine operates closer to peak efficiency over the drive cycle. Using the compressor (powered by waste exhaust enthalpy) to force the piston down under boosted conditions to further reduce pumping losses or possibly produce useful work during induction. An improvement in mechanical efficiency (MECH) due to a reduction in friction losses associated with the decreased capacity. Reduced overall heat losses due to exhaust heat recovery and reduced surface areas for heat transfer. Improved mixing and fuel atomization associated with the higher intake temperatures and pressures. Improved in-cylinder turbulence (tumble and swirl patterns) due to higher intake pressures and gas velocities which promote faster burning rates and hence improved combustion. Engine downsizing by applying turbocharger or supercharger technology to maintain equal power has been described across a wide range of commercial and passenger vehicles in a number of references [12, 52, 53, 64, 83, 88, 115, 127, 130, 171, 172, 174, 191, 202, 248, 249]. Petitjean et al. [174] captured these benefits through the data analysis of several hundred family sedan production vehicles, released over a ten-year period. They concluded on average over the past ten-years, that turbocharging had enabled SI gasoline engines to downsize by approximately 30% for the same brake power output. This reduction in capacity had resulted in an average 8-10% fuel economy improvement with the boosted intake also improving engine torque and vehicle acceleration.

16

Petitjean et al. [174] also ran vehicles over the NEDC to capture the benefits of progressive downsizing and turbocharging, with results for three different engine configurations shown in Figure 2.11. In these tests, a 19.6% reduction in fuel consumption was achieved with a 40% downsized engine capacity. These results become even more impressive when one considers that the gains were achieved with a CR reduction of two points to avoid abnormal combustion, as no complex knock preventive strategies were used. Hence, further efficiency gains could be achieved if MAP and/or CR could be increased whilst further downsizing to maintain equal power. TC downsized engines also offer other benefits besides obvious engine efficiency gains. When engine packaging and overall powerplant weight reduction is improved, there are further enhancements in vehicle efficiency and dynamic performance. The reduced engine size also offers mass reductions in engine-out exhaust emissions prior to catalyst light-off, with reduced need for stratified lean burn strategies to improve efficiency. Thus standards for regulated emissions can be met using conventional aftertreatment methods (TWC) at stoichiometric operating conditions.

11 10 9 8 7 6 5 4 3 2 1 0 Baseline 3.0L NA Baseline 3.0L Turbo 2.4L 20% Downsized

10.5 9.0 10.7 9.1

Fuel Consumption (L/100km) Compression Ratio

8.6

8.5

Turbo 1.8L 40% Downsized

Figure 2.11: The effects on CR and fuel consumption over the NEDC for NA and various capacity TC downsized engines with equal power output in the same test vehicle [174].

17

Although downsized engines have been proven capable of matching the performance of larger NA engines with the addition of boost as previously described, some drawbacks do exist. Low speed performance is dictated by the rate at which the turbocharger can supply the required airflow to match the performance of the larger engine. Furthermore, the downsizing extent is limited by the deliverable pressure ratios the turbocharger can achieve. An additional factor is that the increased specific output places greater strain on the internal components of the engine due to the increased combustion loading with intake boosting and possibly increased inertia forces due to engine speed increases. processes. This may increase unit cost and complexity as there is a need to In-service operating costs may also increase. Furthermore, more These redesign internal components using improved materials and manufacturing elaborate control systems are needed to prevent component failure.

measures are required to ensure reliability and durability over the engines life cycle. The higher pressures and temperatures associated with TC engines also increase the occurrence of uncontrolled combustion, mainly knock in the end-gas region, which further deteriorates engine performance and reliability [98, 123].

2.7 Combustion
The combustion process in an internal combustion engine is extremely complex, involving the development and propagation of an unsteady turbulent flame through a mixture whose pressure, temperature and turbulence intensity are continually changing due to the piston motion and the influence of the flame itself. The combustion parameters such as peak pressure and flame propagation time are functions of pressure, flame speed, piston motion, flame geometry, turbulence and heat transfer. The study of flame propagation using these parameters This ensures meaningful results are requires skillful interpretation with the averaging of many cycles used to calculate the burning history for a typical cycle. the definition of each of these as follows. extracted. A combustion event can be defined as either normal or abnormal, with

18

Normal combustion is defined as a combustion process in which: The combustion event is initiated solely by a controlled spark discharge event. The flame front propagates completely across the combustion chamber. The flame propagation is at a relatively uniform speed and repeatable. Abnormal combustion can refer to a variety of situations in which one or more of the above definitions does not hold. This includes processes where: The flame fails to completely consume all of the charge (partial burns and misfires). A hot spot in the combustion chamber causes surface ignition, which can occur prior to the spark initiating combustion or in the unburned gas region. Some or all of the charge is consumed at extremely high rates (spark knock from flame acceleration or auto ignition). Figure 2.12 displays in-cylinder pressure cycle traces highlighting normal (no knock) and abnormal combustion (spark knock) taken from the test engine.

12 Abnormal Combustion 10 Raw Pressure (MPa) 8 6 4 2 0 -180 Normal Combustion Ignition

-140

-100

-60

-20

20

60

100

140

180

220

Crank Angle (deg.)

Figure 2.12: Normal and abnormal (heavy spark knock) combustion in the test engine. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11, 12 BTDC spark timing, peak knock amplitude = 3 MPa.

19

2.7.1 Abnormal Combustion

This section briefly outlines the phenomena of abnormal combustion, with more detailed descriptions found in the literature [95, 98, 211, 229]. Spark knock is caused by the auto ignition of the mixture before the flame front consumes the charge. Auto ignition is caused by excessively high local temperature and pressure in the end-gas. Auto ignition before the burning flame consumes the charge, accelerates the flame front causing rapid heat release, which results in pressure oscillations within the cylinder (Figure 2.12) as the heat can not be quickly dissipated. The unsteady pressure waves (reflected at the cylinder walls) initiated by the rapid heat release break down the quench layer, exposing the combustion metal surfaces to possible damage. Spark knock or auto ignition is not to be confused with surface ignition. Surface ignition occurs before the spark or shortly afterwards in the slowest burn region as opposed to knock, which occurs late in the cycle. Surface ignition is a result of hot spots within the combustion chamber igniting the mixture. Hot spots are usually caused by a previous knocking cycle or poor combustion chamber design (geometry and cooling). Pre-ignition from the spark plug electrode is a common form of surface ignition. As shown in Figure 2.12, heavy knock is characterized by a rapid rate of pressure rise which can be up to 20 times higher than normal combustion, however less than the rise rate for a detonation wave [92]. The propagation of the ensuing pressure shock waves are at velocities in the order of 1,000 m/s. This compares with 15 m/s for turbulent flame speeds [120, 229]. Heavy knock must be avoided in engines as it leads to audible engine noise which increases the vehicles noise, vibration and harshness (NVH). damage can occur due to the high thermal stresses. Furthermore, depending on the knock location and severity, component structural or surface Common damaged components due to abnormal combustion include cylinder head gaskets, piston and cylinder bore assemblies, bearings and spark plugs, which can all lead to significant and costly engine damage [16, 98, 140, 229]. Furthermore, knock also reduces TH due to an increase in heat transfer through the combustion walls. For

20

knocking pressure oscillations greater than 0.5 MPa, Lu et al. showed that the heat flux increases almost linearly with the knock amplitude [140]. Hence, the tendency to knock is highly dependent on the engine design and operating variables which influence the end-gas temperature, pressure and the time spent at high values of these two properties before flame arrival. Thus the tendency to knock is decreased through reductions in end-gas temperature, with further detail given in Section 2.7.2.

2.7.2 Knock Reduction Methods

An increase in the fuels octane number is one method used to reduce the engines knock propensity. Hence, the quality of fuel is commonly used to characterize the knock reduction methods effectiveness. The effects of different parameters on knock reduction can be based on the extent to which they are capable of lowering the octane number requirement (ONR) of the engine. Table 2.1 shows the effects of various parameters on the ONR of an engine. In the literature, most of the results published on ONR are based on experimental results [90, 154, 169, 195, 217, 220, 226, 264]. The range tested from which these results were developed is also presented in Table 2.1. It is worth noting that most parameters exhibit a strong linear relationship within the range tested, however extrapolating ONR outside of this range should be performed with caution.

2.8 Summary
This chapter provided a brief background on the increasing global vehicle population together with growing fuel consumption and global warming concerns. Various technologies are examined on a cost versus fuel consumption basis, with turbocharging and engine downsizing shown to be a viable near term solution. The possibility of turbocharging small engines for downsizing purposes is also discussed, enabling the engine design (Chapter 3) to accommodate the high loads associated with this type of application.

21

Table 2.1: The effect of various parameters on the ONR of a SI engine [90, 154, 169, 195, 217, 220, 226, 264]. Parameters Spark Advance Octane Number Requirement (ONR) increase 1 ONR / 1 knock limited spark advance increase 1 ONR / 7C peaks around 5% rich of stoichiometric, decreases 2 ONR / 0.1 increase 3 - 4 ONR / 10 kPa Range Tested 0 - 30 CA Reference [90, 195]

Intake Air Temperature

20 - 90C

[90, 195] [154, 195, 226] [195, 226] [90, 154, 195, 217, 264] [195]

Air-Fuel Ratio (AFR) Manifold Absolute Pressure Compression Ratio

12 - 26 AFR

85 - 135 kPa

increase 5 ONR/ CR

5 - 12 CR

Exhaust Back Pressure

increase 1 ONR / 30 kPa

0 - 65 kPa

Coolant Temperature

increase 1 ONR / 10C decrease 1.4 ONR/300 m decrease 2.5 ONR/300 m decrease 1 ONR when increasing relative humidity from 40% to 50% at 30C increase 6 - 9 ONR over life of engine increase up to 12 ONR depending on driving style decrease 1 ONR / 1% H2 added decrease 4 ONR when DI used over PFI decrease up to 5 ONR as squish area increases decrease up to 15 ONR from cylindrical to modern type (hemispherical head)

70 - 110C 0 - 1,800 m 1,800 - 3,600 m -

[195]

Altitude

[90]

Humidity

[90]

Engine Deposits Excessive Oil Consumption Hydrogen (H2) Addition

0 - 25,000 km

[90]

0 - 12% H2 added 0 - 67% squish area 7.8 - 11 CR

[90]

[226]

Type of Fuel Injection

[169]

Increasing Squish Combustion Chamber Shape

[220]

[217]

22

CHAPTER 3
P O W E R E D B Y W A T T A R D

Engine Design

3.1 Introduction
In this chapter, the engineering design of a highly TC inline two cylinder small engine intended for use in FSAE competition is presented. Due to the intricacy and complexity of the system, it is difficult to discuss all aspects of the engine design. Rather, this chapter is focused on the most salient points, with emphasis placed on key features and novel aspects which were vital in ensuring robustness and reliability for the highly TC package. If further detail is required, the reader is referred to Appendix P, which displays detailed assembly and component drawings of the engine design. This chapter begins with an outline of the theoretical analysis used to set the capacity and configuration for this rules-compliant FSAE engine. In addition, computer aided design (CAD) and validated computational fluid dynamic (CFD) models were used to substantiate expected performance gains. The use of a few OEM components is also outlined as they reduced costs, lead times and the complexity of the project. Additionally, the detailed design and analysis of some major components are described. The design process for the new components incorporated proven techniques involving computer aided engineering (CAE) analysis, empirical data and detailed calculations. A special feature is the detailed description of a novel gasketless interface created for an open deck cylinder block design. Lastly, final engine specifications are listed together with a description of the assembly of the test engine.

23

3.2 Design Brief

A clean sheet approach was taken in designing this engine specifically for SAEs student formula race-car competition against the specific FSAE project objectives outlined in Chapter 1.1.1. the Formula. The design brief for the engine featured reduced capacity when compared to 0.6 L maximum capacity engines commonly used in These maximum capacity engines were typically OEM high performance motorcycle engines in four cylinder configuration (Appendix B). Downsizing had obvious packaging advantages including large reductions in mass, physical size and centre of gravity (COG) height, which all contributed to improving the dynamic performance of the Formula vehicle [13, 16, 151]. The specifically designed downsized engine also allowed the positioning of many major components, including manifolds and auxiliaries, to suit FSAE applications. Furthermore, as described in the theoretical analysis (Section 3.3), the overall engine performance curve (including peak power) could also be improved if the mandatory 20 mm diameter intake restriction could be choked at low engine speeds, thus extending the maximum power speed range.

3.3 Theoretical Analysis

From the beginning, based on theoretical analysis, the design was focused on accommodating the mandatory intake flow restriction (Appendix D) which limits peak power [13-15]. Consequently, the capacity and configuration were influenced by air consumption limitations.

3.3.1 Engine Capacity and Configuration

The choice of engine capacity was largely dictated by the FSAE performance objectives of the project as detailed in Appendix B.6. These aims included achieving near constant power over half the speed range together with the opportunity to increase peak power when compared to typical 0.6 L four cylinder reference engines. achieved. With appropriate design decisions, these aims could be

24

With peak brake power limited to ~60 kW due to the limited air consumption (Appendix D.5), peak performance gains at the choked condition were only expected to be minimal and due to a combination of thermal efficiency improvements together with friction loss reductions. The strategy to reduce friction losses involved limiting the maximum engine speed to 10,000 rev/min as losses increased at the square of the speed increase [259, 261]. Furthermore, if the swept capacity could be reduced whilst keeping the rotating and reciprocating components as small as possible and still maintaining choked operation, delivered power would also increase due to the reduction in frictional losses associated with the smaller capacity [13, 15, 16, 113]. The use of a suction device downstream of the restrictor (Figure D.1) allowed the maximum mass flow through the restrictor to be maintained over a wide speed range through delivering air at regulated boost, which would enable the above mentioned constant power performance aims to be achieved. Turbocharging was the preferred method of intake boosting over mechanically supercharging due to the documented TH improvements [42, 123, 174, 191, 248] together with the high pressure ratios that were achievable. flow operating conditions. Furthermore, turbocharger boost limitations dictated the volume of the swept capacity while still maintaining choked Hence, the proposed engine design concept is comparable to downsized engines found in large automobile diesel and gasoline applications but on a reduced scale, with intake boosting used to compensate for the swept capacity reduction and additional mass benefits associated with the smaller engine. Whilst the requirement to achieve these results with sufficient suction on the compressor intake (needed to achieve sonic flow in the restrictor) would appear to be an excessive condition for a non-racing application, in reality it dictates better than usual turbocharger specification and performance. To determine the swept capacity, a compressible flow model was created to calculate intake air consumption based on the maximum mass flow that could be achieved in the choked restrictor condition (Appendix D.3). Hence, the engine capacity was selected with the aid of Figure 3.1, which shows the predicted volumetric efficiency (VOL) needed to maintain sonic flow through the intake restriction for varying engine capacities and operating conditions.

25

400

300

Turbocharger limited maximum VOL

VOL (%)

200

100

0 0

0.20 0.25 0.30 0.35 0.40 0.50 0.60

2000

L L L L L L L
4000 6000

Experimentally observed choked flow (Suzuki GSX-R600 FSAE trim)

8000

10000

12000

14000

16000

Engine Speed (rev/m in)

Figure 3.1: Predicted engine air consumption needed to maintain choked flow operating conditions for varying swept capacities and operating conditions with model validation from previous experimental results.

A validation point for the model simulation is also shown in Figure 3.1, with a dynamometer experimental result from the in-house developed Suzuki GSX-R600 engine in FSAE trim. A slight VOL over-prediction is seen from the model, attributed to flow loss effects on the dynamometer associated with vena contracta [58] and pulsed flow through the intake restriction [14, 245]. On the basis that a pressure ratio of ~3 (~250% VOL assuming losses) is the expected turbocharger limit (Appendix E.2.2), an engine size between 0.4 and 0.45 L was selected with an operating speed ranging from 6,000 to 10,000 rev/min. The engine configuration was heavily influenced by the turbocharger and engine packaging objectives. A single cylinder provided mass and packaging benefits but posed large problems with turbocharging and cylinder filling. Two or a larger number of cylinders reduces the flow velocity fluctuations experienced by the exhaust system and turbine but adds mass and complexity together with increased frictional losses due to the greater piston rubbing areas. The chosen configuration featured a two cylinder in-line configuration as a compromise between the competing effects, but with an uneven firing interval (0, 180 crank angle (CA) ) which was selected for engine balance reasons.

26

Cylinder bore and stroke dimensions were then determined, with the stroke length dictated by the maximum speed limitation of 10,000 rev/min. This equated to a stroke length of ~58-60 mm with the maximum mean piston speed (MPS) limit set at 20 m/s to ensure engine reliability [220]. Hence to maintain the required swept capacity, a bore dimension of ~68-70 mm was required resulting in a bore stroke ratio of ~1.2. This ratio had previously been proven to promote adequate incylinder mixing and consequently good combustion in four valve designs [214, 215, 220]. Maintaining a large bore size relative to other small engines (< 60 mm) also reduced the heat losses because of the lower surface to volume ratio, which potentially improves TH. However, the smaller bore size when compared to usual passenger vehicle applications reduced the knock propensity due to the shorter flame travel distance, which helped improve TC engine reliability. Furthermore, potential performance improvements were also possible through MAP and/or CR increases associated with the reduced knock likelihood.

3.3.2 CFD Simulation

In order to enumerate the expected performance gains suggested in Section 3.3.1 and access the viability of a small capacity, downsized, two cylinder package for FSAE, a 1-dimensional CFD model was created in conjunction with Karagounis [113]. The CFD code employed was Ricardos WAVE [185], with further detail and description behind the software including modeling concepts and equations listed in Appendix F. To validate the application of the WAVE software for a known set of model parameters, including the effects of flow restriction, modeling was completed for the Suzuki GSX-R600 [113] in FSAE trim, for which extensive experimental data was available for comparison. Figure 3.2 displays close correlation between the modeled Suzuki and experimental results. Comparisons validate the authors ability to configure the inlet and exhaust node sequence and conditions together with entering flow and combustion parameters. With this experience, a WAVE model was constructed for the proposed two cylinder design with model block diagrams for both engines shown in Figure 3.3.

27

60 50 Brake Power (kW) 40 30 20 10 0 4000 New Design (Simulated) Suzuki GSX-R600 FSAE Trim (Simulated) Suzuki GSX-R600 FSAE Trim (Experiment)

5000

6000

7000

8000

9000

10000

11000

12000

Engine Speed (rev/min)

Figure 3.2: Simulation results for the proposed two cylinder TC design and the Melbourne University FSAE teams previous engine (NA Suzuki GSX-R600), with Suzuki experimental results used for model validation. Both engines flow restricted.

Initial model performance results were poor due to the odd firing (uneven) inlet and exhaust flows and the turbocharger interaction, as described in Chapter 6. As the engine capacity and configuration including the bore and stroke dimensions were set from Section 3.3.1, no simulated engine optimization was completed in these areas. Rather, to mitigate the odd firing turbocharger interactions, the simulation model was used to improve performance by optimizing the intake and exhaust manifold geometry together with camshaft profiles (Chapter 6.5). This was required due to a shortage of published data and design rules on turbocharging odd fire, two cylinder configurations. A performance comparison between both the Suzuki and the proposed new engine design is also displayed in Figure 3.2. It is noted that the Suzuki had undergone extensive experimental dynamometer and simulation development [93, 113, 187], with best achieved results displayed for fair comparisons. Results highlight the smaller engines slight improvement in peak power due to thermal efficiency improvements and friction loss reductions previously suggested. However, more evident is the engines ability to maintain the choked flow operating condition over a wide speed range (6,000 to 11,000 rev/min), resulting in a near flat power curve and hence large improvements over the Suzuki baseline. of the FSAE project performance objectives. These findings confirmed the previously suggested performance expectations and the feasibility

28

Suzuki GSX-R600 FSAE Trim

Proposed Downsized Concept

Legtend
- Fuel Injector - Restrictor - Pipe - Pipe Junction - Infinite Volume - Pipe Junction - Cylinder - Compressor - Turbine

Figure 3.3: Ricardo WAVE model block diagrams used to predict FSAE engine performance for the flow restricted condition. (Upper): NA Suzuki GSX-R600 model used for initial software validation. (Lower): Proposed highly TC engine model created to explore the expected performance gains from engine downsizing.

29

3.4 Achieving Balance in the Design Process

The design approach centered around overcoming the challenges associated with highly TC engines together with cost and time limitations. To achieve project objectives, compromises were required between three primary factors including engine mass, packaging and reliability as well as cost and time. The philosophical steps in the design were as follows: 1) Set parameters 2) Undertake rough theory 3) Detail theory 4) Size components 5) Analyze and optimize 6) Check and conclude

3.4.1 The Feasibility of OEM Engine Adaption

The possibility of adapting an off-the-shelf inline two cylinder engine which met the criteria set in the theoretical analysis was initially explored. However, this concept was quickly dismissed as no OEM production engine was capable of withstanding the expected 25 bar BMEP arising due to the ~300 kPa MAP that was needed to choke the flow restrictor at mid-range speeds. Furthermore, available production engines were generally motorcycle based and hence already near their performance limit, having high CRs and engine speeds. Consequently, it was unlikely that these engines would be over-designed to the extent that reliability could be maintained at 25 bar BMEP, with the maximum BMEP of these NA engines already limited to the high value of 15 bar [221]. The feasibility of strengthening the internal components of a motorcycle based engine was also explored, but dismissed as no modern lightweight units which fulfilled the requirements set from the theoretical analysis (Section 3.3) existed in the marketplace. This background work effectively meant that a new engine needed to be constructed using suitable OEM components where possible, with particular focus placed on the rotating and reciprocating assembly (Section 3.5) due to the high specific output associated with highly TC applications.

30

3.4.2 OEM Component Adaption

Although a new engine needed to be constructed, it was highly desirable to use as many off-the-shelf OEM components as possible. This ensured components could be quickly sourced and easily replaced during development, while keeping costs and lead times to a minimum. Furthermore, the likelihood of component failure was also reduced as the design and manufacturing elements had been proven up to known values. If OEM components could not be sourced or failed to meet the new operating requirements, new components were designed and manufactured with techniques described in Section 3.4.3. Major OEM based components are now briefly described including the crankcase, cylinder head and transmission, which all required varying levels of modification, with further information supplied in the detailed drawings of Appendix P. Crankcase Previous work by Milliken explored engine durability aspects of TC and NA racing engines [150]. In this work, Milliken found that there was a reliability tradeoff between engine displacement, speed and charge density. As the charge density was anticipated to be high due to turbocharging, a larger capacity, inline two cylinder crankcase with a larger bore size than that required from the theoretical analysis was adapted to improve engine reliability. The bore diameter was then reduced to compensate for the increased combustion pressures resulting in similar crankcase loads. Peak cylinder pressure was used as the measure to identify combustion loads, with turbocharging anticipated to produce up to 40% higher peak cylinder pressures than NA operation during normal combustion at high boost levels [92, 171]. This value was later confirmed as a good estimate when compared to experimental data (Figure 9.11). To facilitate the design process, an existing older style two cylinder crankcase which had many benefits for the highly TC package due to its sturdy construction was adapted. above larger A Kawasaki ER500 crankcase was finally selected, fulfilling the displacement requirements while still available in todays

marketplace, enabling components to be sourced locally. Although the crankcase was well suited to the application, all other components were replaced to enable

31

high BMEP operation. In addition, further crankcase modifications were required for the lubrication system as well as the rotating and reciprocating assemblies to suit the newly designed components. Transmission An integral part of the engine design involved the transmission as this was contained within the selected crankcase. The transmission also allowed the wide speed range engine power characteristic to be matched to FSAE application. Hence the criteria for selecting the required number of gear sets and subsequent ratios were largely dictated by vehicle speeds and engine performance. The major event in FSAE is the Autocross, which has an average speed of around 55 km/h which typically involves a speed range varying from 30 to 90 km/h. In vehicle acceleration over this speed range, typically two or more gear up shifts and consequential downshifts are required. It is noted that automatic gear Even if gearshifts are made shifting in motorsport major competition typically takes about 30 ms, so in one lap up to 0.5 seconds may be lost to gear shifting. clutchless, with the drive train absorbing the shocks, the torque pulsations unsettle the vehicle stability, particularly in turns. Hence, reducing the number of gearshifts has advantages in increasing the power delivered, which can be used by the skilled driver [13]. Furthermore, it follows that an engine with constant power over the speed range has a potential advantage in the autocross event by eliminating gear shifting and running in only one gear. limitations. Consequently, extensive matching analysis was completed to determine the required number of gear sets and ratios to stay at nearly constant power during vehicle operation, with further detail outlined in Appendix G. A three speed The existing transmission was selected, with gear ratios shown in Table G.1. However, in the acceleration event, one gear change may be required depending on traction

transmission within the crankcase was then extensively modified to suit the new application. This involved keeping the existing layout but heavily modifying the 6 speed original design. Hence, the gears and change mechanism were altered with new gear sets to give a wide ratio 3 speed transmission, compatible with the constant power concept of the engine.

32

Cylinder Head To enable high BMEP output, a modern four valve, pent roof combustion chamber cylinder head was required, due to well accepted advantages in both air consumption and combustion. Extensive airflow testing was carried out on a steady flow rig (Appendix H) in the search for a suitable cylinder head that fulfilled the required airflow, CR and bore spacing requirements (set by the crankcase). As no suitable inline two cylinder heads could be sourced, the feasibility of using a cylinder head from a four cylinder configuration was initially explored with flow tests (Appendix H.5). A four cylinder Suzuki RF600 cylinder head was selected due to excellent flow results, while also fulfilling the packaging, bore spacing and timing chain requirements. From airflow experiments, the Suzuki cylinder head gave an inlet mean flow coefficient of 0.4 (Figure H.5.3) with a 0.53 critical Mach index value. This critical Mach index was favorable as it did not exceed the 0.6 maximum value [220], allowing high port velocities which improved in-cylinder tumble and swirl without restricting air consumption. As the Suzuki cylinder head was designed for a slightly reduced bore diameter, the combustion chamber exhibited increased squish areas around the periphery when adapted to the larger bore of the inline two cylinder. This improved in-cylinder turbulence in the end-gas region and hence was beneficial in reducing knock propensity [229], a well documented and expected problem in the highly boosted design. Furthermore, the compactness of the small combustion chamber reduced the volume contained in the cylinder head in comparison to the cylinder volume, allowing relatively high CRs (~13) to be achieved with a simple flat piston crown. The desired CR could eventually be matched to suit the operating conditions with piston crown modifications which reduced the CR (Section 3.5.3). A heavy downside of the Suzuki cylinder head was the extensive modifications that were required to adapt it to the two cylinder configuration. This involved removing cylinders 1 and 4 which exposed the oil and water passages which created sealing challenges. Other retained cylinder head components such as the valve train also created problems, requiring the modification and manufacture of many components, as displayed in the drawings of Appendix P.

33

3.4.3 New Components

To integrate the aforementioned OEM based engine components, a large number of newly designed parts were required. This included the rotating and reciprocating assembly (Section 3.5), as maintaining mechanical integrity during engine operating at high boost levels and hence high specific output was critical to the success of the program. For the purposes of this engine program, it was important not too push the limits in mechanical design too far as reliability was paramount. Well proven empirical design rules and techniques (based on previous experimental data), safety factors and extensive CAE analysis were incorporated in the design process. Ease of assembly and maintenance were also factored into the design to minimize complexity and engine down time during engine development (Chapter 5). Design for manufacture was also an important consideration during the design evolution process due to cost and time limitations. the majority of required components. times involved with cast, forged or Component design where possible focused on utilizing simple machining and fabricating techniques to manufacture These manufacturing methods were extruded prototype components. preferred as the author was proficient in these areas, which avoided the long lead Manufacturing methods were selected to meet individual component loading requirements with complex casting and computer numerical control (CNC) machining only used when there were outstanding advantages. For highly stressed components, especially in the rotating and reciprocating assembly, machining from billet or forged materials was critical to ensure material consistency. Casting was generally avoided due to the limited numbers of parts required and the complexity of the process, which requires moulds and patterns. However, this manufacturing technique was adopted for all covers which required only limited mechanical integrity but allowed the use of thin walled (~3 mm) magnesium alloy to reduce mass. The rocker cover pattern which was fabricated from mild steel and the resulting magnesium alloy casting is shown in Figure 3.4.

34

Figure 3.4: Rocker cover patterns and magnesium alloy casting.

3.5 Rotating and Reciprocating Assembly

The design of the rotating and reciprocating assembly centered on maintaining component reliability under the harsh operating conditions associated with highly boosted, high speed engines. design (Figure 3.5). The design initiated from a load analysis as described, with components highlighted in the sectional view of the final engine

3.5.1 Load Analysis

A load analysis was completed using an analytical model which incorporated both inertia and pressure loading throughout crankshaft rotation. This enabled worst case loading conditions to be established, which set upper design limits on conditions individual components had to withstand. A safety factor was then applied to the worst loading condition to compensate for other factors which could not be accurately quantified, such as abnormal combustion or component misalignment. To enable the design of the reciprocating assembly, load analysis was performed on an individual cylinder basis as the crankshaft had four main bearings (two per cylinder), which was governed by the adopted crankcase. crankshaft (Section 3.5.4). Individual cylinder loads were later fed into the rotating assembly to facilitate the design of the

35

Rocker Cover

Exhaust Camshaft

Intake Camshaft

Coil on Plug

Cylinder Head

Piston

Spark Plug

Piston Rings

Gudgeon Pin

Piston Oil Cooling Jet

Dry Liner + Barrel (Cylinder Block)

Crankshaft

Connecting Rod

SCALE 1:2 Figure 3.5: Sectional view of the final engine design highlighting the rotating and reciprocating components.

36

Initially due to component uncertainty and lack of information, model inputs such as masses were estimated, however were based on empirical data gathered from the literature [8, 220]. As the design and specification of individual components evolved during the design iteration process, the analytical model was updated to determine new loading conditions. This was imperative as all components in the assembly were designed concurrently and hence were affected by one another, with each of the changes contributing to the worst case loading conditions. To determine the gas pressure forces acting on the assembly in the load analysis, in-cylinder pressure data obtained from previous experimental work was utilized to replicate the pressure forces [92, 248], also corresponding to pressure data modeled by WAVE. A peak pressure near 7.5 MPa was expected at the highest boosted conditions under normal combustion. This value was also comparable to previous findings from a dual fuel natural gas diesel platform operating at a CR of 16.5 with a very high pressure rate rise of 400 kPa/CA [266]. A safety factor was then applied to compensate for other factors (abnormal combustion) as previously described. Additionally, the pressure curve was varied in the load analysis model depending on the boost level and hence BMEP. This was required as the maximum boost conditions and hence combustion pressure loading decreased as the engine speed increased due to the choked intake restrictor. Analytical model inertia forces were calculated by modeling the vertical reciprocating motion throughout crankshaft rotation [211], which enabled acceleration rates to be determined from engine speed. Reciprocating components included the piston, gudgeon pin, piston rings and a portion of connecting rod (~1/3 total connecting rod mass, determined from the centre of mass position). Hence forces were then calculated from component masses and continuously updated during the design evolution process due to changing parameters. As the piston was designed to accommodate anticipated CR changes through piston crown modification (Section 3.5.3), the maximum mass corresponding to the highest CR was used in calculations. Figure 3.6 displays the final design loads calculated from the analytical model at the end of the design iteration process. Inertia, gas pressure and combined resultant loads are shown for both worst case operating conditions involving

37

maximum engine speed and maximum BMEP. The upper diagram of Figure 3.6 highlights that the combustion loads are the dominant factor at peak BMEP conditions (only minor inertia force cancellations), with peak loading occurring near peak cylinder pressure (~10 ATDC) as previously documented [155]. At maximum engine speed (Figure 3.6 - Lower), inertia forces dominate with peak loading at top dead centre (TDC) during valve overlap. Both worse case inertia and gas pressure conditions highlight peak resultant loads of ~ 20 kN.

20 15 10 5 Force (kN) 0 -5 -10 -15 -20 -25 -30 0

At maximum BMEP 6,000 rev/min 7.5 MPa peak pressure ~300 kPa MAP

Inertia Force Gas Pressure Force Resultant Axial Force

Compressive Combustion dominated

120 240 360 Crank Angle (deg) 480 600 720

20 15 10 5 Force (kN) 0 -5 -10 -15 -20 -25 -30 0

At maximum speed 10,000 rev/min 6.5 MPa peak pressure ~180 kPa MAP

Inertia Force Gas Pressure Force Resultant Axial Force

Compressive

Inertia dominated

120

240

360 Crank Angle (deg)

480

600

720

Figure 3.6: Crankshaft crankpin loads caused by inertia (reciprocating component movement) and gas pressure (combustion) forces at maximum operating conditions. (Upper): Combustion dominated force at maximum BMEP. (Lower): Inertia dominated force at maximum speed.

38

3.5.2 Connecting Rod

Connecting rod specifications, including the big-end and small-end dimensions, were governed by piston (Section 3.5.3) and crankshaft (Section 3.5.4) designs. A connecting rod length of 116 mm was chosen following design rules which recommended maintaining a minimum connecting rod length to stroke ratio of 2 [77, 219, 220]. This was imperative in preventing high piston jerk loading during mid stroke which placed high resultant thrust loads on the piston skirt. Extensive CAE analysis was utilized during the connecting rod design. Due to the high loads and targeted minimum mass, a high strength 4140 alloy steel [146] with a yield strength in excess of 1,000 MPa was chosen for the design. A safety factor of 2 was also applied to the worst case load calculated from the analytical model (Section 3.5.1) to ensure component integrity. CNC machining from billet material was utilized to manufacture the connecting rods to ensure component integrity as very few were required. Furthermore, this allowed relatively cheap components to be manufactured by specialty manufacturers to the required design. Component fatigue in the connecting rod bolts and the connecting rod itself has previously been documented to be a potential weakness in connecting rod design [3, 150, 216, 220]. Hence, the endurance limit of the alloy steel material (~0.5 of the tensile strength [204]) was not exceeded in the design, with post machining detailing and shot-peening [159] used to reduce the likelihood of crack propagation. Furthermore, very high quality connecting rod bolts (1,200 MPa in order to ensure adequate ARP2000 [22]) were utilized and not oversized

material remained around the connecting rod big-end [220]. Both H and I section connecting rod designs were evaluated with extensive finite element method (FEM) analysis. Figure 3.7 displays model stress analysis The results for compressive and tensile loading conditions for both designs.

connecting rod was analyzed as an assembly with gudgeon pin and crankshaft big-end component representation. This ensured the validity of the FEM analysis by reducing the deflections of the pin and big-end and the effects of strain in the connecting rod. Component surface contact and boundary conditions were evaluated on a node to node element basis.

39

H - Beam
Little-end

I - Beam

A A

B B

Big-end

Y-plane

X-plane

Gudgeon pin axis

Section A-A

Section B-B

H - Beam

I - Beam
MPa 1000

H - Beam

I - Beam

750

500

250

Tensile Loading (40 kN)

Compressive Loading (40 kN)

Figure 3.7: Evaluation of H and I section beam connecting rods. (Upper): Equal mass CAD models with beam cross section detail. (Lower): FEM model analysis comparing von Mises stress distribution with x 100 distorted profile shown.

40

H - Beam

I - Beam
MPa 600

H - Beam

I - Beam

450

300

150

Bending Load (5 kN) X-plane

Bending Load (2 kN) Y-plane

Figure 3.8: FEM model analysis comparing von Mises stress distribution for both connecting rod designs under bending loads (misalignment or abnormal pressure forces) with x 10 distorted profile shown.

Y-plane
End-gas

IN
X-plane (Pin axis)

End-gas

EX

Areas of highest loading due to abnormal combustion

IN

IN

EX
End-gas

EX

Two valve hemispherical chamber (favouring H - beam design)

Four valve pent roof chamber (favouring I - beam design)

Figure 3.9: Various combustion chamber geometries highlighting highest loading locations due to abnormal pressure forces (spark knock) thus favoring various connecting rod designs to minimize deflection and ensure bearing reliability.

41

Figure 3.10: (Left): I section connecting rods during the manufacturing process. (Right): Manufactured items to specified designs and drawings.

Results from the left lower diagram of Figure 3.7 show similar tensile loading stresses for both equal mass H and I section designs due to the similar beam cross sectional area. However, at the same mass, the I section rod had lower compressive stresses midway up the beam at peak load as can be seen in the lower right diagram of Figure 3.7. In theory, the connecting rod is purely a member under compressive or tensile loading, with the piston skirt absorbing the jerk load (thrust force) due to the gudgeon pins rotation about the little-end axis. In practice, an element of connecting rod side loading due to misalignment, whipping and abnormal pressure forces can exist and create bending or twist loads. These bending effects were analyzed, with Figure 3.8 displaying FEM analysis results. The left images of Figure 3.8 highlight that the I section connecting rod is advantageous in resisting bending in the plane of angular rotation (X-plane) associated with whipping. This has benefits in preventing connecting rod fracture in the case of big-end bearing failure due to subsequent high bending loads. Furthermore, the superior stiffness of the I section design in the X-plane (Figure 3.8 - Left) was beneficial under abnormal pressure forces (end-gas spark knock) due to their typical location, in four valve, pent roof designs. However, the H section was beneficial in Y-plane bending (Figure 3.8 - Right), which was

42

advantageous under unsymmetrical loading caused by the piston axis being out of line with the connecting rod axis. Furthermore, the abnormal pressure forces due to the subsequent knock location in two valve designs make the H design preferable for these configurations (Figure 3.9). Consequently, the I section design was adopted due to compressive load benefits and improved suitability for the modern pent roof chamber amongst other factors highlighted, with manufactured items displayed in Figure 3.10. Previous high BMEP applications [66, 77] had exposed the little-end bearing in the connecting rod as a potential weakness, due to problems associated with inadequate gudgeon pin cooling and lubrication. This had resulted in gudgeon pin discoloration due to high temperatures which caused component fretting and galling. Hence, a more forgiving, thin walled bronze bush was inserted into the connecting rod little-end to avoid steel to steel contact and thus improve surface contact lubricity. Furthermore, extensive FEM analysis was completed on the little-end oil hole location, as the conventional single central oil hole was exposed as a potential weakness with minor stress concentrations (Figure 3.11). Hence, the single oil hole was replaced by two smaller oil holes with the same orifice area, as depicted in Figure 3.11.
Stress concentration

MPa 500 400 300 200 100 0

Single Oil Hole

Dual Oil Hole

Figure 3.11: FEM model analysis comparing von Mises stress distribution for various connecting rod little-end oil hole locations (30 kN tensile load). (Left): Single central hole in high loaded region and resultant stress concentration. (Right): Two smaller holes.

43

3.5.3 Piston Assembly

Extensive stress and thermal FEM analysis was utilized during the design of the reciprocating piston assembly, which consisted of the piston, gudgeon pin and piston rings. The mechanical integrity of the assembly was crucial to the success of the project due to the high BMEP output associated with ~300 kPa MAP operation. In previous boosted high output applications, forged aluminum pistons had proved superior and hence more reliable over cast versions [66, 219, 220, 248]. Consequently, the piston design utilized a forged 2618A aluminum alloy [146] blank (Figure 3.12 - Left), as the mechanical and thermal properties of conventional cast aluminum versions from OEM production engines were unsuitable. Well documented piston problems included top land failures due to high combustion temperatures and knock intensities. Furthermore, manufacturing custom pistons for the specific high output application allowed the piston design to focus on maintaining reliability without compromising the mechanical integrity of other components in the rotating and reciprocating assembly. The solid skirt piston design relied heavily on the FEM analysis, however the design process adhered to recommended piston design ratios found in the literature [175, 219, 220]. Furthermore, the design utilized TC piston data from the Lancer EVO-7 WRC supplied by Mitsubishi Australia [152]. All this information aided in determining the location and sizing of key elements such as the gudgeon pin diameter and placement, skirt profile and depth, piston ring locations and piston crown thickness, as further detailed in the drawings of Appendix P.

Figure 3.12: The design and manufacture of forged aluminum custom pistons were required due to the high boost levels employed. (Left): Off-the-shelf forged piston blank prior to machining. (Right): Manufactured piston design with single compression ring.

44

As the piston underside geometry could not be varied due to the blank forging (Figure 3.12 - Left), the design targeted weight reduction by minimizing the required piston crown thickness. Provisions for crown thickness variations were also required as the CR and combustion chamber shape necessitated experimental dynamometer development (Chapter 4.6). Hence, pistons were designed and manufactured with a flat top crown (Figure 3.12 - Right), with simple machining processes later allowing the CR to be reduced from 13 to 9 (Chapter 5.4.3), which resulted in a minimum crown thickness of 4 mm. To ensure crown reliability under high mechanical and thermal loads, oil cooling of the piston underside was utilized as this had been found to be required in previous high BMEP applications [152, 175, 220]. Hence, a single oil jet was positioned to direct oil flow towards the hotter exhaust side (Chapter 5.4.4). A fully floating gudgeon pin design, retained in the piston with spring wire snap clips was utilized. The pin was manufactured from an EN36A high tensile steel alloy [146], which was only case hardened to minimize wear while still maintaining some elasticity under the high loads to prevent brittle failure. The gudgeon pins were ground to individually suit each of the custom pistons supplied. This ensured adequate clearances were maintained to prevent surface fretting and galling between the dissimilar materials, previously documented to be a problem in high BMEP applications [66, 77]. Extensive FEM analysis was completed to determine the gudgeon pin diameter and geometry, as depicted in Figure 3.13. Analysis was completed on the basis of beam theory under maximum gas pressure load [220], with a safety factor of 2 incorporated to allow for abnormal combustion and misalignment. A 20 mm diameter pin was selected, which was oversized when compared to typical ratios (pin diameter / cylinder bore diameter [220]) found in modern performance engines [214, 215], but was required to stiffen the assembly due to the increased BMEP. To combat the increased mass of the large diameter pin, the wall section was reduced while still maintaining mechanical integrity. Furthermore, the end geometry was optimized to reduce mass, with tapered ends implemented as shown in Figure 3.13 and the drawings of Appendix P.

45

MPa 800

600

400

200

Straight Gudgeon Pin (15 MPa Gas Pressure Loading)

Tapered Gudgeon Pin (15 MPa Gas Pressure Loading)

Figure 3.13: FEM model piston analysis comparing von Mises stress distribution for various gudgeon pin designs.

The initial piston ring design did not utilize a conventional three ring sealing pack, commonly found in OEM designs [219]. Rather, to reduce friction losses and increase brake output (Section 3.3.1), a two piston ring pack consisting of a single compression and a conventional 3 piece expander type oil ring was used. The location of the top sealing compression ring was determined by the requirement of a 4 mm piston land thickness [152], which minimized crevice volume but ensured piston integrity at the high boost pressures. The single compression ring design had previously proven successful in NA high speed applications following developments to overcome seal leakage, oil consumption and heat transfer problems [175]. unknown. troublesome. However, the effects on these troublesome areas when combining the single compression ring design with high boost pressures were Hence, the piston design incorporated provision for a second compression ring, which could be machined into the piston if development proved

46

Previous compression ring studies documented that oil control is a more important attribute than compression control for the second compression ring [219]. Hence, the main function of the second compression ring is to scrape oil off the cylinder wall, with only 20% attributed to sealing the combustion gas that leaked past the first compression ring (~1/20 gas pressure difference across the second ring when compared to the top ring) [219]. Therefore, a positive scraper top compression ring design with 1 of taper was specified, as this type of ring had previously been documented to provide adequate combustion sealing and oil retention in two ring piston designs [175]. A 1 mm ring thickness was chosen, which was 20% thicker than modern top ring designs for similar bore sizes [214, 215]. However, the increased ring section improved piston ring reliability under heavy knocking conditions and also resulted in a higher cylinder wall contact area, which improved the piston heat transfer into the water cooling system [55, 119]. Thicker section rings were also considered but dismissed due to ring flutter problems caused by the increased ring inertia at high engine speeds [175, 220]. A spherical graphite (SG) cast iron material was selected for the top compression ring due to compatibility with the cast iron bore and ductile properties which reduced the likelihood of brittle fracture under heavy knocking conditions [175, 219, 220]. As no commercial top rings of the required configuration could be sourced, components were manufactured by specialty suppliers to drawings shown in Appendix P.

3.5.4 Crankshaft
The design layout of the four main bearing crankshaft was largely dictated by the OEM based crankcase. Hence, all main bearing geometry including dimension and location were OEM based values, with the crankshaft featuring a 36 mm diameter main journal pin. The OEM based sprocket assemblies required for camshaft and integral transmission drives were also integrated into the crankshaft design for packaging and simplicity reasons. A specialty EN30B alloy steel material [146] suited to crankshaft manufacture was utilized, with a yield strength in excess of 1,000 MPa. In-house CNC and manual machining was heavily used in the manufacture of the specialty component, with heat treatment, grinding and balancing processes outsourced.

47

Extensive analysis was completed during the crankshaft design in order to ensure robustness at high BMEP and engine speeds. Special emphasis was placed on optimizing crankpin dimensions and fillet radii, which were critical in increasing the mechanical integrity of the crankshaft [8, 84, 179, 220]. A FEM model was created based on previous studies [84, 155], with loading conditions for both cylinders at various CA positions applied, which produced resultant stresses due to both bending and torsion (Figure 3.14). Loading for the FEM analysis was determined from the analytical model created for the load analysis as described in Section 3.5.1. It is noted that the forces applied to each cylinders crankpin were independently calculated for a wide range of CA positions due to the varying gas pressure and inertia loads associated with the uneven firing frequency (0, 180 CA). Figure 9.1 displays FEM analysis results highlighting the final crankshaft designs angular elastic deformation and resultant stresses due to increased loading throughout rotation. Figure 3.14 also highlights the higher peak stresses in the second cylinders crankpin, caused by the torque transferred from the first cylinder due to the transmission drive path. Consequently, the main and crankpin journal radii were varied to reduce stress concentrations. A 3 mm crankpin radius was chosen, which was a compromise between improving the structural integrity of the crankshaft and allowing sufficient bearing edge side clearance.
Cylinder 1 crankpin

Transmission drive

MPa 100

75

50

25

Cylinder 2 crankpin

Figure 3.14: FEM crankshaft evaluation with bending and torsional loads. Note the higher stresses in the second cylinders crankpin due to the transmission drive path.

48

The final crankpin journal diameter of 38 mm was selected from the FEM analysis after completing the design iteration process. This diameter allowed the crankpin to overlap with the main journal pin which significantly improved the crankshaft stiffness [220]. The chosen value was conservative but correlated to previous relationships established for cylinder displacement and BMEP parameters [8, 220]. Furthermore, the crankpin journal diameter enabled the use of high quality, standard size, OEM, plain bearing shells (Trimetal-F780 [21]) suited for highly loaded applications. Specialized surface heat treatment (nitriding) of the crankpin journals was avoided as this did not improve the mechanical integrity of the crankshaft. In addition, the lack of surface treatment enabled the crankpin journals to be re-ground upon damage to suit alternative/oversize bearing shells, which could avoid the manufacture of a new crankshaft. Extensive dynamic analysis was also completed, with a 100% rotating, 0% reciprocating balance ratio employed in order to minimize rotational inertia and thus improve engine acceleration [14]. It was envisaged that the pure vertical vibration due to the piston motion would produce a high vibration couple, however the associated NVH was dismissed due to the engines original intended FSAE application. Hence the crank cheek was designed with only sufficient material near the crankpin to avoid a heavy balance counterweight mass. The counterweight was designed to provide adequate balance with targeted minimal mass, with simple geometry employed for manufacturing purposes. The crank cheek faces were also knife-edged to reduce windage losses as a wet sump system was selected to reduce the complexity of the lubrication system (Appendix I.4). The crankshafts rotating mass was then fine tuned by balancing statically and dynamically, using bob-weights attached to the crankpin, in order to simulate the rotating portion of the connecting rod mass [41]. Detailed drawings of the crankshaft design are displayed in Appendix P with Figure 3.15 depicting two manufactured crankshaft versions as further detailed in Chapter 5.7.

49

Figure 3.15: Manufactured crankshaft version. (Left): Original design. (Right): Altered crankshaft design as detailed in Chapter 5.7.

3.6 Cylinder Block Design

An open deck design was an obvious choice for the cylinder block due to its well published advantages in cooling and end-gas knock avoidance [55, 176]. As the design brief featured high pressure ratio turbocharging, all design decisions were based on maintaining engine reliability. Sealing a thin walled open deck design has previously been proven to be difficult in high BMEP NA engines, due to a lack of structural integrity in the block [108, 201, 262]. Excessive bore distortion is the underlying problem, which significantly increases piston blow-by. Therefore, preventing bore distortion was critical to the success of this program, as the initial development commenced with single compression ring pistons (Section 3.5.3). Procurement of a cylinder head gasket was problematic due to the engines prototype nature. Any form of gasket featured long lead times and significant expense because of the tooling and development required, with no guarantee of durable sealing. Due to the type of turbocharging employed, manufacturers were very hesitant in supplying any components because of the high probability of gasket failure. Heavy knock was expected to be encountered during the calibration phase, with accompanying piston land and head gasket failures well documented for high BMEP engines [15, 98, 108, 220]. The author also expressed concerns to gasket manufacturers over unacceptable limits in conventional head gasket tolerances and bore alignment. This poor alignment could lead to bore and/or gasket edges being exposed and becoming hot spots within the cylinder, thus increasing the likelihood of spark knock and/or

50

surface ignition. Poor alignment also caused increased crevice volume along the periphery of the chamber, which detracts from burning the available charge energy. This reduces engine performance and increases engine-out HC emissions [17, 98]. Applying current gasket technology to this particular engine was considered to be of high risk for the reasons previously outlined, and as such, a new design was sought and developed. The new design concept avoided the use of a conventional head gasket and relied on the correct amount of face pressure generated by interference between the cylinder head and block to seal the interface. This had advantages in improving the structural integrity of the A detailed description of the novel inherently weaker open deck design.

gasketless interface design for an open deck cylinder block is now outlined, including extensive stress and thermal FEM analysis [17].

3.6.1 Gasketless Interface Design

Design Brief The design criteria for the gasketless interface focused on preventing combustion gas leakage under a wide range of unforgiving engine operating conditions. This included the extremes of pressure and temperature associated with this highly TC application and the potential for sustained heavy knock. The following criteria were adhered to when designing the gasketless interface: Sealing integrity must be sustained over the engines life cycle, expected to be several years and hundreds of hours of operation. The interface design should not affect the longevity of other associated components. The cylinder bore cylindricity must be retained in the assembled state, thus minimizing blow-by rates. The design must be reusable and not suffer any sealing deterioration over its lifespan, irrespective of multiple removals during development. The interface arrangement and other associated components need to be manufacturable using standard methods, at a cost similar to other specialized engines.

51

 In the case of interface failure, preventive measures must be designed into the system to avoid water and/or oil entering the cylinder and causing significant engine damage. Load Conditions and Assumptions The load conditions the interface design was required to withstand were dictated by several assumptions, listed as follows: Peak Pressure: In-cylinder pressures of 7.5 MPa (Section 3.5.1) for normal combustion, with a safety factor of 2 applied to ensure component integrity under heavy knocking conditions. A 15 MPa peak pressure was fairly conservative [98], however the severity of knocking pressure oscillations for these operating conditions were largely unknown. Symmetry: Each cylinder is assumed to be independent due to the engine design featuring a central camshaft timing chain, depicted in the exploded views of Appendix P. This design necessitates each cylinder to have its own water jacket. The cylinder head is retained by a total of eight bolts, four per cylinder. Loading: To ensure sealing, the cylinder head bolt tension is required to This corresponds to 56 kN of force per

overcome the combustion force.

cylinder when the peak pressure condition is applied over the piston bore area. Temperature: The average bulk water temperature is assumed to not exceed 100C. Interface Contact Area: This area is dictated by the piston bore diameter and the selected production liners, resulting in a surface contact area of 18 cm2 between the cylinder head and liner. This arrangement results in a surface pressure of 31 MPa at the interface due to the force generated by the cylinder head bolt tension. [262]. Stiffness: The cylinder head and crankcase were assumed to be infinitely stiff in comparison to the barrel assembly. Consequently, the barrel assembly conforms via elastic deformation to suit the top and bottom constraints. This is fairly conservative in comparison to the 15 MPa value published by Yamaguchi et al. for a new cylinder head gasket design

52

 Thermal expansion: As the cylinder head and barrel were manufactured from aluminum, differential thermal expansion effects of the cast iron liner flange compared with the aluminum surrounds and resultant interface loading were assumed to be negligible when compared to combustion loading. to be minimal and hence ignored due to the thin walled liner section. Barrel and Liner Design The gasketless interface, as indicated by the name, requires no specific gasket or components to seal combustion gases between the cylinder head and block assemblies. This task is instead performed by the barrel and liner assembly, which must also fulfill its normal duties associated with piston motion and heat transfer. Before including the interface effects into the barrel and liner design, decisions were finalized to fulfill requirements associated with cost, performance, reliability and delivery times. These are listed, in order of importance. The block assembly was constructed of a one piece barrel with removable dry liners. This reduced development time and costs in the case of bore damage. An open deck water cooling jacket design was chosen to allow full peripheral water cooling around the cylinder bore and up to the cylinder head deck surface. However, a closed deck design has been proven to be more effective in cylinder head gasket sealability due to its superior structural integrity [262]. The removable liners chosen were production items, manufactured from grey cast iron. These liners had the potential to be replaced by Nikasil coated aluminum liners which have more favorable friction and heat transfer properties upon further development. The barrel was CNC machined from billet due to the low volumes required. A 6000 series aluminum, specifically 6061-T6 [146] was chosen due to its heat transfer properties, high yield strength, excellent machinabilty and reparability. The water cooling jacket design was dictated by manufacturing techniques and packaging restrictions. This resulted in water cooling the top 75% of the piston stroke. Cooling the bottom portion of the bore was considered to have only a minor effect, due to the fourfold reduction in local heat flux values between the top and the bottom of the cylinder [55]. The temperature induced axial liner loading on the cylinder head was also expected

53

Water Cooling Jacket

Piston

Gudgeon Pin

50-60 m Cylinder Head Interference

O-ring Detail

Barrel Connecting Rod

Dry Liner

Figure 3.16: Section view highlighting the cylinder block assembly (barrel and liner) featuring the gasketless sealing arrangement with the cylinder head not installed.

The final cylinder block design (barrel and liner assembly), incorporating the gasketless interface, is shown in Figure 3.16. For the reader to gain some visual understanding of the assembly and how it interacts with associated components, exploded views are displayed in the drawings of Appendix P. For the interface to successfully seal the combustion gas, continuous surface pressure, rather than line pressure on the cylinder bore decks is required [17, 262]. This was achieved by shrink fitting the cast iron liners into the aluminum barrel and then machining the liner surface proud in comparison to the barrel deck surface. This ensured that the surface pressures would be highest between the liner and cylinder head surfaces when the cylinder head was installed, thus creating a robust seal. The purpose of shrink fitting the liners into the barrel was to provide the high positional accuracy and alignment, needed to ensure the liners remained in position under all operating conditions. A H7-p6 locational interference fit was chosen, which ensured the liners remained fixed in the barrel under worse case conditions. These conditions corresponded to engine motoring at the maximum

54

bulk water temperature of 100C.

Under all other operating conditions, firing

ensured higher liner temperatures and thus increased the shrink fit interference. Table 3.1 displays the upper and lower bounds of the fit, with tight tolerances on drawings needed to ensure a positive interference at the motoring condition.
Table 3.1: H7-p6 locational shrink fit upper and lower limits at ambient temperature and worse case conditions.

H7-p6
Maximum Minimum Average

Shrink fit interference (m) at ambient, 20C

+93 +37 +65

Shrink fit interference (m) at motoring, 100C

+33 -23 +5

The liner fitting process involved oven heating the aluminum barrel to 80C and chilling the cast iron liners, with the change in dimensions via thermal effects allowing sufficient clearance for assembly. This process could also be reversed if the liners needed to be replaced. The temperatures associated with shrink fitting were below the aging temperature of the aluminum and thus did not affect the mechanical properties. The gasketless interface has the potential to seal combustion gas, cooling water and lubrication oil as per the requirements of a conventional head gasket. However, some redundancy was built into this design to ensure reliability in case of failure due to the engines prototype nature. The primary purpose of the gasketless interface was to seal combustion gas, with o-ring seals independently used to restrain cooling water and oil. For a limited time, these o-ring seals could also seal combustion gas in the case of interface failure. O-ring grooves were positioned nearer to the water jacket, aiding in their longevity due to the lower local temperatures. This also ensured that adequate surface contact area could be retained to seal the interface. A Viton o-ring material was chosen due to its low cost and ability to withstand continuous temperatures above 200C. sectional view highlighting the o-ring locations is also shown in Figure 3.16. Figure 3.17 displays a sectional view of the assembled engine with the gasketless interface highlighted. As can be seen, the cylinder head loading from the bolt tension, has forced the portion of the barrel in contact with the liner to compress A

55

around the periphery of the bore. The compression has generated a marginally deformed seal between the cylinder head and the liner due to the high face pressures, preventing the liner from moving radially. The evidence of this seal is shown in Figures 9.17 and 9.18, with the liner residual machining marks imprinted on the aluminum cylinder head face. In comparison, a conventional metal or fibrous based gasket allows some radial liner movement due the compliant nature of the gasket material [108, 201, 262]. Thus, the gasketless design is expected to improve the structural integrity of the weak open deck design. For the interface to remain serviceable throughout its design life, only elastic compression was permissible, thus allowing the assembly to return back to its original unloaded position as depicted in Figure 3.16. The correct amount of interference required to remain in the elastic region while ensuring adequate face pressure to generate a robust seal was determined using FEM analysis, with results shown in Figure 3.18.

BLOCK PORTION IN CONTACT WITH LINER Figure 3.17: Assembled section view highlighting the cylinder block assembly (barrel and liner) sealing arrangement. The cylinder head causes the barrel to elastically distort, thus creating a robust seal.

(A)

56

With the previously determined loads applied to the FEM model, a liner vertical displacement of 50-60 m was simulated via the elastic deformation of area A, as obvious in Figure 3.17. This value is the vertical height difference between the liner and barrel decks needed to create a robust seal. The resultant stress from the cylinder head bolt tension creates peak stresses in the region of 50 MPa in the aluminum barrel, well below the 276 MPa yield strength of the material. In the FEM analysis, the cylinder head and crankcase are assumed to be infinitely stiff and the liner is assumed to move axially, due to the high forces generated by the cylinder head bolt tension.
STRESS
MPa 120 100 80 60 40 20 0

DISPLACEMENT
m 60 50 40 30 20 10 0

Figure 3.18: FEM model analysis simulating cylinder head loading on the barrel and liner assembly. (Left): Von Mises stress distribution. (Right): Resultant displacement.

3.6.2 Torque Plate Design

Torque plate honing is a method commonly used in closed deck cylinder block manufacturing, where the loading from the cylinder head bolt tension around the threaded holes in the block distorts the deck. This results in bore distortion [201] as the threaded holes are generally in close proximity to the bore. This bore distortion can be significantly reduced by attaching a torque plate fixture to the cylinder block to simulate the distortion prior to its final manufacturing process [82, 189]. A torque plate is essentially a dummy cylinder head with bore access holes to allow access for the final honing operation.

57

Torque plate honing has been used extensively in the high performance aftermarket and motorsport industries. In recent times, it has become more prevalent in OEM production engines. Creating a bore that is distorted (if at all) to the final operating condition when machining has obvious advantages in reducing piston ring blow-by and oil consumption, as a result of the improved initial seal generated between the piston rings and the bore. This significantly reduces the run-in time and improves the cycle life of many components including the piston, piston rings and bore surface, thus increasing engine reliability and longevity [189]. As per the new gasketless design, significant barrel and liner elastic deformation causes the bore to distort when the cylinder head is assembled, due to the high face pressures needed to seal the interface. Torque plate honing can be used to compensate for this distortion thus returning the cylinder bore to correct dimensions. However, its application differs as near zero deflection is associated with the threaded bolt holes due to the open deck design, with the majority of distortion associated with the interface.
Force F SYMMETRY PLANE F F F

D1 D4 INTAKE: CYL 1

Deflection D3

INFINITELY STIFF LINER

Figure 3.19: Cylinder head stiffness measurement, needed for torque plate design. (Left): Schematic highlighting loaded locations (F) and deflections (D). (Right): Physical testing performed.

58

The engine differs from conventional one piece block designs, in that it consists of multiple components including a split crankcase featuring an integral clutch, gearbox and barrel assembly. This makes it more difficult to apply torque plate honing to this type of engine. It was not feasible to complete the process with the cylinder block attached to the crankcase due to the compactness of the design and bore access restrictions. However, it was assumed that the crankcase acted as a rigid one piece body due to the high number of bolts used at joints. Thus, the crankcase was replaced by a stiff fixture which restrained the bottom of the barrel but also allowed bore access. Physical load deflection testing was completed in order to finalize the torque plate design and determine the material and thickness required to simulate the cylinder head deck stiffness. FEM analysis was avoided due to the difficulties associated with modeling the cylinder head casting irregularities and inconsistent material properties. The schematic and experimental setup is displayed in Figure 3.19, showing the load locations (F) and resultant deflections (D1-D4). The load was applied via a hydraulic press, with a dial indicator used for deflection measurement. A near infinitely stiff fixture, replicating the interface contact region was used as a substitute for the barrel and liner assembly, thus ensuring all deflection occurred in the cylinder head
25 D1 (inner) Force F at Bolt Holes (kN) 20 15 10 5 0 0 5 10 15 20 25 30 35 40 Head Surface Vertical Deflection D (m) D2 (inner) D3 (outer) D4 (outer)

Figure 3.20: Results from load deflection tests performed on the cylinder head to determine the relative cylinder head stiffness needed for the torque plate design.

59

Results from load deflection testing are shown in Figure 3.20, with trends suggesting that the deck stiffness varies between the inner and outer bolt locations. This was anticipated as the cylinder head featured open ends with exposed water jacket detail, shown more clearly in the exploded views of Appendix P. Using the cylinder head experimental data, specifications concerning the torque plate were finalized with the aid of the FEM analysis shown in Figure 3.21. Final dimensional thickness was determined to be 40 mm when manufactured from mild steel via laser cutting, with the surface ground flat to replicate the cylinder head deck surface.
VERTICAL DISPLACEMENT
m

SYMMETRY CONSTRAINT

D1

D4 F = 15 kN

60

45

30

15

D2

D3
A

Section A-A
0

Figure 3.21: FEM model analysis simulating torque plate effects to determine the thickness needed to represent the cylinder head. (Left): Vertical displacement. (Right): Vertical displacement along section A-A.

3.6.3 Thermal Analysis

A thermal FEM analysis was completed for the engine with Stryker [155] to determine the effect of the gasketless design on the surface temperatures of the cylinder head and the liner at the operating conditions of peak pressure and exhaust blow-down. The critical areas were assumed to be the bridges (exhaustexhaust, intake-intake and intake-exhaust). To analyze these areas, two sets of models were created with CAD images depicted in Figure 3.22.

60

Figure 3.22: and CAD models used in the thermal FEM analysis, also highlighting the division of the cylinder bore into five thermal regions.

The first set contained cylinder models, created along the plane of symmetry through the exhaust valve bridge to the intake valve bridge. One model featured the gasketless interface, while the other had a 1 mm thick gasket between the cylinder head and block. Both valves were closed in this set. The model created by sectioning through the exhaust valve and intake-exhaust valve bridges was used in the second set. This set contained four models, distinguished by having a gasket or not and whether or not the exhaust valve was open 2 mm or closed. Boundary Conditions The models were created to represent conditions during combustion while the models were created to represent conditions during combustion and exhaust blow-down. The boundary conditions applied for each of the models are listed in Table 3.2. The combustion phase boundary conditions were estimated using the following criteria: The convective heat transfer coefficients used for the combustion surface were calculated using the operating conditions (Tgas, Pcylinder) at peak pressure with the Woschni correlation [256].

61

 The convective correlation for the top 20 mm of the bore surface were assigned values equal to half that of the combustion surface [156]. The exhaust port convective coefficient was calculated using the equation given by Caton et al. for the exhaust valve closed [44]. The fixed bore temperatures and the cylinder head and bore water jacket surface temperatures were assigned values based on Cloughs work [55]. For the exhaust phase, the boundary conditions were determined using experimental temperature and pressure data at the 2 mm open exhaust valve position. Values were applied to calculate a combustion surface convective coefficient using the Woschni correlation [256]. The convective coefficient for the exhaust port calculated using an equation published by Caton et al. [44]. The bottom 80 mm of the bore was assigned a fixed temperature [55].

Solution Method

COSMOS-Works was used to mesh the model, input boundary conditions and
solve for a steady state solution. Although the convective boundary conditions were intended to be used with a transient analysis, it was felt that they gave the proper proportional relationship for the different coefficients as applied to the different surfaces. The steady state solution would give surface temperatures This hypothesis was much higher than a transient solution at a specific CA, however the temperature relationship between the models should be consistent. validated by sensitivity analysis - varying the convective coefficient and ambient temperatures associated with the boundary conditions. Although results showed magnitude changes in temperature, the relationship between the models remained consistent. Additionally, the resulting temperature distributions were similar (although with higher values) to those found in the literature [5, 26, 47, 55, 156]. Therefore, the boundary conditions used in this analysis were appropriate for the intended comparative analysis.

62

Table 3.2: Boundary conditions used in all steady state thermal FEM analysis (Upper): Combustion. (Lower): Exhaust blow-down. Note the higher liner temperatures at the bottom of the bore due to the absence of water cooling and exposure to crankcase oil. Boundary conditions associated with combustion BC's associated with combustion
type Cylinder top 20 mm of liner 20-40 mm of liner 40-60 mm of liner 60-80 mm of liner 80-100 of liner block water jacket Valves valve face valve edge upper valve stem lower valve stem Head head combustion surface exhaust port head water jacket gas temp fixed temp (C) (Kelvin) 2853 140 130 135 140 95 convective coefficient (W/m K) 1000

conv temp temp temp temp temp

conv perfect contact temp conv

2853 100 800

2000

200

conv conv temp

2853 800 100

2000 200

Boundary conditions associated with exhaust blow-down BC's associated with exhaust blow-down
type Cylinder top 20 mm of liner 20-40 mm of liner 40-60 mm of liner 60-80 mm of liner 80-100 of liner block water jacket Valves valve face valve edges lower valve stem upper valve stem Head head combustion surface exhaust port head water jacket gas temp fixed C) (Kelvin) temp ( 1300 1300 1300 135 140 95 convective coefficient (W/m K) 800 800 800

conv conv conv temp temp temp

conv conv conv temp

1300 1100 1100 100

1300 1200 1200

conv conv temp

1300 1100 100

1300 1200

63

EXHAUST VALVE

INTAKE VALVE

Figure 3.23: Temperature distribution on the combustion surface for the model, indicating that the adiabatic assumption through the intake exhaust plane can be assumed for the model.

Although the plane through the exhaust-intake bridge that formed the model was not truly adiabatic, results from the model indicate that an adiabatic assumption for this plane would have little effect on the resulting temperature distribution or temperature differences. The relatively symmetrical temperature pattern in the bridge area can be seen in Figure 3.23. Thermal Results A steady state analysis was completed with model results shown in Figures 3.24 and 3.25 and model results in Figure 3.26. Figure 3.24 shows a comparison of head surface temperatures along the plane of symmetry for the gasket and gasketless models at peak pressure. The gasketless model has lower surface temperatures when compared to the gasket model at each location. This occurs as the metal to metal contact at the interface allows increased heat (thermal energy) to flow from the head to the liner. The difference between the models is most noticeable near the interface region and becomes negligible towards the centre of the bore. occur in this region [98, 192]. intake, as expected. This has relevance as it indicates potential It is also noted that at the interface, the reductions in end-gas temperature, thus decreasing the likelihood for knock to temperature differential was higher on the exhaust side when compared to the

64

Gasket (266C)

Gasketless (243C)

600 500 Temperature (C) 400 300 200 100 0 -45 -30 -15 0 15
EXHAUST INTAKE

Gasketless Gasket

INTERFACE REGION

SURFACE AREA EXPOSED TO BORE

30

45

Radial distance from bore centre (mm)

Figure 3.24: Cylinder head surface temperatures at peak combustion pressure along the symmetry plane line for the model. The area highlighted indicates the reduced temperature due to the gasketless interface.

300 Gasketless 250 Temperature (C) Gasket

200
GASKET BLOCK

150

100 -5 0 5 10 15 20 Distance from Bottom of Cylinder Head (mm)

Figure 3.25: Liner surface temperatures at peak combustion pressure along the symmetry plane line on the exhaust side for the model.

65

600 500 Temperature (C) 400 300 200 100 0 0 10 20 30 to Intake Bridge 40 Exhaust Bridge
EXHAUST BRIDGE REGION EXHAUST/ INTAKE BRIDGE REGION

Gasketless (valve closed) Gasket (valve closed) Gasketless (valve open) Gasket (valve open)

50

60

Figure 3.26: Cylinder head surface temperatures for peak combustion pressure and exhaust blowdown models along the planes defining the model.

Figure 3.25 shows the bore surface temperatures as a function of distance from the bottom of the cylinder head on the exhaust side. The gasket model results show a large temperature gradient across the face of the gasket surface due to the low thermal conductivity of the gasket. The results confirm the findings of Figure 3.24, with the lower surface temperatures of the cylinder head resulting in higher liner surface temperatures due to the increased heat flux at the interface for the gasketless model. Two sets of data are presented in Figure 3.26. Both sets depict the cylinder head surface temperatures at nodes along the planes used to form the model, starting at the exhaust valve bridge and ending at the intake-exhaust valve bridge. This figure confirms that the gasketless model has lower temperatures than the gasket model, as previously shown in Figure 3.24. The effect of the variation in the cylinder head wall thickness on surface temperatures is evident by the dip in the head surface temperature plots. It should also be noted that the temperature pattern and values in the vicinity of the exhaust valve bridge are similar for the half and quarter models. The other set of data shown is for the same surface when the exhaust valve is open. The same pattern is evident, although at much lower temperatures corresponding to the exhaust phase boundary conditions. Hence, the thermal analysis results concluded that removing the conventional head gasket improves the heat flow between the cylinder head and block

66

assembly. This is due to the absence of a gasket, which behaves as a heat insulator. The improved heat flow reduces the likelihood of hot spots forming around the periphery of the chamber, thus reducing the propensity for abnormal combustion (spark knock or surface ignition) in the end-gas region.

3.6.4 Manufactured Design

Figure 3.27 displays the final manufactured cylinder block design. The left image displays the barrel after CNC machining with the right image showing the block assembly complete with shrink fitted liners and reduced shank studs [220] ready for engine installation.

Figure 3.27: (Left): Manufactured aluminum cylinder barrel. (Right): Barrel with shrink fitted cast iron liners ready for engine installation.

3.7 Manifold Design

Both intake and exhaust manifolds incorporated Watsons kinetic energy conservation (KEC) rolling flow plenum chamber design. In this design, the kinetic energy of the flow is assumed to be conserved in a vortex about the axis of the plenum [13-15]. A flow path of Watsons design can be seen pictorially in Figure 3.28, where the inlet and outlet ducts are offset and placed tangentially to the plenum walls. Watsons KEC log style rolling flow design had successfully been applied to intake manifolds [13-15], but until now has not been implemented to exhaust manifold designs. This concept was suited to the exhaust manifold as the design of the plenums in between the compressor, engine and turbine were critical to offset the flow pulses, as an uneven firing interval was selected (Section

67

3.3).

Optimization of both the intake and exhaust plenum systems were

completed through an extensive series of simulations using Ricardos WAVE, with results presented in Chapter 6.5. Concepts for improving intake manifold air consumption and cylinder filling are now discussed. Optimizing the length of the intake runner in order to take advantage of resonant wave tuning has been well documented in the literature [94, 98, 220], with Figure 3.29 displaying the dynamic pressure effects measured at the intake valves. This resonant wave, when correctly timed in the induction process, improves air consumption and cylinder trapping while reducing the exhaust residual gas content within the cylinder [14, 60, 94, 98, 220]. Resonant wave instantaneous pressures in the order of 180 kPa absolute have previously been documented in high BMEP NA racing engines [102, 246]. In boosted applications, resonant tuning is generally less important since the engine airflow can be increased using the compressor. However, resonant tuning can be used to help improve air consumption at engine speeds where boost levels are limited by compressor delivery rates, especially at lower engine speeds in TC applications. Hence, it was intended to use resonant tuning at low engine speeds in order to improve engine performance at conditions where the restrictor had not yet limited airflow. Consequently, a wave propagation theoretical model was developed to explore the resonant wave effects at various operating conditions.
Plenum Chamber

Inlet Outlet

Conventional

KEC plenum design

Figure 3.28: (Left): Conventional plenum design. (Right): Watsons KEC log style rolling flow design where the kinetic energy of the flow is conserved in a vortex about the axis of the plenum.

68

180
Absolute dynamic port pressure measured at the intake valve (kPa)
3 pressure waves per engine cycle corresponding to peak resonant tuning Increasing engine speed

140

100

60

IVO
20 0 180 360 Crank Angle (deg) 540

IVC
720

Figure 3.29: Experimental data from a NA racing engine showing the effect of engine speed on the dynamic pressure waves in the inlet tract @ WOT [102].

Figure 3.30 displays the effective tract length (centerline distance from the intake valve to the highest trumpet edge) and corresponding engine speed for the primary and secondary resonance generated from the theoretical model. motion specifications outlined in Table 3.3.
Effective tract length (m) Resonance Inlet Runner Primary Length (mm) for a 3 Wave Tract 12000
0.25

It is

noted that the results only apply to the proposed two cylinder engine with valve

Effective tract length (m) Secondary Resonance Inlet Runner Length (mm) for a 4 Wave Tract 12000
11000
0.20 0.22 0.24 0.27 0.30 0.35 0.40 0.50 0.60 0.80

11000 10000 Speed (rev/min) 9000 8000 7000 6000 5000 4000 3000 0

0.27 0.30

10000 Speed (rev/min) 9000 8000 7000 6000 5000 4000 3000

0.35 0.40 0.45 0.50 0.60 0.70 0.90

20 40 60 80 Temperature (C) Ambient Temp. (degC)

100

20 40 60 80 Temperature (C) Ambient Temp. (degC)

100

Figure 3.30: Effective inlet tract lengths created from a wave propagation theoretical model highlighting the tuned engine speed for peak primary and secondary resonance at varying ambient conditions.

69

However, the results may serve as a guide to determine the tuned effective tract lengths for other applications. Gas temperature effects associated with ambient conditions and injector location result in changes to the speed of sound, which affect wave propagation rates and thus tuned lengths. The primary resonance, associated with peak resonant tuning corresponds to three residual waves forming per engine cycle as highlighted in Figure 3.29. The secondary resonance corresponds to four residual waves per cycle and is not as dominant [60]. An effective tract length of 400 mm was initially selected as a starting point for the manifold design and subsequent experiments, providing peak primary and secondary resonance at mid-range speeds in the ~ 5,000 - 7,000 rev/min region. This region was critical to achieving the near constant power objectives, placing high boost demand on the turbocharger and hence an obvious area where resonant tuning could be effectively used. Longer tract lengths gave improved lower speed performance but severely compromised FSAE vehicle packaging. Tract length analysis through WAVE also showed that intake resonant tuning had little effect on performance at operating regions where air consumption was flow restricted. However, resonant tuning was shown to improve air consumption in regions where the turbocharger could not provide enough airflow to reach the flow restricted limit. The PFI intake manifold design featured a combination of polymer and metallic materials to reduce mass and simplify the manufacturing complexity, with detailed drawings shown in Appendix P. The intake runners were manufactured from a Nylon66 material using the selective laser sintering (SLS) process. This enabled the fuel injector bosses to be easily incorporated into the design, with the SLS runner simply constructed from a 3-dimensional CAE model. The intake plenum chamber was fabricated from thin wall aluminum tubing, with inlet runner trumpet geometry incorporated into the design. The volume of the plenum chamber was an important consideration due to the odd firing and consequent intermittent pulses through the flow restriction, with simulation results used heavily in the design (Chapter 6.5.2). The plenum chamber also incorporated individual cylinder inlet runner trumpets, with airflow bench testing (Appendix H) finding that it was

70

advantageous to have a bell-mouth trumpet entry into the runner to minimize flow losses as also documented by Richardson et al. [188]. The multi-piece inlet manifold design also incorporated the possibility for tract length variation at the interface of both materials, which could be used to fine tune the inlet manifold in dynamometer experiments. The feasibility of intercooling the boosted charge was also explored, but eventually dismissed due to the upstream throttle location and the possibility of reduced transient response at low engine speeds due to mass transport delays. Furthermore, the expected temperature reductions from intercooling were not large when considering the boosted intake temperatures (Chapter 6.5.2) and intercooler effectiveness. In addition, the large aluminum manifold surface area and high turbocharger efficiencies were expected to reduce intake air temperatures. Developing the engine without an intercooler gave mass, packaging and cost benefits together with simplifying the complexity of the intake system. Furthermore, although power increases are associated with intercooling [143, 252], these benefits were not large for this particular setup due to the limited airflow associated with the flow restriction. Hence, performance improvements due to the cooler intake charge were caused by possible CR increases due to likely knock suppression with lower cycle temperatures [98]. Pre-injecting fuel into the restrictor throat was also examined in an attempt to increase air consumption at the choked mass flow limit. The possible increases in air consumption at the choked limit are associated with the increased charge density due to the lower air temperatures at the restrictor, caused by the phase change latent heat when vaporizing the liquid fuel [220]. Some concerns were raised about the engines transient response with pre-restrictor injection, even though the turbocharger compressor is documented to assist in fuel breakup and charge mixing [248]. However, the long mixing length after the compressor had the potential to cause fuel pooling and wall wetting along the length of the intake manifold. Hence, it was envisaged that the extra injector would only be operated at wide open throttle (WOT) operating conditions at engine speeds where the restrictor limited airflow.

71

Table 3.3: Test engine specifications. UniMelb WATTARD TYPE Parallel twin, 4 stroke SI, Liquid-cooled, Integral clutch/ transmission 433.8 mL 69 x 58 mm Unequal (0, 180 CA) 9-13:1 with piston crown modification Pent roof central spark plug Aluminum head/ barrel/ crankcase/ piston, Cast iron dry liner, Alloy steel connecting rod/ crankshaft 4 x 36 mm main journal, 2 x 38 mm crankpin journal, 20 mm gudgeon pin, 116 mm connecting rod length 8-valve DOHC IVO: 20 BTDC, IVC: 70 ABDC EVO: 60 BBDC, EVC 15 ATDC Wet sump, Dry sump NA, SC (Roots), TC (Garrett GT-12) Carburetion, Sequential PFI 2.5L KEC plenum manifold Motec M4 98-RON pump gasoline Multi wet plate 3 speed constant mesh 47 kg dry

CAPACITY BORE x STROKE FIRING INTERVAL COMPRESSION RATIO COMBUSTION CHAMBER MATERIAL

ROTATING/ RECIPROCATING ASSY

VALVE ACTUATION VALVE TIMING (Chapter 6.5.3)

LUBRICATION (Appendix I.4) INDUCTION (Chapter 4.5) FUEL DELIVERY (Chapter 4.5) EXHAUST (Chapter 6.5.1) ENGINE MANAGEMENT (Chapter 4.2.5) FUEL (Chapter 4.4) CLUTCH (Appendix I.2) TRANSMISSION (Appendix G.2) MASS BARE

Figure 3.31: (Left): Measured, inspected and cleaned components prior to assembly. (Right): First complete engine assembly prior to dynamometer installation.

72

Analysis to determine the viability of pre-restrictor injection at the choked condition was performed, assuming all fuel was injected into the restrictor and completely vaporized (Appendix D.4). Calculations showed the potential for an 11.4% charge increase at the choked limit, however, once accounting for the vaporized fuel portion the mass flow rate of air increased by 9.5%. Furthermore, a major benefit of pre-restrictor injection was found to be a temperature reduction at the intake valves (~20C), as it was beneficial to cool the intake charge through fuel vaporization prior to entering the compressor. This has benefits in reducing the knock propensity. The pre-restrictor injection concept was later rejected due to its added complexity, which made it difficult to implement into a FSAE vehicle.

3.8 Final Specifications and Assembly

Table 3.3 displays the final engine specifications, with information in parenthesis indicating where further detail not outlined in this chapter can be found throughout the thesis. Prior to engine assembly, all components were carefully inspected, detailed and measured to ensure components lay within specified tolerances (Appendix P). This ensured adequate engine clearances when In addition, all assembled to ensure reliability and component longevity. monitored during engine testing and development.

component measurements were documented to enable wear rates to be Figure 3.31 displays components prior to assembly and the first completely assembled engine.

3.9 Summary
In this chapter, the detailed design and analysis was presented for a small capacity (0.43 L) inline two cylinder configuration which was capable of withstanding the high output (exceeding 25 bar BMEP) associated with highly TC engines. The design philosophy placed emphasis on maintaining engine reliability under harsh operating conditions while maintaining design simplicity for ease of manufacture, low cost and time constraints. Such an approach integrated both OEM and custom made components into the final engine design as no suitable OEM production engine could be sourced and adapted to the high output operating conditions. The design and manufacture of new components to suit TC operation also led to some interesting concepts and novel designs.

73

The importance of using proven design techniques involving empirical data, calculations and CAE analysis during the design iteration process has clearly been demonstrated. These tools enabled improved understanding which led to enhanced reliably at reduced costs and lead times. It was difficult to include all aspects of the engine design in this chapter. Rather, the detailed design and analysis of only key features which were vital in ensuring reliability for the highly TC package are discussed, with further detail surrounding all components supplied in the engineering drawings of Appendix P. Particular focus in this chapter was placed on the rotating and reciprocating assembly components due to the elevated loads associated with high BMEP and engine speeds. The block design, which included a novel gasketless interface for an open deck cylinder block, has also been presented. A major advantage of the gasketless design was the improvement of the structural integrity of the inherently weak open deck block design. This was achieved by increasing the local face pressure at the interface between the liner and cylinder head deck surface. Improving the structural integrity was critical to the success of the TC engine program due to the unforgiving operating conditions associated with near 300 kPa MAP operation. The gasketless design also provides a cheap and reliable solution to frequent cylinder head removal, as no components require replacement. Furthermore, the design overcomes costly and time consuming problems associated with supplying speciality head gasket components, and is therefore ideal for a prototype engine design. Moreover, the application of the gasketless interface provides performance, efficiency and emission benefits due to the elimination of any crevice volume near the interface which can exist when conventional head gaskets are used. Extensive thermal analysis also concluded that removing the conventional head gasket improves the heat flow between the cylinder head and block assembly. This is due to the absence of a gasket, which behaves as an insulator. The improved heat flow reduces the likelihood of hot spots forming around the periphery of the chamber, which reduces the propensity for knock to occur in the end-gas region. The possibility of gasket failure due to abnormal combustion is also eliminated with the gasketless design.

74

This chapter highlights that the mechanical aspects of the component design can be overcome for a highly TC engine package and hence are not the performance limiting factors. This is demonstrated with engine components being designed to withstand the high mechanical loads associated with near 300 kPa MAP operation and engine speeds exceeding 10,000 rev/min. This was achieved by utilizing good design techniques coupled with high quality materials and manufacturing processes. Although the cost of these manufacturing and material expenses are higher (density basis) when compared to current OEM production engines, it is feasible that overall unit costs may be comparable to OEM engines with similar power output as significantly less material is required due to the swept capacity and cylinder reduction [9, 12]. Furthermore, although it is possible that in service operating costs may also increase, the annualized whole of life cost is expected to decrease when considering the fuel economy savings [9, 12, 43].

75

76

CHAPTER 4
P O W E R E D B Y W A T T A R D

Experimental Apparatus and Methodology

4.1 Introduction
This chapter outlines the engine experiments, including the objectives used to formulate the test modes (NA, SC and TC) and desired operating conditions. The test methodology is also discussed, which includes the test matrix together with the test sequences and procedures. The original intent of the project requirements centered on FSAE, with aims of achieving success in competition by using a superior downsized engine package when compared to conventional OEM motorcycle units. However, initial results achieved with the new engine design warranted further examination, in particular defining the performance limitations for small gasoline engines (~0.5 L).

77

To carry out this small engine investigation, experiments were designed to operate the test engine over a wide range of speeds with varying MAP, associated with NA and boosted modes. The experiment test sequence was chosen to accommodate the development of the test engine together with the engines installation into FSAE vehicles. To develop the engine and obtain experimental results, a stationary engine test rig was commissioned with subsystems including instrumentation and data acquisition outlined in this chapter.

4.2 Engine Test Rig

In order to commission the new engine design and test the hypothesis, a stationary engine test rig was commissioned by the author, making use of an already existing dynamometer for load absorption. From the onset, the test rig underwent extensive development over the testing period to cater for the required needs. However, initial versions simply enabled engine operation to reduce the development complexity while commissioning the new engine design. Subsystems of the commissioned rig are outlined below, with the final version of the test rig together with schematic shown in Figures 4.1 and 4.2. Dynamometer and control unit (existing) Engine cradle (commissioned) Dynamometer-engine coupling (existing but developed) Water cooling system (commissioned and developed) Exhaust system (commissioned and developed) Fuel supply system (commissioned and developed) Electrical system (commissioned and developed) Instrumentation and data acquisition (commissioned and developed) Fuel and air flow measurement Exhaust gas and air-fuel ratio (AFR) analysis Blow-by measurement Engine cooling water flow measurement In-cylinder pressure measurement

The following sections provide information on the test rig subsystems, with further detail surrounding development and calibration outlined in Appendix J.

78

In-cylinder pressure computer (DAQ)

Blow-by rotameter

Warning gauges/lights, individual cylinder control.

Throttle

Dynamometer control unit Cooling fan

ECU computer Engine (TC mode)

Cylinder 1 EGA Exhaust system

Cylinder 2 EGA Fuel supply and measurement

Water cooling system

Dynamometer

Coupling

Engine cradle

Wiring loom junction box

ECU and CDI modules

Figure 4.1: Final version of the developed experimental rig. (Upper): Control panel. (Lower): Experimental rig testing facility.

79

Charge Amplifier Cylinder Pressure

PC Data Acquisition

CA

Fuel Rig (Tank and Beam Balance)

Exhaust Gas Analyser (NOx, HC, CO, CO2, O2) and AFR. CYL 1

Exhaust Gas Analyser (NOx, HC, CO, CO2, O2) and AFR. CYL 2 Exhaust Intake Plenum Chamber
1

Plenum Chamber

Cooling System

MAP MAT

Head Barrel CA Blow-by

`

Connected to Central Cooling Tower Water Supply

Coupling Eddy Current Dynamometer Crankcase Flywheel

Ignition Control

Engine Cradle Load Cell Reference - Synchronization Dynamometer Controller

Injector Control for PFI modes

Engine Control Unit (ECU)

PC Controller

Figure 4.2: Experimental setup illustrating the basic schematic layout of the engine, including controllers, sensors and data acquisition systems.

80

4.2.1 Dynamometer and Control Unit

A water cooled, eddy current type Heenan and Froude brake dynamometer as depicted in Figure 4.3 was used to absorb load. Rated at absorbing 110 kW As the between 2,400 and 6,000 rev/min, it features a water-cooled field stator and rotor to dissipate the heat generated as a result of the torque absorption. system did not feature motoring capabilities, individual cylinder fuel and ignition disconnection allowed motoring traces (Appendix J.8) to be established in the non-firing cylinder. Hence, the crankshaft was driven by the corresponding firing cylinder with fuel and ignition disconnection activated by the operator through the control panel as depicted in Figure 4.1. Large plenum volumes in the intake and exhaust systems isolated cylinders, thus minimizing interference effects. Randomizing the fuel and ignition disconnection also allowed Morse tests to be conducted at near operating in-cylinder temperatures to estimate individual cylinder performance and friction losses [211].

Figure 4.3: Heenan and Froude Dynamatic MK-1 eddy current dynamometer and OZY-DYN type MISD109-01 control system.

An OZY-DYN type MISD109-01 controller allowed constant speed engine testing at user defined dynamometer speeds, with the load varying by altering the throttle position (TP) and hence manifold pressure. The control system required dynamometer shaft speed and torque inputs via a magnetic speed sensor and load cell unit. Quarterly load cell calibration was performed by applying known masses to a beam at a fixed distance from the dynamometer pivot axis, ensuring

81

results were accurate and repeatable.

Further detail outlining brake torque Relative humidity,

measurement and calibration is given in Appendix J.2.

calculated via wet and dry bulb temperatures, together with barometric pressure allowed raw brake data to be corrected to ISO-3046 international standards for internal combustion engine performance [19].

4.2.2 Dynamometer-Engine Coupling

As a result of the engine design, featuring an integral clutch and transmission within the crankcase, coupling of the engine crankshaft directly to the dynamometer proved unfeasible. Consequently, the dynamometer shaft was coupled to the transmission output shaft, with testing completed in final gear (2.98:1 reduction ratio) to ensure minimal differential speeds between the crankshaft and the dynamometer. The reduction ratio also allowed experiments to be completed over the entire speed range due to the dynamometers maximum speed limitation of 4,000 rev/min. Hence all brake data presented in this thesis corresponds to the performance at the transmission output shaft and not at the crankshaft. Performance at the crankshaft is expected to be marginally higher, due to the reduction in friction losses associated with driving the transmission components. However, differences are anticipated to be less than 3% over the engine operating range [159]. Experiments were initially completed using an already existing direct coupled splined shaft [187]. However this coupling method proved troublesome due to misalignment issues between the dynamometer and engine transmission shafts, eventually resulting in dynamometer bearing seizure and subsequent engine failure. Further detail outlining the coupling development is given in Appendix J.3, with the final iterative solution featuring a chain drive with dynamometer shaft torsional damping. This eliminated shaft misalignment and avoided direct coupling to ensure engine reliability in the case of reoccurring dynamometer seizure. Brake data correction was needed to compensate for the higher losses associated with the chain drive. couplings is given in Figure J.4. The calibration correlating the two drive

82

4.2.3 Engine Cradle

An engine cradle was designed and manufactured to mount and locate the test engine to an already existing test bed, with CAD and manufactured images displayed in Figure 4.4. The design minimized the complexity and downtime in installing and removing the engine, with provision for overhead crane access. Mounting the engine to the cradle underwent considerable development in order to minimize the vibration effects and associated problems (Chapter 5.5 to 5.7), with success eventually achieved with elastic 70 durometer rubber bushes. Consistent positional location was achieved using tapered dowel pins, allowing quick engine, cradle and test bed alignment for torque transmission, which minimized the likelihood of coupling failure (Section 4.2.2). As the cradle had no mass constraints, the design incorporated ease of manufacturing with high safety factors in order to sustain resultant loads and vibrations associated with high speed engine operation.

Figure 4.4: Engine cradle used to mount and locate the test engine to the dynamometer. (Left): CAD image. (Right): Fabricated cradle frame, housing the test engine.

4.2.4 Cooling system

Initially, an external water to air heat exchanger (radiator) from a small passenger vehicle was used to control engine coolant temperatures. However, the initial system was quickly replaced with an existing open water to air cooling tower as the radiator was unable to dissipate the required energy (~20-30% of the fuel

83

energy [95]) due to insufficient airflow through the system. The cooling tower regulated water temperatures by replacing a portion of the hot water with cold, supplied by the laboratory cooling system. This system worked successfully and was later configured to electronically control water temperatures by triggering the opening of the control valve through the engine control unit (ECU). However, after engine reliability concerns associated with water cooling during vehicle operation (Appendix I.5), the test rig cooling system was further developed to be more representative of the vehicle system. This involved adding a header tank swirl pot and pressurizing the test rig cooling system to 2.3 bar absolute. Pressurizing the cooling system involved the implementation of a water to water heat exchanger, which was submerged in the existing cooling tower, allowing temperatures to be controlled as previously described. Pressurizing the cooling system raised the boiling temperature by 25C [208], which minimized local film boiling and steam pockets enabling engine operation at elevated temperatures to ensure engine reliability. No further cooling system problems were recorded during dynamometer or vehicle operation. Further detail behind the test rig cooling system is given in Appendix I.5 and J.6, with drawings and schematics displayed in Appendix P.

4.2.5 Electrical System

A Motec M4 ECU was used to control ignition settings for all modes and the fuel delivery rates for the PFI system. The ECU featured individual cylinder sequential ignition and injection control. This allowed for the manipulation of injector timing and pulse width together with spark timing (ST) at user defined mapping sites. Mapping sites were functions of engine speed and MAP, varying in 500 rev/min increments from 2,000 to 11,000 rev/min. MAP sites were arranged in 10 kPa increments, starting from 20 kPa absolute to the maximum MAP condition at WOT (Table 4.5). Besides fuel and ignition control, the ECU also featured four outputs which could be configured to control auxiliary systems based on sensor inputs. During certain phases of testing, these outputs were used to control the electric water pump, fuel pump, thermatic electric fans, cooling tower water flow valve, turbocharger wastegate and to trigger external warning lights. Operator interface

84

was achieved through a personal computer connected to the ECU via an RS-232 link. This allowed operating parameters to be altered in real time for open loop calibration, whilst monitoring sensor inputs. The ECU also allowed data logging, with information later viewable through Motec Interpreter software [158]. The ECU required sensor inputs for reliable engine operation. Both engine crank angle (reference) and cycle position (synchronization) were detected using two GT-101 hall-effect sensors. Significant development was needed to allow the correct operation of the reference and synchronization sensors due to the high speed operation and crankshaft torsional vibrations associated with the odd firing twin configuration (Chapter 5.3.2). A Delco 3 bar MAP sensor sampling from the end of the plenum chamber allowed the ECU to monitor manifold pressure. In order to minimize transient delays, the connecting passage was minimized with a restriction placed in the sampling line to dampen the pressure fluctuations caused by the inlet resonant waves. The manifold air temperature (MAT) sensor was Water and oil temperature sensors were Other ECU inputs which allowed located centrally in the intake plenum chamber to ensure free stream stagnation air temperature measurement. positioned near the highest temperature regions in order to provide sufficient warning in the case of overheating. Appendix P providing further information. The non-wasted spark ignition system included a two channel M&W Pro-12 capacitive discharge ignition (CDI) module connected to coil on plugs (COP) as outlined in Table 4.1. The system required varying levels of ignition energy during development for performance and reliability reasons as outlined in Chapters 5.3.1 and 9.5.1. Establishing the ECU crank index reference relative to TDC for injector and spark timing was achieved with an ignition timing light (strobe light triggered from the first cylinders spark discharge). A removable lead positioned between the spark plug and COP allowed timing light access. The ECU spark discharge was set at TDC, with the crank index then altered until the TDC indentations highlighted by the timing light were visually aligned at idle speed. Appendix J.8. The ECU phasing relative to TDC is estimated to lie within 0.5 CA as further described in compensation and enhanced reliability are outlined in Table 4.1, with drawings in

85

A wiring harness [139, 210] connected all components of the electrical system via a junction box, including sensors, fuel and ignition systems, together with the control panel, as displayed in Figure 4.1. The control panel included individual cylinder fuel and ignition disconnection, together with gauges and warning lights to monitor system temperatures and pressures.

Table 4.1: Electrical system fuel and ignition modules together with ECU sensor inputs required for engine operation. Fuel and Ignition Modules ECU Ignition module Ignition coil Motec M4 M&W CDI Pro-12 S3 Denso COP 129700-3440 J044

ECU Sensor Input Crankshaft reference sensor Camshaft synchronization sensor MAP sensor MAT sensor Water temperature sensor Oil temperature sensor Throttle position Oxygen sensor Boost control valve (TC mode only) Honeywell GT-101 Hall effect Honeywell GT-101 Hall effect Delco 3 bar 53003 Bosch 0280 130 023 Bosch 0280 130 026 Bosch 0280 130 026 Honeywell M518-1 Bosch LSM-11 0258 104 004 Delco 1604-0749

4.2.6 Exhaust System

Ducting of the engine-out exhaust gas into the laboratory system was initially seen as trivial. However, initial versions resulted in catastrophic exhaust manifold failures (Chapter 5.6) due to solid mounted exhaust ducting. Failures were caused by high speed engine vibrations, which resulted in exhaust manifold vibrations which conflicted with the almost fixed laboratory system. These problems were overcome by installing a short length of flexible stainless steel piping, positioned on the outlet of the exhaust manifold to isolate both systems.

86

4.3 Instrumentation and Data Acquisition

4.3.1 Fuel and Air Flow Measurement
Fuel consumption measurement was achieved gravimetrically using a beam balance mechanism, with a schematic of the designed and manufactured system displayed in Figure 4.5. Injector calibration at duty cycle intervals allowed fuel To accurately quantify flow rates for the PFI modes to be measured independently of the beam balance, thus allowing results to be cross-checked for accuracy. injector flow rates, the fuel pressure was varied using a rising rate regulator with MAP reference. This ensured the pressure across the injector remained constant for all MAP conditions, allowing fuel flow determination from the calibration. Further detail concerning the beam balance and fuel calibration is given in Appendix J.4. The mass flow rate of air was not measured directly but instead computed using the known fuel mass flow rate and the AFR. The overall AFR was measured using several different methods, as outlined in Appendix M.4. lambda () value computed. the choked condition.
Beam balance

By using several

methods, the measurements could be cross-checked for accuracy and an average Airflow rates calculated from the AFR were also validated using the known airflow rate through the restriction while operating in

Sliding moment arm

Mass measurement

Counterweight Framework Fuel storage tank

Figure 4.5: Schematic layout of the fuel mass measurement system, calibrated to provide fuel consumption rates.

87

The known airflow rate was determined theoretically (Appendix D.1) and experimentally using the flow bench rig (Appendix H.5.4). It is noted that even though the flow bench experiments were steady state, the pulsing through the flow restriction in the TC mode is negligible in the choked condition (Chapter 6.5.2).

4.3.2 Exhaust Emissions Analyzer

Throughout experimentation, two Auto-Diagnostic ADS-9000 Super Five Gas Analyzers were used to measure dry emission concentrations, specifically hydrocarbons (HC), carbon monoxide (CO), carbon dioxide (CO2), oxygen (O2) and oxides of nitrogen (NOX). These exhaust concentrations were used to calculate AFR by performing a chemical balance of the reactants and products. Sampling each cylinders exhaust gas allowed individual cylinder exhaust products to be analyzed, allowing cylinder-to-cylinder variations to be computed. Appendix M.2 outlines the exhaust manifold sampling locations together with the method used to obtain corrected emissions and calculated AFR. In order to ensure consistent and repeatable results throughout experimentation, both exhaust emissions analyzers were calibrated at regular intervals using calibration span gas to ensure accurate and consistent operation. The sensitivity of the ADS-9000 to NO had previously been determined with a linear response to the calibration gas [62]. Appendix M.3. However, non-dispersive infra-red (NDIR) HC concentrations for gasoline required correction, with further detail outlined in Automated electronic zeroing was performed during machine warm-up and after long idle periods to account for varying atmospheric conditions within the laboratory.

4.3.3 Blow-by Measurement

Volumetric blow-by measurement was made possible by venting all blow-by gas through a calibrated rotameter. The rotameter was positioned at the front of the dynamometer control panel, ensuring the blow-by rate could be visually monitored in real time whilst controlling the TP and hence MAP. Monitoring blow-by rates during engine run-in (Appendix L.1) was vital in maintaining engine reliability, as

88

the piston seal could be gauged to ensure adequate bed in and sealing prior to applying increasing load. If sudden fluctuations in blow-by flow were observed, development costs and engine downtime could also be minimized as the engine could be rapidly shutdown to limit further damage. This was important as the engine design underwent considerable piston development as described in Chapter 5.4. Further detail behind implementing the blow-by measurement system into both wet and dry sump lubrication systems is given in Appendix J.5.

4.3.4 Cylinder Pressure Measurement

An air-cooled Kistler 601-B1 piezoelectric pressure transducer connected to a Kistler 504E dual mode charge amplifier was used to measure in-cylinder pressure in the first cylinder. The transducer was calibrated statically using a dead weight tester [131], with the calibration curve given in Appendix J.7. The pressure transducer was mounted in the cylinder head, with the diaphragm exposed to the combustion chamber. Flush mounting the pressure transducer Hence, the was initially dismissed due to access constraints in the compact four valve, pent roof combustion chamber and the limited transducer body length. pressure transducer was initially non-flush mounted according to Kistler remote mounting instructions [122]. However, excessive resonance within the connecting passage and clearance chamber produced signal oscillations (Figure J.11), which were eventually eliminated by flush mounting the pressure transducer. Flush mounting was achieved by extending the length of the pressure transducer assembly using a National Instruments PC-401A series connector, with the transducer assembly depicted in Figure J.12. Furthermore, a room temperature vulcanizing (RTV) silicon was applied to the transducer diaphragm to minimize the effects of thermal shock [38, 133, 181]. The reader is referred to Appendix J.7 for further details.

4.3.5 Data Acquisition and Pre-Processing

Cylinder pressure analysis in internal combustion engines involves measuring and recording large quantities of data at high sampling frequencies due to the changing cycle pressures and cylinder swept volumes. To monitor these affects,

89

in-cylinder pressure and corresponding crank angle were recorded through two channels of the data acquisition (DAQ) system. The DAQ system incorporated hardware and software supplied by Eagle Technology. This involved a PC-30D analogue to digital converter (ADC) together with WaveView [224] interface software. Due to the varying engine speeds over the range of experimental data points, the DAQ frequency and sampling values were fixed to minimize the changes and complexity throughout the experiments. A maximum sampling frequency of 200 kHz was chosen for the two channels, limited by the DAQ system capabilities. This corresponded to a minimum sampling resolution of 0.6 CA at maximum engine speed, which had previously been proven to be adequate for qualitative and quantitative combustion analysis [92]. Sampling at lower engine speeds improved the resolution, with all data interpolated to 0.5 CA increments in the post processing (Chapter 9.3). This was required for the combustion analysis, which was performed using an Experimental Combustion Burn Rate Analysis (ECoBRA) program [91]. The sample size was fixed at 100,000, which enabled a minimum of 25 cycles to be recorded at 3,000 rev/min. Increasing the engine speed increased the number of recorded cycles, allowing data trimming in the post processing analysis. Post processing detail is given in Chapter 9.3, which further outlines the resampling techniques. To enable PC-based real time data pre-processing, raw analog signals were converted into digital form using a PC-30D ADC. The board has a 20-volt digitization range giving a resolution of 0.004888 volt/count. This translates to a maximum resolution of 4 kPa over a 16 MPa pressure range for the pressure transducer constants defined in Appendix J.7. allows multi-channel operations at high Direct memory access circuitry frequencies. Further sampling

specifications of the ADC board are given in Table 4.2. The interface between the computers ADC board and the BNC cables transmitting the raw signals required an electrical interface junction. WaveView enabled the visualization of raw pressure and crank angle signals allowing pre-processing analysis during development. Output parameters that describe combustion, including peak cylinder pressure and the rate of pressure

90

rise were evaluated in WaveViews oscilloscope mode, which allowed real time data streaming as shown in Figure 4.6. More complex combustion outputs were evaluated in the post processing analysis (Chapter 9).

Table 4.2: PC-30D analog to digital converter specifications. Analog Input No. of input channels Resolution Total system accuracy Differential non-linearity Input ranges Data acquisition rate Internal clock External clock External trigger Block scan mode Interrupts Direct memory access 16 single ended 12 bit, 1 in 4096 1 LSB LSB maximum -5 v to +5 v, 0 to 10 v 5 mHz to 200 kHz A/D Clock 2MHz, crystal controlled TTL compatible TTL compatible, enable or disable conversions Up to 256 channels per block PC Interface 16 single ended Duel channel 16-bit jumper schedule

5 4 Volts (V) 3 2 1 0 0.03

Crank Angle Encoding Pressure Signal

0.04

0.05 0.06 Time (s)

0.07

0.08

60 CA

Figure 4.6: Motoring raw cylinder pressure with crankshaft angle encoding from WaveView. Time versus voltage, NA - PFI, 6000 rev/min, WOT, CR = 10.

91

4.4 Experiments
Experiments were completed using Shell Optimax as the base fuel, a commercial high octane (98-RON) pump gasoline with specifications given in Appendix C.1. The use of a high quality fuel in experiments also followed present trends which suggest increases in gasoline quality, as ethanol and other high octane constituents become more widespread. The high fuel quality also complied with the regulated fuel permitted in FSAE competition. Based on the research foundations described in Chapters 1 and 2 and consistent with the research objectives, the hypotheses to be tested are outlined below.

4.4.1 Experimental Objectives

Experimentally explore the effects of CR variations on performance, efficiency and emissions for a small engine (~0.5 L) across engine speed and MAP domains. Compare carburetion and PFI fuel delivery systems in a small engine. Experimentally explore the effects of varying NA, SC and TC modes on performance, efficiency and emissions for a small engine (~0.5 L) across engine speed and MAP domains.

4.4.2 Experimental Results Analysis

Develop an understanding of the air consumption effects of an odd firing engine fitted with and without an intake flow restriction. Compare experimental trends and peak results to larger bore engines over engine speed, MAP and CR domains. Calculate combustion parameters for a small high speed engine, comparing the effects of engine speed, MAP, CR and fuel delivery variations. Interpret the results in terms of the performance potential and operating limits of small engines for the purpose of replacing larger engines found in compact sized regular passenger vehicles. Define the extent or the swept capacity reduction ratio to which larger engines can be reliably downsized while still maintaining equal power. Determine the factors limiting performance for small engines.

92

4.5 Test Modes

For the experiments (Section 4.4), four test modes were defined by variations in the induction system with drawing assemblies displayed in Appendix P. These modes included: (A) NA with carburetion (B) NA with PFI (C) SC with PFI (D) TC with PFI Variations in the induction system included two different fuel delivery systems, involving carburetion and PFI (Table 4.3), as the majority of small engines still operate with the outdated carburetion system. A modern OEM carburetor system from a small engine with similar bore and stroke dimensions [215] was adapted to suit the test engine. The PFI fuel system was designed using standard components from a variety of manufacturers to suit the engine configuration.
Table 4.3: Fuel delivery specifications for the carbureted and PFI systems. Carburetion
Type Brand Part Number Quantity Fuel Pressure Fuel Entry Downdraught, constant vacuum Mikuni BDSR36 1 carburetor / cylinder Gravity fed High

PFI
Sequential Injection, 4 orifice nozzle Bosch 0-280-156-124 1 injector / cylinder 500 kPa across injector High or low

The use of differing fuel delivery systems also necessitated alternative intake manifolds between the test modes. The OEM carbureted system was fitted with no plenum chamber with both trumpets open to atmosphere, as depicted in the upper diagram of Figure 4.7. Consequently, it would have been difficult to implement the flow restriction or boosted operation in this mode, which was unnecessary and not part of experiments (Section 4.4). An intake manifold was designed and developed for the PFI fuel delivery system (Chapter 3.7 and 5.5). Although the intake manifold geometry was optimized for TC operation (Chapter 6.5.2), the geometry was applied over all PFI modes, with the intake plenum volume, intake runner length and intake restriction fixed at

93

constant values.

To accommodate the compressor into the intake system for

boosted operation, the NA manifold layout was altered with modifications to the duct entry, as displayed pictorially in Figures 4.8 and 4.9. The PFI intake manifold also enabled the 20 mm diameter intake flow restriction to be installed, which was mandated for FSAE competition. Further intake system details are given in Table 4.4, with associated NA views shown in Figure 4.7.

Figure 4.7: CAD and actual NA modes with fuel delivery variation. (Upper): Mode A - NA with Carburetion (single carburetor shown in CAE model). (Lower): Mode B - NA with PFI.

94

Table 4.4: Intake manifold geometry details for all test modes to suit the varying fuel delivery systems. Mode (A)
Fuel Delivery Intake Manifold Design (Appendix P) FSAE Flow Restriction Intake Plenum Volume (L) Primary Intake Length (mm) Carburetion Drawing 0012A NO N/A 200 Drawing 0012B YES 4.5 400

(B) & (C)

PFI

(D)
Drawing 0012C YES 4.5 400

For the SC mode, boost was achieved by coupling an externally driven Roots type supercharger to the NA - PFI mode. A maximum boost level of half an atmosphere was desired to examine potential knock problems and ensure engine reliability at the high CR. Coupling of the variable speed positive displacement blower to the engine was trivial, and was achieved via installing a flexible duct from the outlet of the supercharger to the PFI intake assembly, as shown in Figure 4.8. To accommodate the supercharger, the air filter was repositioned to the inlet side of the compressor. Thus, MAP was controlled by varying the throttle position, as the throttle body/restrictor system was positioned between the supercharger and the intake manifold. Driving a supercharger compressor with the crankshaft would have been difficult due to the crankcase design, which featured an integral clutch and transmission. Thus, the supercharging objectives were achieved with a separately driven compressor for simplicity reasons as no supercharger could easily be adapted into the engine design. Engine air supply was delivered on an as required basis. This method was chosen over an external air accumulator tank as it was more representative of supercharging effects. These effects included higher intake temperatures which enabled realistic operating limits to be quantified. Humidity effects were also more representative as the water vapor was less likely to condense into liquid form due to the higher intake temperatures. Changes in humidity had previously been shown to affect TH and knock limits together with NOX emissions, as the inert water vapors enthalpy reduces the charge and combustion temperatures [92, 220].

95

Variable speed drive

Electric motor

Throttle

Outlet duct

Engine (PFI mode)

Roots type supercharger

Air filter

Figure 4.8: Externally driven Roots type supercharger assembly. Mode C - SC with PFI.

One drawback of the external supercharging system is the unquantified effect on

TH. Consequently, all TH values presented in this thesis for the SC mode exclude
the electrical energy required to drive the Roots compressor. The electrical energy is not included as the energy discrepancies between external electrical energy and engine power are too dissimilar. Engine parasitic losses associated with driving the compressor (belt drive) are also expected to be negligible [66, 159] in comparison to the engine brake power output due to the fairly constant blower efficiencies. For the TC - PFI mode, a Garrett GT-12 turbocharger was utilized as displayed in Figure 4.9 and outlined in Chapter 6. Although the air consumption in the TC PFI mode was limited by the flow restriction, the desired TC mode operating objectives did not vary significantly when compared to typical TC applications, as found in passenger vehicles. The objective in TC applications is to always improve low speed performance as boost at high engine speeds is relatively easy to achieve. However, achieving boost at as low a speed as possible is imperative in improving vehicle drivability. Hence, although the flow restriction limited the maximum airflow and affected the engine speed at which peak boost occurred, it did not influence the rate at which boost could be delivered. As previously outlined, the objective for FSAE competition was to maintain a choked restricted inlet at the lowest possible engine speed to broaden and

96

increase the delivered power, which improved vehicle drivability [13-16]. Hence, pressure ratios at WOT were dictated by airflow limitations and corresponding engine speeds. Maximum boost levels were dictated by turbocharger performance maps and were expected to reach near 300 kPa MAP, which was needed to produce sonic flow at mid-range speeds. Once operating in the choked restricted condition, decreasing MAP levels matched the increasing engine speeds to maintain a constant airflow rate. These operating conditions are specific to flow restricted configurations like FSAE and differ from typical turbocharger applications, which usually maintain constant MAP once desired boost levels have been achieved due to knock or mechanical limitations. Table 4.5 highlights the maximum desired intake conditions across all test modes. Maximum operating conditions at WOT for the NA modes were limited to atmospheric conditions, with the aid of intake resonant tuning to improve air consumption. For the PFI modes, air consumption effects due to the flow restriction are assumed to be negligible in the NA and SC modes as engine air consumption rates are below choking limits. Hence, the flow restriction is only assumed to affect performance in the TC mode due to the desired operating conditions and initial project performance objectives associated with FSAE.

Figure 4.9: Mode D - TC with PFI, CAD model and final developed version.

97

Table 4.5: Defining the test modes desired maximum operating conditions. Mode
(A) NA - Carburetion (B) NA - PFI (C) SC - PFI (D) TC - PFI

Desired Maximum MAP (kPa) at WOT

100 100 150 Maximum achievable (flow restriction limited)

4.6 Test Methodology

4.6.1 Test Matrix
The experimental test matrix employed to test the hypotheses is shown in Tables 4.6 and 4.7.
Table 4.6: Engine speed versus MAP test matrix at the CR closest to the HUCR. MAP / Speed
<=50 kPa 60 kPa 70 kPa 80 kPa 90 kPa 100 kPa 120 kPa 140 kPa 150 kPa 160 kPa 180 kPa 200 kPa 220 kPa 240 kPa >=260 kPa 4,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

5,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

6,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

7,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

8,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

9,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

10,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

SC-PFI TC-PFI*

SC-PFI TC-PFI*

SC-PFI TC-PFI

SC-PFI TC-PFI

SC-PFI TC-PFI

SC-PFI TC-PFI

SC-PFI TC-PFI

TC-PFI*

TC-PFI*

TC-PFI*

TC-PFI*

TC-PFI*

TC-PFI*

TC-PFI*

* MAP dependent on achieving boost, limited by turbocharger or flow restriction

Table 4.7: Engine speed versus CR test matrix at the MAP defined in Table 4.6. CR / Speed
4,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

5,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

6,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

7,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

8,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

9,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

10,000 rev/min
NA-CARBS NA-PFI SC-PFI TC-PFI

9-13

98

The strategy employed across the test matrix was to standardize all settings except the variable being tested. This allowed accurate determination not possible in a multi-variable change approach such as Taguchi methods which can reduce the number of test points but offer less precision [153]. Standard settings included the valve motion properties (valve timing) together with the employed tuning strategy. Exhaust manifold geometry was also held constant over all The modes with exhaust duct changes allowing turbocharger implementation. MAP parameters included: Mode variations CR variations The maximum CR was limited to 13, which was the highest achievable with a flat top piston in the pent roof combustion chamber. The decision to use a flat top piston was based on manufacturability, with simple machining processes allowing reductions in CR. 5.4.3). The effects of DI, EGR and intake charge cooling were not investigated during experiments due to their added complexity. improve. If these strategies were to be implemented, performance, efficiency and/or emissions results would only Eliminating these variables enabled the target parameters such as engine speed, MAP and CR to be more thoroughly examined across the test modes in the time available. Hence, variations ranging from 9-13 were made possible through piston crown modifications to a set of custom forged pistons (Chapter

test variables requiring engine modification across the range of engine speed and

4.6.2 Test Sequence

The testing sequence was chosen to accommodate the development of the new engine together with testing the research hypothesis through the test matrix. Initially, the engine completed a run-in (Appendix L) to ensure adequate bed-in of all components to minimize wear rates, improve reliability and stabilize performance [112]. The run-in process was followed whenever the engine was reassembled.

99

Testing commenced at the highest achievable CR of 13. This allowed the CR to be reduced after completing experiments for the CR investigation (Table 4.7), as piston quantities were limited. When knock was encountered, the first stage of knock control relied on traditional tuning methods involving varying degrees of spark retard and/or fuel enrichment [42], albeit with the penalty of increased fuel consumption. If knock control could not be achieved through the ECU using these tuning strategies or further piston development (Chapter 5.4), experiments across subsequent modes were abandoned at the test CR. The CR was then decreased in accordance with the knock severity, with modes subsequently retested at the reduced CR to find the limits of operation. The knock (KL) and damage limit (DL) previously published by Rothe et al. [192], were used to determine if CR reductions were needed to ensure engine reliability. Table 4.8 defines the knock amplitude (KA) used to quantify limits [192], with the KA defined as the zero to peak pressure of the high pass filtered cylinder pressure.
Table 4.8: Knock and damage limits previously defined by Rothe et al. [192]. Defining Knock and Damage Limits Knock Amplitude (KA) Knock Limit (KL) Damage Limit (DL) Zero to peak pressure of the filtered cylinder pressure 1% of cycles with KA > 4 bar 1% of cycles with KA > 20 bar

Mode investigation commenced with NA operation as this had major advantages in developing the engine design under reduced combustion loading, which minimized development time and costs by reducing the likelihood of engine failure. Lighter loading also allowed a base calibration to be set, with increased error safety margins when compared to boosted modes as the limits of operation were explored. NA performance could also be gauged and compared to the literature as no information could be found on highly turbocharging engines of this scale or configuration. This a realizable consequence when downsizing engines found in small passenger vehicles [12]. NA performance also allowed a repeatable baseline to be established to evaluate the engine design for future testing.

100

Carburetion was chosen for initial test mode investigation as this significantly reduced the complexity during the initial calibration phase. Ignition timing was adjusted with the ECU while fuel mixtures were self-controlled after initial carburetor jetting. This was extremely helpful in troubleshooting other areas, as the system was independent of all electronics and sensors and thus focus could be placed on developing the internals of the new engine design. Once engine reliability could be ensured, calibration was performed in accordance with the tuning strategy (Section 4.6.4). After successful data collection, sequential PFI was fitted followed by boosted operation to find limits of operation. Boosted operation commenced in the SC mode as the highest useful compression ratio (HUCR) was expected to decrease for the TC mode due to the higher boost levels. Supercharging also enabled parametric constraints to be established to define the engines limitations under mild boost levels with no exhaust influence. The SC data also provided valuable insight into the engine modifications needed to sustain high pressure ratio turbocharging [16, 248]. The calibration complexity was also reduced in the SC mode, allowing a base fuel and ignition table to be established for boosted MAP values. Establishing some calibration tables in more forgiving surroundings also reduced the likelihood of engine failure in subsequent turbocharger testing.

4.6.3 Test Procedure

With the dynamometer direct coupled to the transmission due to the engine design (Section 4.2.2), engine starting and shutting down was completed with the clutch manually engaged while in final gear. This process together with testing was completed by two operators. Decoupling the engine from the transmission made starting and shutting down less troublesome, with the dynamometer inertia minimizing the likelihood of engine stall during transmission engagement. Clutch disengagement during shutdown also minimized the shock loading through the system as the dynamometer inertia did not drive the engine crankshaft through the transmission. Variations in water coolant temperature can have a significant effect on emissions,

MECH and in-cylinder heat transfer characteristics.

Therefore, day to day

101

operation involved engine warm-up to 85-90C, with this time minimized through an electric heater integrated into the test rig cooling system. Although the test rig cooling system enabled the water coolant temperature to be regulated prior to engine operation, sufficient time was given for the in-cylinder surfaces to reach operating temperature prior to any data collection. load while monitoring the torque output. This also allowed the lubricating oil to reach operating temperature as the engine was operated at light Consequently, the oil temperature exceeded the regulated water jacket temperature prior to any data collection. After warm-up, engine calibration was performed in accordance with the tuning strategy as outlined in Section 4.6.4, starting at low load and speed conditions. The base table was initially calibrated with load increases due to the dynamometer constant speed limitations, followed by speed increases. This process minimized the likelihood of engine failure as knock limited regions were gradually explored at high CR, while avoiding high knock intensities.

4.6.4 Tuning Strategy

The employed tuning strategy was dictated by the engines intended application, which was initially for FSAE competition. This resulted in a tuning strategy with variation as Formula rules did not govern specific emissions. Consequently, power, efficiency or knock control was targeted for varying MAP, as stoichiometric AFR was unnecessary for efficient TWC operation. At light and medium loads, lean and stoichiometric mixtures were used to improve efficiency and reduce fuel consumption. Richer mixtures were used at heavier load conditions associated with brake performance, with the effects on NA BMEP displayed in Appendix M.6. This improved brake output and provided component protection due to the reduced combustion temperatures, which further enhanced engine reliability. Fuel injection timing was varied to achieve maximum brake torque (MBT injection timing), using HC emissions as an indicator of mixing along with CO and CO2 as indicators of combustion efficiency (defined as the useful energy released throughout this thesis). For assessing cylinder-to-cylinder air-fuel distributions, the exhaust gas composition was measured in each cylinders exhaust header using two NDIR five-gas exhaust analyzers, as described in Appendix M.

102

The ignition tuning strategy involved finding the minimum spark advance for maximum brake torque (MBT-ST) or in the case when ignition timing was knock limited, the knock limited spark timing (KL-ST). The employed tuning strategy has significant relevance to FSAE applications, where one of five dynamic events features fuel economy performance. Hence, the tuning strategy allowed the engine to conserve fuel when the driver required little power, with significant power increases upon driver demand. Details of the tuning strategy are given in Table 4.9.
Table 4.9: Tuning strategy employed over the varying load conditions. LOAD CONDITION Light
Corresponding BMEP (kPa) Corresponding MAP estimate (kPa) Targeted Spark Timing PFI Injection Timing < 300 < 60 1.1-1.2 MBT MBT

Medium
300 - 600 70 - 80 1 MBT MBT

Heavy
> 600 > 80 0.9 MBT or KL MBT

4.7 Summary
This chapter describes the engine test rig which was commissioned and calibrated by the author and used to complete the engine experiments. The instrumentation and data acquisition that was needed to accurately measure and record data is also outlined. Furthermore, this chapter clearly identifies the experimental objectives which led to the decision to test various engine test modes for experiments. included. A brief description of the motivation behind the testing sequence is given, as experimental data collection had to evolve with engine development. The test procedure, which ensured experiments were accurate and repeatable, is also outlined and underpinned the engine development found in the next chapter. The methodology behind experimental testing is also explained, with the test matrix

103

104

CHAPTER 5
P O W E R E D B Y W A T T A R D

Engine Development

5.1 Introduction
This chapter outlines the development required for test engine operation following the initial design, manufacture and assembly described in Chapter 3. Considerable engine development was completed over an 18 month period, the majority on the constant speed engine dynamometer test bed. The engine was also installed into successive Melbourne University FSAE vehicles (Figure 5.1) for transient development, evaluation, tuning and finally for competition. As countless steps were involved in the engine development, this chapter mainly describes novel and interesting aspects surrounding highly turbocharging this inline twin configuration. Further detail surrounding transmission and engine subsystem development including cooling, lubrication and in-vehicle development is given in Appendix I. The majority of engine development was completed to improve reliability and to solve the numerous problems associated with running high levels of boost. Hence, engine development continued concurrently with experiments until the later stages of testing. However, relatively few problems were encountered in NA operation after engine commissioning. In this chapter, the ways in which development was accelerated through modeling tools and FEM analysis is outlined. This has enhanced understanding which has enabled improved performance and reliability at reduced costs and lead times.

105

Figure 5.1: Test engine fitted into Melbourne University FSAE vehicles for further engine development, testing and validation. (Upper): 2003 vehicle. (Lower): 2004 vehicle.

As the thesis describes the engine reliability development throughout, Table 5.1 serves as a thesis location directory for the reader to gather information surrounding the specific problem that needed to be overcome. Table 5.1 also distinguishes where the engine development took place (dynamometer test rig or while the engine was fitted to a FSAE vehicle) in chronological order to outline the sequence in which problems were solved.

106

Table 5.1: Engine reliability development sequence in chronological order associated with NA and boosted engine testing. Development Problem to Overcome Thesis Location

2003 (Apr-Nov) Dynamometer Testing

Electrical system Calibration Exhaust manifold ECU sensor and signal errors Ignition coil failures NA fixed speed tuning Fatigue and thermal cracking Manufacturing method High bore distortion Combustion gas leakage Thrust bearing failures Timing chain resonance Excessive vertical vibration Balance weight failure Piston skirt and liner contact Piston land / ring failures due to knock Water cooling Oil control Section 5.3 Appendix 4.6.4 Chapter 4.2.6 Section 5.6 Section 5.2

Gasketless interface

Camshaft and timing chain Engine balance and vibration Piston assembly Turbocharger

Section 5.8 Chapter 5.7 Chapter 5.4 Chapter 9.6 Chapter 6.4

2003 (Dec) NA Vehicle Testing

Injector location Lubrication Water cooling Calibration Poor throttle response Wet sump oil surge Engine overheating NA transient tuning Section 5.5.2 Appendix I.4 Appendix I.5 Appendix 5.5.2

2004 (Jan-Oct) Dynamometer Testing

Water cooling Lubrication Calibration Turbocharger Piston assembly Electrical system Exhaust manifold Intake manifold Clutch Transmission Flow circuitry and pump cavitation Dry sump blow-by and oil scavenge TC fixed speed tuning Oil control Boost control Piston land / ring failures due to knock Ignition coil failures Fatigue and thermal cracking Leakage and fatigue cracking Oil cooling, overheating and slippage Gear disengagement Appendix I.5 Appendix I.4 Appendix 4.6.4 Chapter 6.4 Chapter 5.4 Chapter 9.6 Section 5.3 Section 5.6 Section 5.5 Appendix I.2 Appendix I.3

2004 (Dec) TC Vehicle Testing

Calibration TC transient tuning Appendix 4.6.4

107

5.2 Barrel and Liner Gasketless Interface

5.2.1 Manufacturing and Assembly
Due to the short time scale of the engine program and budget limitations, there was little choice but to manufacture the barrel and liners concurrently. This was expected to be problematic due to the unknown bore distortion effect from shrink fitting the liner into the barrel. However, this was expected to be minimal due to the slight interference between both parts. plates were installed prior to honing. As both items were machined from billet, drawings were issued with tight tolerances on critical dimensions to ensure the correct interference when assembled. To minimize time loss, the barrel and liners were finished machined with a +50 m bore honing allowance to later match the custom forged pistons, which as yet had not been delivered. This method of manufacture proved to be problematic and expensive. Although both barrel and liner bores were concentric and consistent in diameter along the length of the bore prior to assembly, significant bore distortion occurred when the liners were shrink fitted. the top being 25 m larger than the bottom. This variation was a result of inconsistent hoop stresses generated as a result of the varying barrel wall thickness, with the top allowing more radial deflection than the bottom. This is confirmed in Figure 5.2, which displays the resultant stresses and deflections generated via the shrink fit, with the FEM analysis performed after assembly. The analysis clearly shows that all displacement occurs radially, with higher stress bands in the liner. This occurs as the barrel allows less deflection due to the increased wall section, which is reinforced in the middle and lower portions. However, the upper unsupported water jacket region allows both components to radially deflect, with similar local stress bands. Although the bore had remained concentric, it was no longer consistent in diameter, with It was assumed from earlier FEM analysis (Chapter 3.6) that the majority of distortion would occur when the torque

108

STRESS
MPa 60 50 40 30 20 10 0

DISPLACEMENT
m 50 40 30 20 10 0

Figure 5.2: FEM model analysis simulating the H7-p6 shrink fit between the barrel and liner. (Left): Von Mises stress distribution. (Right): Resultant displacement.

The bore diameter inconsistencies were further magnified after torque plate installation, as confirmed by the FEM analysis in Figure 5.3. The analysis showed that the torque plate loading effects are dominant, but highlighted that the shrink fit effects cannot be ignored. concentricity. This bore inconsistency, increased the honing difficulty, with the finished bore size varying by 25 m in diameter and This was not ideal, but with approaching deadlines, engine assembly and commissioning continued.

STRESS
MPa 70 60 50 40 30 20 10 0

DISPLACEMENT
m 80

60

40

20

Figure 5.3: FEM model analysis simulating both the shrink fit and torque plate effects. (Left): Von Mises stress distribution. (Right): Resultant displacement.

109

5.2.2 Initial Testing

From the onset, piston blow-by rates were high with no significant reduction after initial engine run-in and several hours of operation. Blow-by rates significantly increased as load increased, thus testing stopped for disassembly and inspection. Figure 5.4 displays some findings of concern, with close inspection revealing abnormal carbon deposits on the cylinder head and liner mating surfaces. This could only be explained by combustion gas leaking into this area as a result of the failed interface. The high blow-by rates observed during testing were a result of leakage past the top piston ring, which is also highlighted in Figure 5.4. This problem was made worse as the first iteration piston design featured only one compression ring. The high blow-by rates could be attributed to bore dimensional irregularities associated with the manufacturing process previously described. However, the failed interface could not be explained at the time.
Interface Leakage

Piston Ring Leakage

Figure 5.4: Failed gasketless interface with clear signs of leakage between the cylinder head and liner mating surfaces. High piston ring leakage is also evident.

5.2.3 Failure Analysis and Rectification

Two major problems were evident after initial development. The first involved how to manufacture the assembly within specified tolerances. The other involved the amount of interference needed to produce the required face pressure to ensure a robust interface seal.

110

The manufacturing procedure was altered to accommodate the deflection from the shrink fit and torque plate loading. This included an extra 0.5 mm allowance on the liner inside bore, which had only a minor effect on shrink fit deflection. This extra allowance allowed the final bore machining to be completed after shrink fitting, with the addition of torque plate loading. The following manufacturing process was devised which ensured the bores remained concentric and consistent in diameter when the cylinder head was installed. 1. The liners outside diameter (OD) was machined to match the barrel inside diameter (ID) allowing for the associated shrink fit. 2. A 0.5 mm extra allowance was left on the ID of the liners and the interface sealing face. 3. Liners were installed into the barrel by shrink fitting. 4. The liner deck surface was machined to specified tolerances above the barrel deck surface to achieve the required interference. 5. The bottom fixture was installed followed by the torque plate. 6. The liners ID bore was machined and then honed to match the pistons. Following the manufacturing process development, the interface failure was investigated, with findings suggesting that the interface failure was caused by insufficient face pressure between both surfaces. The logged data confirmed that the measured in-cylinder pressures were well below the worse case design pressure, with additional error margin in the safety factor assumption as outlined in the design (Chapter 3.6.1). However, original calculations assumed the cylinder head was a rigid solid body with no deflection. This assumption was proven to be incorrect by the physical load deflection tests performed for the torque plate design (Chapter 3.6.2) and later confirmed by the FEM analysis of Figure 5.5. The deflection analysis of Figure 5.5 highlights the 20 m variation in the interface region. This value is similar in magnitude to the experimental test value (Figure 3.20) and is a result of the cylinder head modifications (Chapter 3.4.2), which allow more elastic deformation in the outer regions. The deflection corresponds to the interface failure region, due to the low cylinder head face stiffness. However, successful sealing was apparent over the majority of the interface.

111

STRESS

Section A-A
MPa 0 20 40 60 80

A
DISPLACEMENT

Section A-A
m 0 10 20 30 40

A Figure 5.5: FEM model analysis simulating the gasketless interface effects on the cylinder head. (Upper): Von Mises stress distribution. (Lower): Vertical displacement.

The cylinder head stress analysis of Figure 5.5 shows that the loading due to the interface produces only elastic deformation. The simulated peak stresses of 40 MPa are well below the elastic limit for the material. Physical measurement of the cylinder head after disassembly also confirmed that there were no changes in surface flatness after initial testing. Thus, the interface leakage was resolved by increasing the interference to 80-90 m in order to accommodate the cylinder head deflection. Further FEM analysis confirmed that the material yield strength was not exceeded.

112

5.2.4 Interface Performance

The integrity of the gasketless interface was tested over the extensive experimental program, including steady state dynamometer testing combined with vehicle transient testing. The design proved very successful, performing faultlessly with no sign of any interface combustion gas or fluid leakage after initial development. Furthermore, the piston ring seal and resulting blow-by rates for the conventional three ring piston design (Section 5.4) in both NA and boosted configurations were also comparable to previous published data for closed deck block designs [163, 244, 253]. This suggests that the previously documented bore distortion problems that occur in highly boosted engines and the concomitant structural integrity reductions of the open deck design can be overcome with the gasketless design [17]. The sealing integrity of the interface is further demonstrated by the extreme conditions it withstood during engine testing. Appendix I.5). This included test conditions featuring sustained high knock intensities and engine overheating (~130C: Sealing integrity was maintained for peak cylinder pressures exceeding 10 MPa, with combined knock amplitudes of 6 MPa (Chapter 9.6). The integrity of the interface is further highlighted by the piston failures (Figure 9.17) and elevated combustion temperatures (Chapter 9.5) it withstood without failure. Further experiments also found that raw engine-out HC emissions were significantly lower than those of similar engines operating at comparable conditions (Chapter 7.6.1). This was caused by two factors involving more uniform cylinder head and block surface temperatures (Chapter 3.6.3) and the absence of any crevice volume near the bore periphery that exists with conventional head gaskets. This amount of crevice volume is largely dependent on the alignment between the bore, gasket and combustion chamber and is eliminated in the gasketless design. Since the interference and resultant face pressure between the liner and cylinder head produces only elastic deformation, no adverse flatness effects on the mating surfaces of the assembly were recorded and no long term effects are anticipated. This is despite the frequent removal of the cylinder head during development for

113

inspection and engine overhaul purposes. However, some adverse effects of the design do exist. As a result of both surfaces being in direct contact, the potential for local imprinting and/or fretting to occur on both surfaces exists, with local imprinting occurring on the cylinder head surface during experiments due to the cast iron liner deck surface finish (Figures 9.17 and 9.18). This however, could be argued to be no worse than in a conventional head gasket design. The absence of any gasket material also eliminates weakspots that may cause gasket failure due to abnormal combustion (Chapter 9.6). or other failure. As proven from experiments, the gasketless interface has the potential to maintain combustion sealing until piston

5.3 Electrical System

5.3.1 Ignition System
A M&W Pro 12 CDI ignition system as described in Chapter 4.2.5 was used to provide spark plug energy. The system was originally rated at 100 mJ of ignition energy per firing sequence but after several COP failures, due to excessive ignition energy, the system was de-rated to 50 mJ by the manufacturer. This system was adequate for NA testing, however for boosted operation, combustion stability worsened with increasing levels of boost, which reduced brake performance. Furthermore, the poor combustion stability could not be improved with spark timing variation (Figure 9.14), with Chapter 9.5.1 giving further combustion insight. In an attempt to improve combustion stability, ignition components including the COPs and spark plugs were replaced with no improvements observed. Furthermore, various spark plug types and heat ranges were tested, including the surface discharge type, which only worsened combustion stability. Consequently, the ignition energy was returned to 100 mJ, with immediate improvements in combustion stability and hence power, as further outlined in Chapter 9.5.1. Unfortunately, this solution was not feasible due to the recurrence of COP failures at this energy as previously described. To remedy this, the COPs were replaced with ignition leads and large external Ferrite coils. However, this resulted in ECU errors as signals from sensors were

114

intermittently altered by the high tension (HT) noise emitted from the ignition leads. Efforts to eliminate this problem included shielding sensor cables and the use of Magnecor [144] ignition leads which suppressed radio frequency interference (RFI) and electromagnetic interference (EMI). The HT noise caused problems ranging from intermittent misfires which resulted in decreased maximum torque over much of the speed range. Hence, the COP system was reverted to 100 mJ of ignition energy, but the coils were replaced every 15 hours prior to coil burnout failure. Although this was not a permanent solution, the time and cost limitations of the project did not permit the manufacture of specific higher current COPs for the application, which would have provided a long term solution.

5.3.2 Reference and Synchronization Signals

Problems were encountered with ECU reference and synchronization signals due to the crankshaft torsional vibration and speed fluctuation during individual engine cycles (Figure 9.1). These problems were likely enhanced by the high engine speeds and the odd firing twin configuration. Initially, magnetic flux sensors giving an output voltage based on the change in magnetic flux (d/dT) were used. Sensors were mounted to rely on changes in radial distance and hence flux rather than end on due to problems previously found [93]. These problems involved instances where the sensor to pickup gap varied due to shaft end-float, which varied the signal strength. However, the radially mounted magnetic sensors also responded to crankshaft deflections, which tended to change the signal strength depending on the load. Hence, Hall effect sensors were implemented which gave a clear 0-5 volt square wave output signal to the ECU. For the crankshaft reference trigger signal, six equidistantly spaced slots were machined into the flywheel, with a limitation on the depth as the electrical power generator system was housed within the assembly. However, missing or duplicate signal problems were encountered, with the development process to overcome these shown pictorially in Figure 5.6. These problems were eventually solved by deepening the defined trigger edges to 3 mm, lengthening the pickup to avoid duplicate signals and changing the trigger (slot) geometry to a square edge to allow for the crankshaft end-float.

115

Initial Design Circular slot

Second Iteration Lengthened slot + deeper

Final Design Square edge slot

Figure 5.6: Crankshaft synchronization trigger edge (slot) development needed for high speed operation due to the higher torsional and cyclic velocity fluctuations associated with an odd firing twin configuration.

Due to the high speed of the camshaft and space restraints associated with the central position timing chain, a synchronization signal trigger wheel could not be mounted to the camshaft. Rather, a novel synchronization method was used, which relied on a cam lobe to be used as the signal source. Development of this concept began with a makeshift arrangement which simply rotated the camshaft at high speed in a lathe, hence proving the concept before applying it to the engine. With this setup, the synchronization trigger was only used by the ECU to define the cycle position (crankshaft revolution in the cycle), requiring only coarse camshaft position accuracy (1/2 distance between crank trigger = +/- 30 CA) for reliable ECU operation. It was feasible that some error could occur due to the changing distance between the camshaft lobe and the pickup sensor, resulting from varying valve spring loads or camshaft lobe wear. However, no problems were recorded with the setup until the cam lobes were repaired with a layer of non-magnetic nickel based material (Colmonoy5 [234]) after camshaft damage due to valve failure (Chapter 5.4.5). The problem was rectified by reducing the clearance between the sensor and camshaft lobe to enable pickup of the magnetic substrate. would give cost savings. The success of the synchronization setup demonstrates the potential of its application to production engines, which

116

5.4 Piston Assembly

5.4.1 Piston
Two piston designs, distinguished by the number of piston rings were used throughout development. Table 5.2 outlines the two (Type A) and three (Type B) ring piston versions, with slight alterations to the upper piston geometry required to accommodate the varying designs, mainly the top land thickness to suit the various piston ring packs (Section 5.4.2). The variations in the top land thickness were negligible in comparison to the total crevice volume, hence no changes in performance and only slight changes to HC emissions were recorded between the various top land geometries tested. Although the upper piston geometry required alteration, the skirt profile and pin location remained unchanged amongst piston designs. The changes to piston configurations and top land thicknesses were found necessary in the boosted development. Initially, NA engine commissioning and development was completed with piston type (A), with a minimum top land thickness of 4.0 mm. However, due to piston failures encountered due to knock (Figure 9.17), piston development centered on increasing the top land thickness and later increasing the number of sealing rings to improve piston longevity. Hence, as piston quantities were limited, the pistons used for later NA testing followed this trend. Hence NA and boosted experiments were completed with both piston designs, allowing effects to be documented.

Table 5.2: Varying piston designs used during development, differentiated by the number of compression sealing rings. Piston Design
(Type A) Two piston rings (One compression ring)

Appendix P, Drawing #
0112A 0112C 0112D 0112E

Top Land Thickness

4.0 - 5.8

(Type B) Three piston rings (Two compression rings)

0112B 0112F

4.0 - 4.5

117

An interesting finding from the piston ring location for piston type (B) is now discussed. To further improve piston reliability in the case of knock, the top land thickness was increased in piston type (B). To achieve the increase in top land thickness in the limited space, the three piece oil ring was relocated 3 mm down. Thus the oil ring had its entire section exposed to the piston pin end, with portions of the bottom oil rail not supported by a ring land (Appendix P). This design move, although radical proved successful with no effect on oil ring longevity or measured oil consumption in the following six months of testing, with no failures or oil ring rail bending recorded. To prevent cantilever bending on the oil rail ends, the rail was adequately positioned with side load from the oil expander ensuring no oil rail rotation. This finding concerning the oil ring location enables the piston pin to be relocated towards the piston crown. This would permit the use of a longer connecting rod which reduces piston side loading and hence friction or enables the block height to be reduced, thus reducing the engines mass and COG. Experimental tests found that the single compression ring of piston type (A) worked successfully throughout NA testing after initial piston development, which mainly involved determining adequate clearances (Section 5.4.6) required for operation. The performance variation between both piston designs in NA modes due to the higher friction losses of piston type (B) were also negligible, with experimental tests finding no clear performance differences. Generally, the entire piston ring pack contributes to ~25% of the total friction loss [98, 220], corresponding to ~2-3% of the IMEP. Hence, the performance effects due to a 25-40% reduction in piston ring friction due to the absence of a second compression ring were difficult to quantify, lying within the range of dynamometer error. Thus, the performance findings are not unexpected after friction breakdown analysis [10]. In the NA modes with the single compression ring design of piston type (A), blowby and oil consumption rates were comparable to piston type (B) at WOT, with the test engine data also similar to other engine experimental data [98, 163, 225, 253]. The suspected key reason for similar blow-by is the adequate control of bore distortion at high BMEP (cylinder pressure) due to the barrel design (Chapter

118

3.6). Furthermore, the use of a more pliable thin section piston ring enables the ring to more uniformly conform to the bore from the gas loading, which improves the seal. However, at light load, where the cylinder pressure and hence the gas loading on the ring is reduced, higher blow-by rates were recorded for the single compression ring design of piston type (A). The higher blow-by at part load is despite very tight side clearance in the piston ring groove (5-10 m) to minimize ring flutter effects at high speed. However, one detrimental effect of the thin ring is the reduced pressure on the bore due to the reduced cross section and hence stiffness. The higher blow-by at reduced loads for piston type (A) increases fuel consumption and crankcase emissions, making the single compression ring design impractical in current production vehicles. However, the design may perhaps be feasible in other applications due to the reduced costs and friction losses. NA developments also revealed that once the single compression ring of piston type (A) failed to adequately seal, large amounts of blow-by were recorded with plumes of white smoke emitted from the exhaust system. The white smoke was caused by excessive oil consumption, as the single ring not only had to seal combustion gas, but also had to provide adequate oil retention in the absence of the omitted second compression ring. Sealing problems were associated with damage to liner, piston and piston ring components together with ring flutter. Boosted operation revealed very different findings to those from NA operation. Many problems and failures were recorded with piston type (A), which did not permit adequate engine operation at mild levels of boost. associated with higher pressures and heat loads. These problems included increased blow-by and oil consumption due to the increased distortion These effects led to heavy knock and hence piston land failure as seen in Figure 9.17. Although significant development to reduce the propensity for knock was tried (Chapter 9.6.2), the defining change was the move to a more conventional piston design featuring two compression rings (Type B). This reduced blow-by and prevented oil entering the combustion chamber. Furthermore, the increased ring contact area with the bore significantly increased the path area for heat dissipation [95, 98] over the single compression ring of piston type (A). The combustion benefits are demonstrated by the changing cylinder pressures and reduced combustion variability displayed in Figure 5.7.

119

30 10 Raw Pressure (MPa) 8 6 4 2 0 -180

20 15 10 5 0 -5 220

-140

-100

-60

-20

20

60

100

140

180

Crank Angle (deg.) 30 10 Raw Pressure (MPa) 8 6 4 2 0 -180

20 15 10 5 0 -5 220

-140

-100

-60

-20

20

60

100

140

180

Crank Angle (deg.) 7 6 Raw Pressure (MPa) 5 4 3 2 1 0 -180 30

20 15 10 5 0 -5 220

-140

-100

-60

-20

20

60

100

140

180

Crank Angle (deg.)

Figure 5.7: Effect of varying the number of compression rings (heat flux and oil control effects) and fuel quality on combustion. TC - PFI, 7000 rev/min, 200 kPa MAP, CR = 10, 12 BTDC spark timing.

120

Knock Amplitude (MPa)

Piston Type (B) 2 Compression Rings (98-RON)

25

Knock Amplitude (MPa)

Piston Type (A) 1 Compression Ring (112-RON)

25

Knock Amplitude (MPa)

Piston Type (A) 1 Compression Ring (98-RON)

25

The varying pressures of Figure 5.7 illustrate that reduced temperatures from increased heat dissipation out of the end-gas and the concomitant oil control from the three ring piston design, featuring two compression rings, reduces knock. The middle diagram of Figure 5.7 also highlights that significantly increasing the fuel quality does not eliminate knock if the heat can not be dissipated and the oil can not be prevented from entering the combustion chamber. Hence, piston rings have been shown to be a very important engine design factor in controlling knock, and thus performance and efficiency in boosted engines. Further improvements are also possible through the ring number and location, as boosted engines generally carry over conventional designs from NA engines. Furthermore, results from experiments showed that slightly varying the ring land thickness for piston type (B) had little effect on reducing knock intensities, highlighting similar heat dissipation out of the top land through the top piston ring. However, the increased land thickness provides increased safety margin as the piston can withstand higher knock intensities before failure. It is also feasible that thicker section piston rings would increase the heat transfer to the cooling system due to the increased piston ring to bore contact, thus increasing the heat flux. This could allow increases in MAP and/or CR before exceeding the DL, but perhaps the ring section is limited by ring flutter at high engine speeds due to the increased mass. The potential also exists for more than two compression rings in an extended piston design after gudgeon pin relocation. Furthermore, an increase in the number of rings is also feasible after oil ring relocation, as was shown to work successfully in experiments. This would allow an unchanged gudgeon pin position, resulting in the same block deck height. MAP associated with highly TC engines. Either way, the piston ring developments have been shown to be critical in making pistons survive at high

5.4.2 Piston Ring Pack

Table 5.3 outlines the different piston ring packs trialed during experiments. The variations in piston ring dimensions, cylinder bore contact profiles and materials that were tested for certain reasons surrounding blow-by, oil control, performance and availability are now discussed.

121

Table 5.3: Different piston ring packs tried throughout development across NA and boosted modes. W: Width, P: Profile, M: Material. Ring Pack
A

Appendix P, Drawing #
0113A

Top Ring
W: 1.0 mm P: Positive taper M: SG cast iron W: 1.0 mm P: Barrel M: Chrome plated steel W: 1.2 mm P: Positive taper M: Plain cast iron W: 1.0 mm P: Barrel M: Chrome plated steel

Second Ring

Oil Ring

0113C

N/A

0113D

W: 2.8 mm 3 piece rail/expander

0113B

W: 1.2 mm P: Positive taper M: Plain cast iron

Single compression ring development commenced with the as-designed positive scraper SG cast iron ring as described in Chapter 3.5.3, which was specifically manufactured for the application. However, after limited supply due to bore scuffing and ring failure (Section 5.4), the top compression ring was replaced with an OEM Wiseco version, which was a chrome plated steel ring with a barrel profile, manufactured for a conventional three ring design. This ring was later replaced by a thicker 1.2 mm positive scraper ring to continue testing due to limited piston supplies, after ring flutter had further widened the top ring groove. Single compression ring NA experiments with all three versions found negligible differences in performance and blow-by after adequate component bed-in. The effect of oil consumption is difficult to accurately quantify due to the relatively short running periods compared to OEM cycles, however little variation also existed between all ring packs. One aspect evident during the development was the reduced bed-in time for all positive scraper cast iron rings, evident by the earlier blow-by reduction, which was the metric for gauging the seal during engine run-in (Appendix L). Furthermore, one effect due to the differing ring materials was the resistance to ring fracture and hence failure due to high knock intensities. The chrome plated steel and plain cast iron materials tended to be quite brittle and hence fracture, with ring pieces breaking away and causing combustion

122

chamber damage. The SG cast iron material was advantageous in surviving the heavy knocking environment, tending to plastically deform but not fracture. Development also focused on ring end gap clearances between the two and three ring configurations. Inspection of top ring components after testing found evidence of ring butting, resulting in a further 0.1-0.2 mm increase in ring end gap clearance for the single compression ring design. These results suggest increased thermal growth for the single compression ring design due to higher piston temperatures associated with the reduced heat flux into the cooling system.

5.4.3 CR Variation
As the test modes required different CRs, the piston crown was modified through simple machining processes (turning) to reduce the CR, with the piston design allowing reductions from 13 to 9 (Chapter 3.5.3). However, for experiments the CR was only reduced to 9.6. The CR was varied in four steps, as dictated by the limiting knock severity (Table 4.8). The varying chamber geometries associated with each CR are shown in Figure 5.8. It is noted that squish areas around the periphery of the chamber were maintained to minimize the variation in squish generated turbulence and vorticity, limiting any resulting combustion effects from varying CRs. However, changes to the turbulent length scale were inevitable due to the changes in the combustion chamber dish depth as the CR was altered.

2400 CR = 13 2000 1600 1200 800 CR = 11 40 50 60 70 80 90 10 CR = 10 00 00 00 00 CR = 9.6 0

Figure 5.8: CR variations ranging from 13-9.6 through piston crown modifications.

123

With the test engine, CR reduction below 9 was also possible with the installation of annealed copper decompression spacers of varying thicknesses (0.9-3 mm), which could be inserted between the crankcase and barrel. Due to the gasketless design, it was unfeasible to place the spacers between the head and barrel. However, the annealed copper decompression spacers were never implemented as CR reduction below 9.6 was not needed during development. Moreover, it is noted that there were problems with this method of CR adjustment as the squish height would increase by the same amount as the spacer, resulting in combustion changes due to reducing turbulence and hence burn rates.

5.4.4 Piston Oil Cooling

The engine was designed with underside piston oil cooling to ensure component reliability at the high heat loads associated with the high power density (Chapter 3.5.3). Originally, a single jet was positioned to direct oil flow towards the exhaust side, as exhaust side surface temperatures have been shown to be higher in comparison to the intake [55, 98] (Figure 9.18). Hence, it was assumed that the piston would benefit from exhaust side cooling, due to previously documented exhaust side piston failures [66, 248] (Chapter 9.6.1).

Carbon deposits

Oil pooling

Adequate Oil Control

Poor Oil Control (Fourfold increase in underside oil cooling)

Figure 5.9: The effects due to a significant increase in underside oil cooling on the single compression ring piston design (Type A). Disassembly after ~5 hours operation at similar conditions. (Left): Dry crown. (Right): Chamber contamination with oil pooling and deposits.

124

However, after initial encounters with knock that resulted in intake side piston land failure (Figure 9.17), substantial piston oil cooling development work was undertaken to mitigate high knock intensities during turbocharger implementation. The first inlet side top land failure was originally thought to have been caused by high intake side piston surface temperatures as the exhaust side received underside oil cooling. However, inlet side top land failure continued to occur after the implementation of a second jet, directed towards the inlet side. side, with a 0.45 mm diameter orifice. failure. Later, the intake side oil flow was increased twofold and then fourfold in an attempt to reduce knock intensities. However, this resulted in no change in knock location and did not reduce knock intensities. Rather, the fourfold increase in oil flow resulted in increased knock intensities due to the inadequate oil control associated with the single compression ring (Piston Type A), which allowed increased quantities of lubricating oil to enter the combustion chamber as seen in Figure 5.9. The increased knock was likely caused by the reduction in fuel quality associated with the oil consumption, which left carbon deposits that further increased the propensity for knock. Furthermore, the higher oil flow increased parasitic losses due to the extra energy required to drive the oil pump. Additionally, the high amounts of oil directed on the piston underside increased the windage losses and at low speeds the high oil flow rate compromised oil system pressure. These effects all contributed to reducing the engines MECH. As piston underside oil cooling had no effect on reducing knock intensities or location, for knock control, cooling the centre piston crown was found to be relatively unimportant in comparison to reducing piston temperatures near the periphery of the chamber, where end-gas knock occurs. Although increasing the underside oil cooling most probably reduced mean piston temperatures, the oil cooling was ineffective in surface temperature control due to the pistons high thermal mass. Furthermore, large changes in surface temperatures are needed to affect end-gas temperatures or flame propagation. Hence, instantaneous This jets orifice size was varied, originally having the same flow capacity as the exhaust However, this development yielded no change in knock severity or location as indicated by further inlet side piston land

125

temperatures and not mean temperatures are important for knock avoidance, which is unachievable with oil cooling. Consequently, oil cooling had little effect on reducing end-gas temperatures and did not reduce knock as temperature hot spots related to surface ignition were not present. described in Section 5.4.1. Instead, it was found that piston ring design is more effective in heat dissipation and hence knock control, as

5.4.5 Piston to Valve Clearance

The piston to valve clearance development is outlined as the limits have been found in this engine, with these limits being applicable to four valve, pent roof, high speed designs. The clearance is important as it defines the required cutout geometry in the piston crown, permitting increased CRs, reduced surface to volume ratio and also combustion burn rate improvements. Originally, the engine featured 1.2 mm intake and 1.7 mm exhaust piston to valve clearance at closest proximity, with the clearance expected to reduce during engine operation due to dynamic component deflection, valve movement control and thermal geometry changes. These clearances were based on experience with two valve per cylinder racing engines and were considerably smaller than those found in production four valve, pent roof designs [214, 215]. No piston to valve contact could be tolerated as contact would bend the head of the valve relative to the valve stem due to the included valve angles, which would in turn cause valve failure and subsequent component damage. The adopted clearances were successful in preventing contact for the NA modes. However, after rapid piston land failure due to boosted knock problems, it was thought that the top land structural integrity could be improved by reducing the local valve cutout depths. This would increase the minimum top land thickness and reduce piston stress concentrations. Hence, the piston to valve clearance was reduced by 0.4 mm on the intake side (1.2 to 0.8 mm) as the piston land had previously suffered failure on the intake side. The reasoning behind the clearance reduction stemmed from the fact that two valve racing engines have larger and heavier valves necessitating larger clearances. Hence, it was assumed that the valve movement of the smaller and hence lighter valves of the test engine could

126

be more accurately controlled due to the reduced inertia, permitting smaller clearances. Unfortunately, this led to engine failure as shown in Figure 5.10 after piston to valve contact at high speed. The heavy damage in the second cylinder made it impossible to determine the level of contact, however the slight contact in the first cylinder suggested clearance limits were only slightly exceeded. Hence the minimum piston to valve clearance to operate successfully for this particular engine was found to be > 0.8 mm on the intake side on the basis that failure occurred at 0.8 mm. Following engine failure, the engine was repaired with no additional development completed to determine minimum exhaust side clearances due to the possibility of further engine damage.
Cylinder 1 Cylinder 2

Valve contact

Valve head embedded in piston crown

Figure 5.10: The effects of insufficient piston to valve clearances. (Upper): Poppet valve piston contact in individual cylinders. (Lower): Subsequent engine damage as a result of valve failure in Cylinder 2.

127

5.4.6 Piston Skirt and Liner

The piston and liner interface, consisting of the piston skirt profile and the piston to bore clearance, required development as the pistons were specifically manufactured for the test engine. The piston skirt geometry, namely the cam and barrel profiles were adapted from a Lancer EVO-7, with technical geometry information supplied by Mitsubishi Australia [152]. The EVO-7 piston profile was used as both engines were boosted and produced high BMEP, with the EVO-7 widely used in rally motorsport applications. In addition, both engines utilized a forged piston in a cast iron liner, with a modern pent roof combustion chamber. Initial engine commissioning began with a 70-90 m piston to bore clearance due to honing difficulties associated with the gasketless interface as previously described (Section 5.2). These clearances were greater than those found in the EVO-7 (+ 25 m), however during engine commissioning, piston seizure was of greater concern than high oil consumption. This concern was caused by the possible higher mean piston temperatures, due to the unproven first iteration piston design which featured a single compression ring. Following engine commissioning, disassembly and inspection revealed no concerns regarding piston skirt longevity, indicated by visual evidence of good piston to bore contact during engine operation. However, after both liners were destroyed (Figure 5.23) and subsequent replacement, the piston to bore clearance was reduced to 60 5 m due to machining tolerances. greater than that of the EVO-7. Upon further NA testing at high speeds (10,000 rev/min), blow-by rates rapidly increased due to inadequate ring sealing. A cylinder compression test revealed poor cranking cylinder pressure, due to the piston liner damage shown in Figure 5.11. The excessive damage is caused by the breakdown of the boundary lubrication film between the piston skirt and liner, resulting in metal to metal contact. The failure was attributed to greater piston thermal growth due to the higher temperatures associated with the single compression ring design, which rejected less heat (Section 5.4.1) when compared to the two compression ring EVO-7. However, it was not anticipated that the reduction in clearance would be a problem, as it was still

128

Figure 5.11: Piston skirt and liner damage during engine development.

To remedy the problem and ensure engine reliability, the piston to bore clearance was increased to 80 5 m for NA operation and 100 5 m for boosted operation. The larger clearance for the boosted mode was to accommodate the higher piston temperatures associated with the increased combustion temperatures and power density. Following clearance increases, no further piston skirt or liner problems were encountered during experiments. One important finding from the development is the negligible difference in oil consumption and blow-by with piston to bore clearance increases of this magnitude, although limited testing was completed when compared to OEM durability cycles.

5.5 Inlet Manifold

5.5.1 Reliability
Intake manifold development for reliability was accomplished through solving problems as they arose in experimental testing. The initial PFI manifold was Although manufactured from a combination of polymer and metallic materials, including Nylon66 runners and an aluminum plenum chamber (Figure 5.13). successful in the NA mode, the higher intake pressures and temperatures associated with boost reduced the mechanical properties of the polymer material. This caused the 8 mm thick mating flanges to eventually buckle, resulting in intake air leaking past the o-ring seal. Initially, steel plates were used to stiffen the polymer flanges with the o-ring seals increased twofold in diameter and relocated to the cylinder head. This proved unsuccessful as the stiffening plates detached from the plastic flanges at the elevated manifold temperatures.

129

The polymer manifold manufactured from Nylon66 using the SLS method was proven to successfully work for this particular engine operating in the NA mode. During engine development, this type of manifold is advantageous as several manifolds can be manufactured quickly and cheaply for evaluation. In addition, the semi-transparent intake runners can withstand some backfire cycles and as the combustion traveling up the runner is visible, it can give insight into abnormal events. Upon long exposure to elevated intake temperatures, the material was found to lose its mechanical properties and fail. To withstand boosted operation, a new lightweight, 1.6 mm wall, aluminum manifold was fabricated with the same geometry as the initial SLS version. However, the aluminum was prone to fatigue cracking (Figure 5.12) from the excessive vibration associated with solid mounting the manifold to the cylinder head. Although this setup had previously worked in four cylinder configurations [93, 138, 187], the vibrations due to the engines odd fire twin configuration exacerbated the problem and significantly reduced the manifolds life. To reduce the effects of vibration, the aluminum manifold was made in three sections; the plenum chamber and a runner for each cylinder, connected by flexible silicone tubing. The plenum chamber was supported by vibration dampening mounts No which were attached to the cylinder head as depicted in Figure 5.13. drawing detail given in Appendix P.

subsequent failures were recorded after the development with further manifold

Repaired fatigue cracks

Continued failure

Figure 5.12: Fatigue crack repairs and continued failures in the lightweight aluminum alloy manifold discovered during development.

130

In summary, experimental testing showed that although successful in NA operation, the polymer Nylon66 manifold could not withstand the elevated intake temperatures associated with boosted operation, thus requiring a metallic system. However, when a lightweight aluminum manifold is fitted to an odd firing twin configuration, some isolation or dampening from the engine is required to ensure manifold longevity.
Initial Manifold Design Final Developed Version

Nylon runners prone to distortion and leakage under boost pressure

Plenum supports and rubber dampening needed to prevent fatigue cracking

Figure 5.13: Initial polymer and final aluminum alloy inlet manifold versions.

5.5.2 Fuel Injector Location

PFI engine development initially commenced on the dynamometer with each cylinders fuel injector positioned low in the intake runner, as depicted in Figure 5.14 Setup 1. However, after several knock induced piston failures, the fuel injector was relocated near the trumpet entry (Figure 5.14 - Setup 2). This has been reported to have brake output advantages due to intake charge cooling effects caused by vaporizing the liquid fuel [14, 94]. The charge cooling is caused by the energy exchange between the gasolines phase change enthalpy and the air with which it mixes. Even when heat transfer from the walls is included, intake charge temperatures can reduce by as much as 22C depending on the MAP, thus

131

increasing air density and VOL [220]. NA dynamometer tests confirmed a 2-3% increase in brake output at WOT when compared to the low injector setup. This intake charge cooling effect was also observed under some ambient conditions with ice forming on the lower part of the polymer intake runner. The two injector locations resulted in different ECU fuel calibrations, with greater individual cylinder trimming needed to accommodate the high injector setup. This was caused by a combination of unequal air and fuel distribution, the latter a result of the high injector setup which was more prone to fuel back flow into the plenum chamber due to the intake resonant waves. This effect was further magnified with the unequal firing interval, with the second cylinders induction offset by 180 CA. Injection timing also needed adjustment between both setups to compensate for the charge transport delays of the longer mixing lengths. MBTinjection timing methodology was used to optimize both setups, as previously described in the tuning strategy (Chapter 4.6.4).

High Injector (Setup 2)

Bell-Mouthed Trumpet Entry

Low Injector (Setup 1)

Figure 5.14: Intake manifold section highlighting the two differing injector setup positions used during development.

132

FSAE vehicle testing commenced with the high injector setup. Initial testing found poor transient response, with lags between throttle tip-in and engine power increase. Transient calibration had previously not been optimized on the steady speed dynamometer. This calibration was intended to be performed using feedback information from the driver, while running a variety of vehicle tests. Unfortunately, transient calibration parameters were restricted by the Motec ECU, which allowed only a single accelerator enrichment value to be adjusted. Maximum enrichment was limited to a 100% increase, based on the fuel main table, by the ECU software.
8000 1.5 1.4 Engine Speed (rev/min) 6000 1.3 Lambda ( ) Throttle Position (%) 1.2 4000 1.1 1.0 2000 Engine Speed 2105 2110 2115 2120 Time (s) 2125 Lambda 2130 2135 0.9 0.8 0 2100 0.7 2140

100 80 MAP (kPa) 60 40 20 0 2100 MAP 2105 2110 2115 2120 Time (s) Thottle Position 2125 2130 2135

100 80 60 40 20 0 2140

Figure 5.15: Fuel mixtures at associated MAP, TP and engine speed during transient vehicle testing for the high injector configuration (Setup 2) depicted in Figure 5.14. Note the fuel mixture excursion from lean to rich on accelerator tip in, caused by fuel wall wetting due to the long mixing length.

133

The logged data from vehicle testing (Figure 5.15) revealed MAP followed TP, indicating increased airflow. However, the engine speed did not simultaneously increase as the recorded fuel mixtures from the oxygen sensor showed a lack of compensation for the sudden increase in airflow. Mixtures initially went lean and then rich outside the measurement range of the wide band oxygen sensor, which limited mixture logging ( = 0.8-1.2) as shown in Figure 5.15. These mixture deviations were only evident on sudden accelerator tip in and caused by the long mixing length associated with the high injector position. This gave the fuel a chance to precipitate after initial vaporization, encouraged by the cold surfaces of the polymer intake runners. fuel handling and ECU calibration. Attempts were made to minimize these effects, with full accelerator enrichment used to avoid the lean region. However, this was unsuccessful with insufficient enrichment for fuel deposition, resulting in only excessive rich regions once the liquid was completely deposited. Ignition advance was adjusted at part open throttle (POT) in an attempt to increase cylinder temperatures to aid fuel vaporization, but had little benefit. Thus the high injector setup was discarded. A strategy combining both high and low injectors could have been used but dismissed due to the increased system complexity and time limitations. In addition, the system would only increase air consumption at less than choked flow operating conditions. If the system was implemented, high injectors would only be used at consistent WOT operation, which improved brake power output. The low setup would be used at all other operating conditions to avoid wall wetting problems previously described and hence improve efficiency. High injector setups are common in motorsport applications [14, 94]. However, the reasoning behind the unsuccessful implementation in this application involves the upstream throttle location and long curved intake runners, tuned for lower speeds. Downstream throttling, which is a feature in the majority of racing Other advantages include engines, aids in improving transient response. This caused wall wetting and/or pooling along the length of the intake runner, adding complexity to the transient

improved fuel breakup and mixing at reduced throttle openings due to the higher local velocities associated with the reduced flow area past the throttle [98].

134

5.5.3 Intake Runner Geometry

For the PFI system, several variations in intake manifold runner geometry were tested. From experimentation with varying trumpet geometry (Figure 5.16), it was confirmed that resonant wave tuning is highly dependent on the effective tract length (L1) coupled with valve timing events. The effective length differs slightly from the actual length (L2) as shown in Figure 5.16. The actual length is defined as the centerline distance from the intake valve to the highest trumpet edge. However, varying tract geometry involving the inner bell-mouth radius (r1) and the splayed angle () can cause minor changes to the effective length, while the actual length remains unchanged. TC experimental results with varying inner and outer bell-mouth trumpet geometry found only minor changes in performance due to the low tuning speeds and the relatively high MAP, making the engine fairly insensitive to trumpet geometry changes. These results also corresponded to steady state airflow bench testing results (Appendix H), which suggested little airflow variation if high discharge coefficients are maintained for trumpet geometry changes. Intake runner length variation was also conducted in TC experiments. Tract

length variation was tested at two intake lengths although the polymer intake system could be incrementally varied (Chapter 3.7). The purpose of the length variation was to experimentally validate the theoretical model created in Chapter 3.7 and cross reference the CFD optimization data (Chapter 3.3.2).
Parallel Trumpet Geometry Splayed Trumpet Geometry

r1

r2

>0

Actual length

Actual length

L2

L22 L

Effective length L1 = A

Effective length L1 < A

Figure 5.16: Varying trumpet geometry highlighting the actual and effective length, governed by the defined trumpet parameters (parameter A is a constant).

135

Hence, only a short (L2 = 400 mm) and long (L2 = 550 mm) runner were tested to verify the extremes of the model at engine speeds which placed high boost demand on the turbocharger (regions which were yet not flow restricted, < 6,000 rev/min). Figure 5.17 displays images and comparisons of both tract lengths, with experimental results shown to correlate with the theoretical modeled data and CFD simulation results. Hence, no further runner length variations were tested during development as the optimization had been completed by simulation.

Short Runner (400 mm)

Long Runner (550 mm)

120 100 VOL VOL (%) 80 60 40 20 0

Theoretical Model Ricardo W ave CFD Experiment

350 400 mm Effective Tract Length

550 mm

Figure 5.17: (Upper): Constructed short and long runner intake manifolds. (Lower): Short and long runner engine air consumption comparison over theoretical, CFD and experimental testing. TC - PFI, 4000 rev/min, WOT.

136

5.6 Exhaust Manifold

Due to the engines odd firing configuration, an exhaust plenum was required to smooth the exhaust pressure fluctuations into the turbine (Chapter 6.5.1). There is little documentation concerning plenum exhaust manifold designs as they are seldom used, except for single cylinder research applications. However, nonplenum manifold failures are well documented [145, 255] and generally engine specific. Hence, exhaust manifold reliability development for the test engine is now outlined. Several exhaust manifold failures were experienced during testing due to the severe cyclic mechanical and thermal loads, exacerbated by the plenum design. Initial failures were a result of non-flexible ducting to the laboratories exhaust system, as previously described (Chapter 4.2.6). Further problems are described leading to the manufacture of three exhaust manifolds to overcome failures. The initial version was fabricated from 1.6 mm wall thickness 316 stainless steel, due to its superior heat conductivity when compared to mild steels. However, manifold cracking problems attributed to out-of-phase thermal fatigue due to the compressive strains that occur at high temperatures [145] quickly led to alternative material selection. Hence, a new mild steel manifold was manufactured to the same geometry, with no further problems occurring in NA testing. Mild steel was chosen due to its resistance to fatigue cracking [13, 15, 204]. However, initial concerns were raised over turbine reliability as the carbon scale that generally forms on the manifold skin has the potential to pass through the turbocharger causing damage to the turbine blades. However, no damaged blades were witnessed during experiments. Attaching the turbocharger to the plenum manifold led to further failures due to higher exhaust temperatures, with the hottest regions of the manifold coinciding with those of highest mechanical stress (Figure 5.18). duct (Figure 5.18), causing crack initiation and manifold failure. The mass of the turbocharger and the vibration placed higher loads on the exhaust plenum outlet

137

Cracks (cyclic mechanical and thermal loads)

Top

Bottom

Thermal Stresses

Mechanical Stresses

Regions of high thermal and mechanical stress concentrations

Figure 5.18: Exhaust manifold development due to the higher thermal and cyclic mechanical stresses at the outlet junction due to turbocharger installation. (Upper): Reliability concerns due to manifold cracking. (Lower): Thermal loads from experiments and FEM analysis highlighting plenum chamber stress concentrations due to vibration.

138

To fully support the mass of the turbocharger, the manifold wall thickness was increased to 3 mm with a vibration support added between the turbocharger and the crankcase to strengthen the system. FEM analysis was also used to further reduce loading on the manifold outlet junction, which led to the incorporation of a stiffening rib. This reduced the mechanical stress by a factor of four over the previous design (Figure 5.18). No further exhaust manifold reliability concerns were witnessed in testing, with Figure 5.19 displaying the final constructed item and the FEM analysis.

Figure 5.19: (Left): Final constructed exhaust plenum design. (Right): FEM analysis resulting in a fourfold reduction in mechanical stresses when compared to the previous design in Figure 5.18 - Lower.

5.7 Engine Balance and Vibration

From the commencement of dynamometer testing, higher than expected external engine vibration was evident. The engine mounts were initially thought to be the source of the problem, hence the Acetal Nylon bushes were replaced with a more elastic 70 durometer rubber material. Rubber mounting was necessary to minimize the vibration effects associated with the inline twin configuration, which caused oil line and manifold component fatigue failures. Changing the mounting scheme resulted in only minor vibration reductions. Consequently, alternative crankshaft counterweight designs with varying balance ratios were tested in an attempt to minimize the external vibration effects. The crankshaft counterweight designs are shown in Figure 5.20 with further details listed in Table 5.4.

139

Table 5.4: Varying crankshaft design specifications used in experiments. Crankshaft A

Appendix P, Drawing # Crankshaft mass Proportion of rotating mass balanced with counterweight Proportion of reciprocating mass balanced with counterweight Bob-weight mass attached around crankpin for balancing 0707A 5.3 kg 100% 0% 0.31 kg

B
0707B 5.9 kg 100% 40% 0.44 kg

C
0707C 5.6 kg 100% 50% 0.48 kg

Crankshaft A (Original design)

Crankshaft B (Crankshaft A with heavy metal inserts)

Crankshaft C (Final design)

Figure 5.20: Varying counterweight geometry designs used to alter the rotating and reciprocation balance proportions for differing crankshafts used in experiments.

The initial crankshaft design (Chapter 3.5.4) simply balanced the rotating mass with the counterweight, neglecting the reciprocating component in order to minimize the rotational inertia and hence improve the engines crankshaft acceleration. Although the counterweight reduced some vibration, the vertical vibration due to both cylinders piston motion remained, which caused the engine to rock longitudinally and components to fatigue. In order to reduce the peak vibration amplitude, the crankshaft was modified to a 100% rotating, 40% reciprocating balance ratio. This was successful in longitudinal vibration reduction, however lateral vibration increased due to the counterweight motion. The balance ratio adjustment was achieved by pressing heavy Mallory metal inserts [41] into the existing cheek (Figure 5.20), which was

140

outsourced, to avoid the manufacture of a new crankshaft. However, upon engine disassembly after several weeks of operation, it was discovered that the heavy metal inserts had begun to dislodge as shown in Figure 5.21. shrink fit was too high, considering the location of the insert. Through FEM The excessive analysis, it was determined that the amount of interference used to create the interference caused the cheek to plastically deform and hence reduced the contact area of the shrink fit, making the insert susceptible to movement under vibration. To avoid the possibility of engine failure due to the inserts dislodging, a new one piece crankshaft was manufactured with a 100% rotating, 50% reciprocating balance ratio. This ensured equal lateral and longitudinal peak vibrations, which were the lowest achievable without the use of a secondary rotating balance shaft. The crankshaft balancing method involved attaching adjustable mass bob-weights to both crankpins and either balancing dynamically on a testing rig or statically between v-blocks. No visible change in vibration or engine power was recorded with either method. Additionally, the crankshaft balance ratio was only shown to affect external engine vibration and not power measured on the fixed speed dynamometer.

STRESS

MPa 1250 1000 750 500 250 0

Figure 5.21: Failed crankshaft (Left): Heavy metal insert movement. (Right): FEM failure analysis (von Mises stress distribution) simulating the interference fit and the resulting plastic deformation in the alloy steel crank cheek.

141

5.8 Camshaft and Timing Chain

After the initial engine commissioning (~10 hours operation), when the rocker cover was removed for inspection, it was evident that both the inlet and exhaust side camshaft thrust bearings had failed. The excessive camshaft side movement caused the camshaft to wear the cylinder head side covers, as shown in Figure 5.22. Initial investigation suggested that the high wear rates were caused by the uneven surface finish of the camshaft thrust faces. Camshaft thrust bearings could not be replaced as they were integrated into the camshaft bearing housing. Hence, the excessive clearance was reduced by nickel plating the camshaft and then machining to suit the widened camshaft thrust bearing groove. Furthermore, upon camshaft removal for repair, wear marks were visible on the side of the cylinder barrel timing chain opening (Figure 5.23). The excessive timing chain side movement during engine operation was thought to be caused by the failed camshaft thrust bearings, with the higher clearance allowing more timing chain side movement. As no further barrel damage due to the timing chain movement was tolerable, the inside surfaces of the barrel timing chain opening were hard anodized to prevent further damage. However, one unfortunate consequence of the outsourced anodizing process, due to negligence, is the erosion of the cast iron liners beyond repair, as also visible in Figure 5.23.

Wear due to camshaft thrust bearing failure

Figure 5.22: Engine damage caused by camshaft thrust bearing failure.

142

Liner erosion

Barrel timing chain opening

Figure 5.23: Consequences of negligence during the hard anodizing process with the eroding of both cast iron cylinder liners with close-up detail.

Following engine repair and subsequent testing, camshaft thrust bearing failure continued to occur. Once again the excessive clearance was repaired by nickel plating the camshaft thrust faces to take up the excessive clearance. To rectify the high wear rates, more lubricating oil was fed into the thrust bearing. This did not solve the problem and subsequent thrust bearing failure continued, as did barrel damage due to the excessive timing chain side movement. After subsequent repair, it was envisaged that the thrust bearing failures resulted from the excessive side loading of the camshaft. This was due to the fluctuating timing chain side movement, which caused camshaft axial loading from chain misalignment. To remedy the problem, a timing chain side guide rail was implemented into the camshaft drive system, as shown in the drawings of Appendix P. All guide rails were manufactured from an oil filled Nylon-6 material [6], which ensured some lubricity and hence minimized wear between the sliding components. No further camshaft bearing failures were recorded in testing. The thrust bearing development is worth documenting as camshaft axial movement is generally not of concern in even fire multi-cylinder configurations ( three cylinders). The problem is specific to engines with few cylinders (< three cylinders) which feature close firing intervals. In these types of engines, there exists the potential for intermittent and therefore uneven camshaft loading due to the valve spring force. Hence, over a portion of the crankshaft rotation, there is

143

no valve spring load on the camshaft allowing its angular velocity to vary in relation to the driving crankshaft. This in turn, reduces the tension on the timing chain which allows more movement, hence causing higher rates of misalignment which may lead to camshaft thrust bearing failures or other problems.

5.9 Summary
In this chapter, the engine developments required to ensure the reliability and robustness associated with highly turbocharging an odd fire, inline twin configuration, were presented. Although, relatively few problems were encountered in NA modes after engine commissioning, ongoing TC engine development was necessary until the later stages of testing. Components requiring major developments for performance and reliability after the initial design (Chapter 3) included the gasketless barrel and liner assembly, intake and exhaust manifolds, the piston assembly and the crankshaft. Problems concerning these components were gradually overcome through an incremental development process by evolving the initial designs. The novel gasketless interface design outlined in Chapter 3.6 has proved to be successful, performing faultlessly with no sign of any interface combustion gas or fluid leakage after initial development. required clearances. Development involved determining the manufacturing processes required to implement the design whilst maintaining the The novel gasketless design has enabled successful engine operation when coupling an open deck barrel design with high pressure ratio turbocharging. The previously documented bore distortion problems that occur in highly boosted engines and the concomitant structural integrity reductions of the open deck design have been overcome, with similar blow-by values recorded when compared to other conventional designs. The success of the interface is further demonstrated, with sealing integrity maintained for test conditions up to 25 bar BMEP, sustained high knock intensities as well as engine overheating. This included peak cylinder pressures exceeding 10 MPa, with combined knock amplitudes of 6 MPa. The sealing integrity of the interface is further highlighted by the piston failures and elevated combustion temperatures it withstood without failure.

144

The piston assembly required the most development to achieve reliable highly boosted engine operation. Both high mechanical and thermal aspects were overcome, with the engine performance not limited by the piston assembly after subsequent developments. Several areas in the piston design were targeted to enable engine operation at 25 bar BMEP. These included the piston ring lands and ring pack, combustion chamber geometry and resulting CR, underside oil cooling and the determination of adequate bore and valve clearances. Furthermore, the developments allowed the combustion operating limits to be extended without piston failure through knock suppression, which enabled higher output. A major step forward during the TC development was reverting back to a conventional three ring pack design, consisting of two compression sealing rings. Although the single compression ring design worked adequately for NA operation, severe oil consumption, blow-by and heat transfer related problems caused numerous piston failures during TC testing. Hence, the piston rings were shown to be the most critical piston design factor, heavily affecting performance and efficiency. It is anticipated that further improvements are also possible through optimizing the correct ring number, thickness and location. Significant manifold and engine balance development was required due to reliability concerns caused by the high mechanical and thermal stresses associated with high speed, high pressure ratio, two cylinder operation. This combination resulted in both intake and exhaust manifold fatigue failures caused by vibrational and thermal loads. These problems were overcome by simultaneously reducing the engines peak vibrational amplitude through crankshaft development and structural improvements to the manifolds, aided by FEM analysis. The engine development outlined in this chapter confirms the design chapters findings (Chapter 3.9) which concluded that the mechanical aspects of the component design are not the engine performance limiting factors. Although, the conclusions in Chapter 3 were founded on analytical and CAE analysis, this chapter further supports these findings with experimental evidence. The thermal aspects of the component design have also been overcome in experimental tests, with the engine successfully operating near 300 kPa MAP and engine speeds exceeding 10,000 rev/min.

145

146

CHAPTER 6
P O W E R E D B Y W A T T A R D

Turbocharging

6.1 Introduction
This chapter outlines the selection, development and performance optimization required to successfully implement a 2004 Garrett model GT-12 turbocharger to the test engine. Turbocharger selection based on various weighted criteria is described together with previous well documented problems associated with oil control in throttled compressors. Oil control is highlighted as a major problem, as the Garrett GT-12s lubrication system was not designed to operate with the compressor at below atmospheric conditions. Throttled conditions occur due to the test engines intake system layout, which features upstream throttling and a regulated intake flow restriction. A novel solution is presented which is applicable to modern The solution is turbocharger designs, using existing piston ring type oil seals.

shown to simultaneously enhance engine transient performance and reliability. Further turbocharger development regarding cooling and boost control is also outlined. Matching the engine to the turbocharger was guided by CAE based analysis tools. Validation of the computer based simulations with experimental results is also presented. Engine matching focus is placed on valve events together with intake and exhaust geometry, particularly to accommodate the odd firing two cylinder configuration.

147

6.2 Turbocharger Selection

Challenges were faced in the selection and procurement of a turbocharger, due to the engines small capacity. A background search revealed very few units which could meet the airflow and corresponding pressure ratio requirements over the engines operating speed range (Appendix E). Two units meeting requirements were the GT-12 and GT-15, which were manufactured by Garrett Honeywell for current day diesel and gasoline platforms. branch with relatively short lead times. Although these units were manufactured offshore, they could be sourced locally through the Australasian Turbocharger selection was heavily influenced by many factors as summarized and weighted in Table 6.1. Further indepth analysis and reasoning behind the decision tables results are given in Appendix E.2. As indicated in Table 6.1, the GT-12 was selected. 12. Although both units were

closely matched, the boost control method and improved cooling favored the GTHowever, both units would suffer from excessive oil consumption under compressor throttling due to the test engines induction system layout and the inadequate turbocharger seal design for this application [15, 75, 252].
Table 6.1: Decision table used in turbocharger selection. Details in Appendix E.2. Garrett GT-12
Factor Availability Matching Implementation Cost Control Mass Packaging Cooling/Lubrication Aggregate Weighting (/5) 5 4 4 3 3 2 2 2 Raw (/5) 5 4 3 5 4 3 3 4 100 Weighted 25 16 12 15 12 6 6 8

Garrett GT-15
Raw (/5) 5 3 3 5 2 1 1 2 78 Weighted 25 12 12 15 6 2 2 4

148

6.3 Oil Control in Throttled Compressors

Previous literature studies have found both advantages and disadvantages in varying the throttle location in TC applications [143, 247, 252]. However, a well recognized problem arises from placing the throttle upstream of the compressor, which results in oil from the journal bearing being drawn into the inlet manifold past the compressor side seal assembly [13, 15, 143, 252]. Nevertheless, FSAE regulations mandate a downstream turbocharger location relative to the throttle. Consequently, the engine was developed with the throttle and intake flow restriction located on the suction side of the compressor. This layout (Figure 6.1) significantly increased the difficultly of implementing a turbocharger into the intake system. At part load or restrictor choking conditions, the manifold vacuum generated would induce engine oil from the turbocharger lubrication system into the compressor housing and thus the intake manifold [13, 15]. This is not only an imposition on the required oil capacity, but the oil consumption also causes major combustion problems ranging from plug fouling to increased probability of preignition and/or end-gas knock. These problems all contribute to a significant reduction in engine performance and reliability, often resulting in clouds of white smoke being emitted from TC FSAE vehicles [15].
Upstream throttle Oil Turbine

20 mm intake restriction Engine Compressor

Figure 6.1: Test engine turbocharger layout and resulting upstream compressor throttling due to the mandatory throttle location and fitment of the intake restrictor.

149

Despite the drawbacks, upstream throttling in TC applications has many benefits including reduced turbo lag and increased spool rates (rotational acceleration) during engine transients [143, 247, 252] without the need for costly and sophisticated variable turbine geometry (VTG) [167]. However, these benefits are normally outweighed by the oil control issues previously highlighted. under non-throttled compressor conditions. The turbochargers found in most current production vehicles are designed to operate As a result of this design layout, these turbochargers feature reliable piston ring type oil seals as there is no suction on the compressor side. This type of seal has many advantages when compared to the outdated carbon face seal design [252]. when implemented as in Figure 6.1. The seal leakage and hence oil flow into the intake manifold is caused by the pressure in the compressor being lower than the engine sump at engine throttled conditions. So although the turbocharger bearing is fed engine oil at a pressure of 3-5 bar, a small portion of the oil drains into the compressor housing (past the inadequate piston ring seal) at conditions where the pressure in the compressor is lower than that in the sump (PCOMP < PSUMP = PATM). This seal leakage is not important under non-throttled conditions as compressor pressures are always equal to or higher than atmospheric. This results in small amounts of intake air bleeding into the turbocharger lubrication drain system, which is recycled with blow-by into the intake manifold or vented to the atmosphere. Thus the modern turbocharger design limits applications to conditions where compressors are unthrottled. The oil control problems associated with compressor throttled applications have previously been overcome using carbon face seals, which have been proven to successfully operate under throttled conditions [191, 248, 252]. However, the Shaft design is more costly to implement, with oil seal longevity and performance becoming major issues, thus limiting the life and reliability of the unit. loading friction and thus spool time are all increased compared to the piston ring design, due to the increased face seal drag. Turbochargers fitted with this design are the only choice for carbureted draw through applications, which by definition are throttled [143]. However, the implementation of fuel injection to engines has However, these seals fail to generate a complete seal, and thus are susceptible to this oil consumption issue

150

resulted in carbon face seals becoming obsolete.

No carbon face seal

turbochargers were available for this particular engine, thus development centered on finding a solution for the Garrett GT-12, which featured piston ring type oil seals as found in most current day turbochargers.

6.4 Turbocharger Development

Applying the Garrett GT-12 to the test engines intake system required significant development as the application differed from what the unit had originally been designed for. The major differences between the applications were the turbocharger location and operating conditions. The development process used to solve the previously outlined oil control problems is outlined, together with cooling and control issues. Figure 6.2 displays a sectional view of the OEM supplied turbocharger scroll prior to development. GARRETT GT-12 TURBOCHARGER (as supplied)
Oil feed (illustrative only) Water cooling jacket Centre bearing

Compressor piston ring seal

Compressor

Compressor seal housing

Turbine piston ring seal

Water cooling jacket Oil drain

Turbine Centre bearing housing

SCALE 1:1

Figure 6.2: Sectional view of the Garrett supplied GT-12 turbocharger with the compressor and turbine housings removed. (Note): Water cooling jacket detail altered and oil feed relocated for illustrative purposes only.

151

6.4.1 Oil Control

Initial turbocharger development involved installing the as supplied unit in order to quantify oil consumption amounts. It was originally thought that the induced oil could be trapped and collected in the intake plenum, as the trumpets were raised from the plenum floor. However, oil losses were excessive, with several liters consumed at idle conditions after a short running period of ~15 minutes. Several strategies of controlling turbocharger oil consumption were attempted and are outlined chronologically in Table 6.2, along with their effects on oil consumption rates.
Table 6.2: Methods and resulting affects for controlling turbocharger oil consumption under compressor throttled conditions. TC - PFI, 2300 rev/min (fast idle speed), 50 kPa MAP, 3 bar turbocharger oil feed pressure. Turbocharger Oil Control Strategy
1) 2) As supplied 1) + limited oil supply flow and increased drain size 2) + check valve accumulator system 2) + externally scavenging OEM supplied carbon face seal 2) + new compressor seal housing with reduced escape area (1 ring) Modified 6) with labyrinth annulus (1 ring) Modified 7) gapless ring seal (2 rings in 1 groove) Modified 7) + 2 rings in separate grooves

Oil Recovered from Plenum (L/hr)

5-8

Comments
High oil pooling and exhaust smoke 1.6 mm oil line restriction fitted, drain size to match turbocharger housing Unsuccessful Successful at idle but not at MAP values below scavenge (sudden tip-out) Increased turbo lag and shaft loading, exhaust smoke upon sudden tip-out Some improvement, 50% reduction in oil escape area Continued improvement Exhaust smoke on sudden tipout Silver soldered centre bearing housing 1.6 mm vent. No exhaust smoke on tip-out

3-4

3)

3-4

4) 5) 6)

0.1

7)

0.7

8)

0.5

9)

0.2

10) 9) with vent between both ring seals

152

The first attempt (Table 6.2: Strategy 2) to solve the oil consumption problem involved limiting the oil supply flow to the turbocharger and increasing the drain size. This provided adequate lubrication and reduced oil losses but was only Due to its success, this strategy was incorporated into all partially successful.

further developments. The second method (Table 6.2: Strategy 3) involved the use of a check valve accumulator system, which attempted to equalize the pressure across the compressor side seal at intake manifold vacuum conditions. At these conditions, oil would be stored in an accumulator pot, which would then empty under boosted conditions. However, high pressure losses across the check valves and high opening (cracking) pressures resulted in the system failing. A schematic of the failed attempt is displayed in Figure 6.3.
Oil In Compressor Intake Restrictor Centre bearing housing

Turbine Throttle Oil Out

Intake Manifold Check Valve 2 Baffle Pot

Check Valve 1 Crankcase Sump Figure 6.3: Schematic displaying the failed oil control strategy (Table 6.2: Strategy 3) to combat compressor throttled problems. Strategy involves creating a pressure balance across the sealing system.

153

Following this, turbocharger oil drain scavenging via an electric external source was attempted. If successful, a specific scavenge pump stage from the dry sump lubrication system may have been implemented on the turbocharger oil out drain line. Scavenging proved successful at MAP values above the scavenge pressure. However, oil losses and exhaust smoke occurred on sudden throttle tip-out, as MAP would drop to 20 kPa. Although unsuccessful, this result indicated that further development should be focused on preventing the oil from escaping, rather than trying to draw the oil out from the drain with scavenge. The next step involved testing an OEM modified unit which became available, retro-fitted with a carbon face seal and supplied by the manufacturer. Although the new unit substantially reduced oil loses, oil consumption was not completely eliminated. The implementation of the carbon face seal also reduced turbocharger rotational acceleration. Turbocharger reliability issues due to the higher shaft loading from the seal drag were also of concern, with several competing FSAE teams experiencing shaft failures in competition [15, 75]. Thus the carbon face seal was not employed in further developments. Subsequently (Table 6.2: Strategy 6), the internals of the turbocharger were modified with development focusing on the compressor seal assembly. A new compressor seal housing with an increased outside diameter was implemented. This design increased the pressure drop across the seal and limited the oil escape path to the compressor. The enlarged housing further reduced oil losses and was incorporated into later designs. Several iterations were investigated, ranging from a 20-70% reduction in escape area. However, as a result of the enlarged housing, the rotating assembly was more prone to surface rubbing as the bearing or shaft deteriorated. This was combated by compressor seal housing material selection and the fact that the two surfaces could act as a plain bearing. The final version was a compromise between the competing effects and settled on a 50% reduction in escape area. Incorporating a labyrinth annulus seal in the housing further increased the pressure drop and was implemented in future developments. Although significant oil consumption improvements were made with this housing and seal package, additional development was still required to eliminate the problem.

154

OEM GARRETT GT-12 TURBOCHARGER (as supplied)

Compressor seal assembly

Single piston ring seal

OEM compressor seal housing SCALE 2:3

FINAL OIL SOLUTION (Strategy 10)

Compressor seal assembly

Vent (1.6mm orifice)

New compressor seal housing

Reduced escape area

Labyrinth annulus

Welded insert (increased land area)

SCALE 2:3

Figure 6.4: (Upper): Sectional view of the Garrett supplied GT-12 turbocharger. (Lower): Sectional view of the redesigned turbocharger to overcome oil consumption problems under throttled conditions. Note: Water cooling jacket detail removed and oil feed relocated for illustrative purposes only.

155

The next stage involved incorporating a gapless ring seal to eliminate the ring end gap. This was achieved by placing two ring seals in a widened groove. consumption was attributed to the ring seal end gap. pressure drop across the sealing system. This was only partially successful and it was concluded that only a minor portion of the oil This suggested further improvement could be achieved by separating the two seals to maximize the However, developments were still constrained by the piston ring seal location, due to the limited area in the seal region as seen in the upper diagram of Figure 6.4. Hence, the land area was increased by machining the centre housing to accommodate an insert which was silver soldered in the housing and later machined to the required dimensions. This allowed two ring seals to be located in series, separated by the annular labyrinth seal as seen in Figures 6.4 and 6.5. The piston ring seals were rotated by 180 with respect to each other, in order to minimize the ring end gap effects. This further reduced oil loss and the final solution used this concept as its basis. The final solution involved venting the labyrinth annulus cavity to atmosphere. This was achieved by drilling a single 1.6 mm diameter hole through the centre bearing housing, as depicted in Figure 6.5. This solved the oil issue because if vacuum was induced behind the first seal, the compressor would induct air rather than suck oil past the second seal. The vent also provided a positive pressure balance across the second seal to ensure identical conditions to those the turbocharger was originally designed for. The final seal design featured excellent oil retention with no oil film in the intake system, even after several months of dynamometer and vehicle operation. This oil retention is further illustrated by the HC emissions, with recorded values as low as 57 ppm C6 at throttled conditions. The final seal design allows the continued use of compressor side piston ring seals, found in modern turbochargers for all applications with no oil consumption effects. However, the design has additional performance benefits when compared to usual modern layouts as it also allows the turbocharger to be placed downstream of the throttle. Consequently, vehicle drivability and transient response is improved without affecting engine and turbocharger reliability. When compared to previous conventional sealing methods (carbon face seals) used in downstream turbocharger applications, the new seal design has enabled

156

reduced unit cost and turbo lag [167] while increasing unit life, reliability and rotational acceleration. This is achieved with the elimination of oil consumption over all operating conditions, thus ensuring engine performance and reliability. Furthermore, exhaust after-treatment effectiveness and longevity is improved by the removal of any catalyst oil poisoning. No turbocharger reliability issues were evident after oil development and subsequent modifications, with the system completing hundreds of hours of dynamometer and vehicle operation over an extended testing period.
CENTRE BEARING HOUSING Silver soldered joint Vent Steel insert

ROTATING ASSEMBLY Dual ring seals Seal housing

Figure 6.5: Strategy 10 - Components used to overcome the oil control issue. (Upper): Modified centre bearing housing with silver soldered insert and orifice vent. (Lower): Rotating assembly with new compressor seal assembly.

157

6.4.2 Cooling
The OEM supplied GT-12 unit featured water cooling in the centre bearing housing as depicted in the sectional view of Figure 6.2. Water cooling prevented problems associated with turbocharger heat soak after rapid shutdown. This had been a major issue in previous turbocharger programs at the University of Melbourne [191, 248]. The problem affects shaft and bearing reliability as the high temperatures can cause the oil within the turbocharger to pyrolize or carbonize, resulting in subsequent bearing failure on restarting. Although turbocharger heat soak could be avoided on the dynamometer during development, it was exposed to this running condition during vehicle operation and as such could not be dismissed. Hence, development commenced with the engine coolant lines re-routed to accommodate the turbocharger. However, significant engine reliability issues arose from the onset of initial engine run-in and calibration in the TC mode at light load operating conditions where the engine had previously operated successfully in NA modes (Appendix I.5). Thus at the time of development, the validity of developing the turbocharger without the centre bearing cooling was explored as its effect on the overall engine cooling system was unknown. turbocharger plumbing. Prior to commencing the turbocharger cooling development on the dynamometer, analysis was performed to explore the effects on unit reliability. Heat transfer and thermal growth changes were analyzed due to the possibility of differing temperatures in the centre bearing housing. This placed more emphasis on dissipating the majority of centre bearing heat through the oil lubrication system. However, this was not dissimilar to the majority of non-water cooled units, which under close monitoring had proven to work successfully under the high heat loads associated with motorsport and other high BMEP applications [66, 171, 248]. Hence to ensure reliability, the centre bearing housing was retro-fitted with thermocouples to monitor inner core temperatures during development. Other benefits included overall unit mass reduction and simplifying the complexity of the

158

There also existed the potential to dissipate some portion of the heat to the atmosphere through the housing surface area. This was achieved by sealing the water cooling jacket within the turbocharger and filling the cavity with a high boiling temperature fluid. This promoted inner core heat transfer through the High body to the atmosphere, as the insulating air gap had been removed.

temperature synthetic oil was used as the medium, with recorded centre bearing cavity temperatures not exceeding 160C, even after heavy heat soak from hard running and rapid shutdown. To further reduce the likelihood of oil pyrolizing or carbonizing, high quality synthetic lubricating oil was used. This synthetic oil has a higher flash temperature compared to the older mineral based oils. No shaft or bearing problems were recorded after hundreds of hours of engine operation. Component wear rates were also monitored upon subsequent turbocharger disassembly, with no signs of abnormality.

6.4.3 Intake Boost Regulation

Initial development commenced with no form of exhaust bypass through the turbine. This was achieved by welding the internal wastegate flap valve closed, thus forcing all exhaust flow through the exhaust turbine. Boost control relied upon correct matching and choked flow through the intake restriction to limit airflow and avoid compressor surge. However, during experiments it was found advantageous to limit the pressure drop to reduce the turbocharger compressor work and hence exhaust backpressure, as there was no need to reduce the suction pressure when the nozzle was already choked. This pressure drop across the nozzle was verified using upstream and downstream absolute pressure sensors, thus verifying when the intake system was running in the choked condition. The possibility of reducing pumping losses or producing useful work out of the compressor by increasing the boost pressure at the choked condition was also explored. However this strategy increased the MAT, which increased the knock tendency and surge probability causing reduced brake output. Thus, the internal wastegate flap valve was used to control the turbocharger, which allowed the exhaust gas to bypass the turbine.

159

Wastegate movement was controlled using a pneumatic diaphragm opposing the force of a mechanical spring, with the diaphragm actuator referencing MAP. The MAP reference signal lines were kept as short as possible to maximize wastegate response, with an inline restrictor ( 3 mm) fitted to dampen pressure pulsations to the diaphragm. The MAP signal was also taken from the side of the plenum chamber, which was less prone to resonant wave interference from the inlet tracts. Initially, the OEM supplied diaphragm relied solely on MAP input to control boost levels, with intake pressure limits pre-set at 150 kPa. However, this system did not meet the variable boost regulation needed at differing engine speed and MAP requirements, associated with the choked intake restrictor. Variable boost regulation was initially achieved by physically altering the spring tension at wastegate opening to achieve the required MAP at a fixed engine speed. This method of control was adequate for maximizing performance during fixed speed dynamometer testing. However, this system would not meet engine transient requirements when fitted to a vehicle.
Bleed valve (ECU controlled)

Vent to atmosphere

Diaphragm actuator MAP signal (dampened via restrictor)

Compressor Wastegate flap valve Turbine

Figure 6.6: Boost control schematic highlighting how the raw MAP signal was manipulated by the ECU. This enabled the wastegate movement to be ECU controlled to achieve the desired boost levels.

160

After further development, final boost control was achieved using an electronic control system [251]. This involved using an unmodified diaphragm, with the MAP reference signal manipulated via an electronic bleed valve [15, 143]. The pulse width modulated bleed valve allowed the pressure sensed by the actuator to be varied by bleeding a portion of the MAP signal to atmosphere. Consequently the wastegate remained closed allowing the turbocharger to produce higher MAP than the actuator had previously allowed. Wastegate response was improved by positioning the bleed valve in the signal line towards the intake manifold, immediately after the inline dampening restriction. The bleed valve was controlled through the ECU, allowing the turbocharger to be calibrated to deliver the required MAP at user defined functions of engine speed and TP. A schematic of the final system is displayed in Figure 6.6. To ensure engine reliability as a result of bleed valve failure, the system was designed to default close. This resulted in no manipulation of the MAP signal, allowing the wastegate to function as per the original OEM design. Other specific TC safeguards were later configured into the ECU to ensure engine reliability for increasing BMEP. This included simultaneous random fuel and ignition interrupts (Chapter 4.2.5) to control intake air temperatures together with rising MAP levels past predefined limits. The system was also configured to trigger a warning light once these limits were exceeded, notifying the dynamometer operator or vehicle driver.

6.5 Turbocharger Performance Optimization

An experimentally validated 1-dimensional CFD simulation model [14, 15, 113] as described in Chapter 3.3.2 was used in conjunction with experimental dynamometer testing to optimize turbocharger engine performance. Parameters known to significantly affect turbocharger performance were investigated with WAVE, including intake and exhaust geometries together with valve events [191, 248, 249].

161

6.5.1 Exhaust Manifold Geometry

Exhaust manifold design including the configuration and dimensions were heavily influenced by the unequal pulse frequency effects attributed to the odd (uneven) firing configuration. The unequal pulsing caused problems ranging from exhaust turbine velocity fluctuations to ineffective cylinder scavenging, which limited achievable boost pressures. The ineffective cylinder scavenging resulted in high residual gas fractions, which decreased performance and increased knock tendency. To mitigate this effect, several exhaust manifold concepts were proposed. The first implemented pulse type turbocharging, with unequal exhaust primary lengths to compensate for the odd pulsing. Pulse type turbocharging is beneficial in even fire applications as it has advantages associated with reducing turbocharger spool time and delivering higher pressure ratios [143, 252]. This is due to utilizing the steady frequency of high kinetic energy gas flowing during exhaust blowdown, where the discharge pressure and velocity are high. However, the proposed unequal length pulse design had limitations in that the exhaust lengths are engine speed dependent and thus the system could only be tuned to eliminate odd pulsing at a given speed. Other issues included the inherent pipe frictional losses associated with the high velocities, which reduced the pressure rise rate at the turbine nozzle. Constant pressure turbocharging, featuring an exhaust plenum chamber, was also investigated as it smoothed the exhaust mass flow into the turbine by dampening the intermittent pulsing. However, the increased manifold surface area increases the heat losses and reduces delivered energy to the turbine [15, 191, 249]. Moreover, supplying constant pressure to the turbine improved efficiencies when compared to pulse systems [113, 143]. stationary applications. This has often resulted in constant pressure systems being optimized to operate at fixed operating conditions for The concept of placing high performance emphasis on one operating point is not dissimilar to the test engine due to the peak power limitation imposed by the flow restriction. Hence, the test engine was configured to produce sonic flow at the lowest possible engine speed to broaden the speed range over which increased power is delivered.

162

The exhaust manifold geometry for both constant pressure and pulse systems were investigated with the help of WAVE. This enabled the competing effects to be examined for this particular odd firing engine, with the simulation results for varying exhaust geometries shown in Figure 6.7. The equal length pulse manifold is taken as the base case for comparisons in Figure 6.7 due to its relative ease of manufacture. Simulation results from Figure 6.7 show significant brake power and torque improvements over the engine speed range when employing constant pressure type turbocharging to this particular odd firing engine. Performance differences between the constant pressure system and pulse system are largest at lower engine speeds, prior to the intake restriction limiting airflow. At the higher engine speeds, the performance difference is reduced due to the higher exhaust pulse frequencies. It follows from the results that smoothing the intermittent exhaust pulses to the turbine significantly improves the test engines performance. These intermittent pressure pulsations associated with the pulse system are shown in Figure 6.8, with simulation results showing large variations in pressure, prior to the mass entering the turbine.
60 70 60 50 (A) (B) (C) (D) (E) Pulse: Equal Length (0.7L volume) [1] Plenum: Log (0.7L volume) [2] Plenum: Sphere (1L volume) [5] Plenum: Log (2L volume) [3] Plenum: Log (4L volume) [3] 40 30 20 10 0 11000 Improvement (%) over Pulse (A)

50

Brake Power (kW)

40

30

20

10

0 5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.7: Simulated results displaying the performance discrepancies between constant pressure and pulse type turbocharging for the flow restricted test engine fitted with various exhaust manifold geometries. [ ] = Ease of manufacture / 5.

163

(A) Pulse System: Equal Length

600
Turbine Housing Pressure (kPa)

500 400 300 200 100 -180

Cylinder 1 pulse Cylinder 2 pulse

180 Crank Angle (deg)

360

540

Figure 6.8: Simulated pressure pulsations in the turbine housing when applying pulse type turbocharging (Model A) to the odd firing test engine. TC - PFI, 6000 rev/min, single engine cycle.

Simulation results from Figure 6.7 also highlight the reduced performance of the equal length pulse system (Model A) when compared to similar exhaust volume constant pressure systems (Models B-C) across all speeds, caused by the increased residual gas content. Performance increases are found from increasing the exhaust manifold volume, as this reduces the exhaust pulsing effect thus reducing the exhaust turbine velocity fluctuations and increasing delivered boost. Simulation results for the constant pressure system with varying plenum volumes (Models B-E: 0.7 - 4 L) shows a 20% improvement in peak torque at conditions where the intake restriction begins to limit airflow by increasing the plenum volume. A 5% improvement in peak power is also recorded in the choked flow operating region. Consequently, a 30% peak torque and 12% peak power improvement is recorded for the constant pressure system (Model E) over the base case pulsed system (Model A). From the simulation results of Figure 6.7, an exhaust manifold geometry with a volume of four liters was selected (x 10 cylinder volume). Larger volumes provided small gains in output but were coupled with increased heat losses due to the increased manifold surface area, which reduced transient response and thus were not used. Furthermore, increasing the exhaust volume inherently increased

164

mass transport times, which may reduce transient response at low firing frequencies (low engine speeds) depending on driver detection. The constant pressure system inherently increased turbo lag over the pulse system due to the reduced amplitude of the exhaust pressure pulsations, which spooled the turbocharger at quicker rates. However, this effect was significantly reduced due to the low rotational inertia of the unit as a consequence of its relatively small size. Turbo lag could also be minimized using other traditional methods [23, 142, 174, 252, 254]. The most efficient plenum geometry was thought to be a sphere, as this has the minimum surface to volume ratio, reducing heat losses. However, simulation results from Figure 6.7 (Models B-C) show that the log style plenum had reduced flow losses for similar volumes and so was employed. The log style plenum also permitted the adoption of Watsons KEC rolling flow design (Chapter 3.7), which had been used successfully in intake manifolds [13-15, 93], but as yet had not been applied to exhaust manifold designs. A series of simulations were also conducted for an even firing configuration (Model F: 0, 360 CA). The purpose of the even fire simulations were to verify the test engines simulation findings and compare the performance results to a more conventional engine and turbocharger package as found in current passenger vehicles. Figure 6.9 displays the effect of firing spacing coupled with pulse and constant pressure systems. Results demonstrate that the peak power of the constant pressure system (Model E) can be matched with the equal firing pulse system (Model F) after the odd firing has been eliminated. Results after minimal optimization also show the potential for the even fire pulse configuration (Model F) to match the overall power curve of the optimized odd fire constant pressure configuration (Model E). The high speed performance differences shown are caused by varying cylinder residuals as both configurations were limited in airflow (choked flow) at these operating speeds. It may be assumed that the exhaust manifold dimensions and valve timing of the even fire pulse configuration could be further optimized to reduce this effect.

165

60

85 75

50 Improvement (%) over Pulse (A) 65 Brake Power (kW) 40 (A) Pulse: Equal Length (Odd Fire: 0-180 CA) (E) Constant Pressure (Odd Fire: 0-180 CA) (F) Pulse: Equal Length (Even Fire: 0-360 CA) 55 45 35 25 15 10 5 0 5000 -5 11000

30

20

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.9: Simulated brake power results comparing firing interval (odd and even) for constant pressure and pulse type turbocharging.

Simulation findings for a flow restricted engine suggest little if any peak and overall power differences between an odd fire constant pressure and an even fire pulse configuration. However, simulation results show that low to mid speed power is heavily dependent on the turbochargers ability to deliver the required boost. In an odd firing two cylinder configuration, this is easier to achieve with a constant pressure system, requiring little parameter optimization to achieve good results. Consequently, the engine developed in the TC mode commenced with the constant pressure design (Model E). However, due to manifold reliability problems previously outlined (Chapter 5.6), the plenum was replaced with the simple two into one pulse system (Model A) to continue development and quantify simulation findings. Both manufactured exhaust manifold designs are displayed in Figure 6.10. With this pulse manifold design, engine torque was reduced by as much as 30% as shown in Figure 6.11, with severe knock problems at previously determined MBT-ST values. In order to avoid the DL for the given CR for the pulse system, boost levels and spark timing were reduced and thus performance suffered, with experimental results for a fixed speed shown in Figure 6.12.

166

PULSE TYPE (Model A)

CONSTANT PRESSURE (Model E)

Figure 6.10: Varying exhaust manifold geometry. (Upper): Two into one pulse system (Model A). (Lower): Watsons KEC constant pressure plenum system (Model E).

167

2400

Expt. Pred. Expt. Pred.

Plenum (Constant press) Plenum (Constant press) 2 into 1 (Pulse type) 2 into 1 (Pulse type)

BMEP (kPa)

2000

1600

1200

800 4000

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.11: Experimental versus simulation BMEP results for constant pressure and pulse type turbocharging. TC - PFI, WOT, CR = 10.

Results from both simulation and experimentation show significant brake power improvements when employing constant pressure type turbocharging to this particular odd firing engine. Thus the initial constant pressure plenum design was once again implemented to improve engine performance, after further manifold reliability development outlined in Chapter 5.6.

1800 1600 1400 BMEP (kPa) 1200 1000 800 600 400 200 0 100

= KL-ST

59 53 47 41 (A) Pulse Type (E) Constant Pressure 35 29 23 17 11

110

120

130

140 MAP (kPa)

150

160

170

5 180

Figure 6.12: Experimental BMEP and spark timing results versus MAP for constant pressure and pulse type turbocharging. TC - PFI, 5000 rev/min, CR = 10.

168

Spark Timing (deg. BTDC)

Spark Timing (deg. BTDC)

6.5.2 Inlet Manifold Geometry

To determine the size of the intake plenum, a series of simulations were conducted, with results for varying volumes shown in Figure 6.13. Simulation results show that the plenum volume is the most influential intake parameter to optimize for a TC flow restricted odd firing engine [14, 15, 113]. As with the exhaust geometry, the odd firing effect influences the intake system resulting in the second cylinder having reduced airflow. This is due to the plenum chamber being partially evacuated by the previous induction with insufficient time for recovery. TC simulated results show no detrimental brake performance effects due to this charge robbing effect as long as the cylinder mass flow variations are reduced to below 5% when the engine is running in the flow restricted choked condition. This is due to the engine being force fed air via the compressor, which is constantly inducing the maximum mass flow through the orifice over all CA degrees. This minimizes the intermittent pulsing through the flow restriction However, if charge robbing caused by the piston motion during induction.

variations are greater than 5% due to reduced plenum volume, brake performance is found to gradually decrease due to residual gas content differences between cylinders and the piston motion having an increased effect on flow through the restriction.

16 6L Intake Volume 2L Intake Volume Airflow Variation (%) [Cylinder (1-2)/1] 12

0 4000

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.13: Simulated results showing cylinder airflow variations for TC operation with varying intake volumes due to varying plenum chamber sizes. Airflow variations associated with charge robbing effects caused by the flow restriction and odd firing interval.

169

300 Indivudual Cylinder VOL (%) 250 200 150 100 50 0 4000

Cyl 1 (Exp) Cyl 2 (Exp) Cylinder Variation (Exp) Cylinder Variation (Sim)

40 35 Airflow Variation (%) [Cylinder (1-2)/1] 30 25 20 15 10 5

5000

6000

7000

8000

9000

0 10000

Engine Speed (rev/min)

Figure 6.14: Experimental individual cylinder VOL and cylinder variations versus simulated results. Variations due to the charge robbing effects associated with the flow restriction and odd firing interval. TC - PFI, CR = 10, 6 L intake volume.

Thus a 6 L intake volume was chosen from simulation results. Larger plenum volumes reduced cylinder variations but did not improve TC brake performance and reduced engine transient response at low firing frequencies. TC experimental results confirm simulation findings, with Figure 6.14 displaying individual cylinder

VOL and resultant cylinder variations as confirmed by each cylinders exhaust gas
analysis. Thus the performance output from individual cylinders is also expected to differ by this margin. Hence individual cylinder calibration was required, with up to a 5% second cylinder fuel trim needed to equalize AFR at WOT for the TC mode. Experimental results in Figure 6.15 show the intake charge temperature effects due to turbocharging, with MAT values displayed. To ensure consistent results, the Bosch MAT sensor was centrally located in the plenum chamber between both cylinders (Chapter 4.2.5). It is noted that the temperature at the inlet valves is likely to be less than that measured in the plenum chamber due to the fuel vaporization cooling effect, which is only marginally offset by the heat transfer from the inlet port walls.

170

The experimental data confirms the previous design decision to avoid intercooling to control intake temperatures (Chapter 3.7). Dynamometer results from Figure 6.15 show MAT rarely exceeding 60C over a wide range of operating conditions. The relatively low intake temperatures compared to other TC setups [66] are due to the large surface to volume ratio of the intake system and the high compressor efficiencies due to the matched operating regime (Section 6.6). pressure ratio due to the fixed mass flow from the flow restriction. Thus the Over the compressor could be matched to operate over a wide speed range and resulting range of fixed speeds, intake temperatures are lowest at 6,000 rev/min for varying MAP, which corresponds to the compressor operating near best efficiency for the given pressure ratio (Figure 6.20). As the speed increases, the intake temperatures increase due to reduced compressor efficiencies and the increased heating effect due to higher residual fraction associated with higher exhaust back pressures (Section 6.6). The controllable intake temperatures achieved during dynamometer testing are further strengthened by transient testing. Logged intake temperatures with the engine fitted to a FSAE vehicle during competition also show similar trends with MAT not exceeding 55C, after completing the endurance event (40 minutes of sustained competition) in 35C ambient conditions.

70

60 MAT (C)

6000 rev/min 7000 rev/min 8000 rev/min 9000 rev/min 10000 rev/min

50

40

30 60 80 100 120 MAP (kPa) 140 160 180

Figure 6.15: Experimental TC intake manifold air temperatures for various engine speeds and MAP. Results highlight the possibility of not intercooling the boosted charge for small engines if matched correctly.

171

Experimental and simulation results for varying intake geometries to improve TC brake performance showed only minor effects in comparison to optimizing exhaust geometries. This is due to the flow restriction, with performance significantly Due to the affected by the rate at which the turbocharger can deliver the required airflow to reach the choked flow operating condition as seen in Figure 6.7. upstream throttle location, the potential exists to improve engine transient response at low firing frequencies (low engine speeds) by minimizing the intake volume. However, reducing the intake volume causes increased cylinder airflow variations due to the intake charge robbing effects described. increases. The transient effects the driver feels are reduced if not eliminated as the firing frequency

6.5.3 Valve Timing - Camshaft Specification

Camshaft geometry, mainly valve timing values were optimized through simulation to improve TC brake performance. Limits were set in camshaft parameters, mainly valve lift and acceleration rates to ensure reliability and avoid valve float as OEM poppet valves and matching valves springs were utilized during development. Values for these limits were predefined from manufacturers specifications [214, 215]. Consequently, valve durations and intake and exhaust interaction during valve overlap were investigated. Optimizing intake and exhaust flows for specific TC operation had previously been documented to improve the performance in small capacity 1.3 L engines [23, 52, 249]. More recent applicable camshaft studies involved the performance feasibility of turbocharging a flow restricted 0.6 L engine [113]. However, findings obtained from previous studies were for four cylinder configurations with equal firing intervals. Other variations included differing Nevertheless, results engine operating speeds and the intake flow restriction.

showed up to a 44% [249] and 45% [113] torque improvement at lower engine speeds by reducing the valve overlap in comparison to the NA valve timing. These findings warranted further investigation, thus camshaft simulation began by varying the valve overlap parameter for this particular flow restricted odd firing engine.

172

The purpose of initial simulation runs was not to target significant gains in peak power as this was flow restricted. More importantly, the objective was to increase low to mid-range torque where the flow restriction had yet not limited airflow. Thus, initial simulation runs were completed with reduced exhaust plenum volumes to see this effect more clearly. Furthermore, performance improvement through camshaft optimization at reduced plenum volumes also explored the feasibility of reducing the exhaust volume, which minimized the heat losses and hence improved transient response (Section 6.5.1). Figures 6.16 to 6.18 display camshaft simulation results, using the previously optimized NA restricted four cylinder valve timing [113] as an initial starting point and hence a base case for all comparisons. Simulation results for varying valve overlap are shown in Figure 6.16 (varying inlet valve opening (IVO) and exhaust valve closing (EVC) positions), achieved by varying the intake and exhaust valve duration while keeping the exhaust valve opening (EVO) and inlet valve closing (IVC) position fixed. The variations in valve overlap were achieved with camshaft profile changes, with the following valve timing notation used to distinguish profiles (EVO:BBDC, EVC:ATDC, IVO:BTDC, IVC:ABDC). Figure 6.16 presents valve overlap simulation results for two differing exhaust volumes. Results show up to a 4% peak power and 7% peak torque improvement by reducing the base case valve overlap from 55 to 35 for TC operation. No detrimental performance effects were simulated for all reduced valve overlap cases, with continued power and torque increases over base cam timing, even when the valve overlap is nearly eliminated (TC - Cam 4: 15 valve overlap). The only performance reduction arises with the increased valve overlap case (TC Cam 1: 60 valve overlap). Hence, the 35 valve overlap value was used in all further simulations. The power and torque increases resulting from the reduced overlap simulations are attributed to improved cylinder filling. At mid-range speeds, improved cylinder filling is caused by the reduction of reverse flow from the exhaust to the intake due to the slight increase in boost. At higher engine speeds, utilizing the induced charge via efficient cylinder trapping through reductions in through scavenging during valve overlap improves output as airflow is limited by the restriction. The simulation results show that the odd firing TC engine requires reduced overlap due

173

to the intake and exhaust interactions which behave differently in NA modes. These results strongly correlate with the literature for small equal fire TC engines, with a 20 reduction in valve overlap needed to compensate for a TC intake system [23, 113].
60 18

50
EVO/EVC, IVO/IVC

14

Brake Power (kW)

40

30

(NA - Base) 77/25, 30/75 (TC - Cam1) 77/20, 40/75 (TC - Cam2) 77/15, 30/75 (TC - Cam3) 77/15, 20/75 (TC - Cam4) 77/ 5, 10/75

10

20 2

10

0 5000

6000

7000

8000

9000

10000

-2 11000

Engine Speed (rev/min)

60

18

50
EVO/EVC, IVO/IVC

14

Brake Power (kW)

40

30

(NA - Base) 77/25, 30/75 (NA - Base) 77/25, 30/75 (TC Cam3) 77/15, 20/75 (TC - Cam3) 77/15, 20/75 (TC - Cam5) 77/25, 20/75 (TC - Cam5) 77/25, 20/75

10

6 20 2

10

0 5000

6000

7000

8000

9000

10000

-2 11000

Engine Speed (rev/min)

Figure 6.16: Test engine simulated brake power results (WOT) with varying valve overlap for constant pressure type turbocharging. (Upper): 0.7 L exhaust volume. (Lower): 1.5 L exhaust volume.

174

Improvement (%) over NA - Base Cam

Improvement (%) over NA - Base Cam

Figure 6.17 displays simulation results for varying IVC positions using the valve timing determined from Figure 6.16 (TC - Cam 3: 35 valve overlap) as an initial starting point. Closing the inlet valve earlier was expected to significantly improve performance for the TC mode [23, 249]. However, simulation results illustrate very small performance change with a 15 reduction when compared to the initial IVC position of 75 ABDC, with slight losses in peak power (-1%) but improvements in peak torque (+4%) simulated. Prior to the intake restriction limiting airflow, the results show signs of following expected trends with slight improvements in cylinder filling, increased available energy to the turbine and consequently increased compressor speed and boost pressure. However, the improvements do not continue in the flow restricted region, attributed to changing intake and exhaust interactions, with the effects from experiments shown in Section 6.6. These results also correspond to previous findings from the TC flow restricted four cylinder engine simulations [113]. A 5 reduction in IVC position to 70 ABDC allowed the peak power to be maintained but slightly reduced the torque improvement when compared to best simulated case. into further camshaft simulations.
60 30 26 22 18 14 10 6 10 2 -2 11000

Hence, a

compromise between peak power and torque resulted in this being implemented

50
EVO/EVC, IVO/IVC

Brake Power (kW)

40

(NA - Base) 77/25, 30/75 (TC - Cam3) 77/15, 20/75 (TC - Cam6) 77/15, 20/70 (TC - Cam7) 77/15, 20/60

30

20

0 5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.17: Test engine simulated brake power results (WOT) with varying IVC timing for constant pressure type turbocharging. 2 L exhaust volume.

175

Improvement (%) over NA - Base Cam

60

34
Improvement (%) over NA - Base Cam

30 50 26
Brake Power (kW)

40

EVO/EVC, IVO/IVC

22 18 14 10 6

30

(NA - Base) 77/25, 30/75 (TC - Cam6) 77/15, 20/70 (TC - Cam8) 70/15, 20/70 (TC - Cam9) 65/15, 20/70 (TC - Cam10) 60/15, 20/70

20

10 2 0 5000 -2 11000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 6.18: Test engine simulated brake power results (WOT) with varying EVO timing for constant pressure type turbocharging. 2 L exhaust volume.

The effects of varying the EVO position were explored using the valve timing determined from Figure 6.17 (TC - Cam 6: 35 valve overlap, IVC = 70 ABDC) as a starting point, with simulation results shown in Figure 6.18. Results showed up to a 3% peak power improvement when reducing the exhaust valve duration by altering the EVO position from 77 to 60 ABDC. The power increase simulated was maintained over the entire flow restricted region due to the reduced energy supplied to the turbine, which could already provide enough compressor work to choke the flow restriction. As a result of the reduced turbine energy, the pumping losses reduced during exhaust blowdown with simultaneous expansion work increases. This effect was small, but nevertheless performance improved ~3% over the flow restricted region when compared to the starting point of Cam 6. Table 6.3 summarizes the camshaft simulation work, with the valve timing and resulting maximum power and torque compared between the initial base camshaft (NA - Base [113]) and the simulated improved version (TC - Cam 10). When compared to the NA base case, a 20 reduction in valve overlap is recorded together with a 27 reduction in exhaust valve duration and a 15 reduction in intake valve duration. Figure 6.18 also compares the performance changes of

176

both camshafts across the varying speed range.

No performance reduction is

recorded for the improved TC camshaft with highest improvements over the NA base at lower engine speeds prior to the intake restriction limiting airflow. Peak performance improvements of up to 10% are recorded at 5,000 rev/min together with a 9% peak torque improvement at 6,000 rev/min over the NA base timing. However, the magnitude of these performance improvements reduces as the brake power is limited by the flow restriction. A 4% peak power improvement is achieved at the choked condition by reducing the residual fraction and pumping losses together with improved cylinder trapping. Simulation performance results for camshaft variation showed fairly insensitive magnitude changes when compared to varying exhaust manifold geometry (Section 6.5.1). Simulation results also highlighted that the camshaft optimization had smaller effects on performance as the exhaust volume was increased. This was due to the exhaust geometry playing a larger role in delivering higher boost pressures at lower speeds (prior to the flow restriction limiting power) in comparison to the camshaft geometry. The exhaust volume plays a larger role than the camshaft for this particular engine configuration as the plenum chamber dampens the uneven flow pulses attributed to the odd firing interval, hence altering the delivered boost. Moreover, the valve timing was shown to improve performance near maximum power (flow restricted region) where the induced airflow was limited by the flow restriction. The performance improvements were associated with reducing the residual fraction and pumping losses together with improved cylinder trapping.
Table 6.3: Summary of test engine TC simulation results with NA and TC camshafts. Valve timing changes and associated improvements. NA - Base Cam [113]
EVO (BBDC) EVC (ATDC) IVO (BTDC) IVC (ABDC) Maximum Power Maximum Torque 77 25 30 75 57.6 kW 73.5 Nm

TC - Cam10
60 15 20 70 60 kW 80 Nm

Improvement 4% 9%

177

Simulation results for camshaft variation found identical performance effects when compared to other restricted and unrestricted even fire studies [23, 52, 113, 249] by altering the NA base camshaft geometry to suit the TC application. Specifically, reducing the valve overlap and intake and exhaust duration for this odd firing engine has produced similar performance effects to even fire engines. However, the magnitude improvements due to camshaft optimization are somewhat limited by the exhaust geometry and flow restriction for this particular odd firing engine.

6.5.4 Final Results

TC brake performance optimization was significantly influenced by CFD tools, with the importance of using simulation as an aid in accelerating engine development clearly demonstrated. These tools have enabled improved TC performance and understanding at reduced costs and lead times. Figure 6.19 compares the experimental and simulated TC performance data for the test engine. Experimental and simulation results show that by successfully implementing turbocharging to a flow restricted engine, near constant power can be achieved over a wide speed range. This is due to the intake restriction, which limits airflow and thus performance when operating at the choked condition. The potential also exists to further improve low speed performance (prior to the intake restriction limiting airflow) with VTG [167]. Experimental results also show that peak power occurs at the lowest flow restricted engine speed (6,000 rev/min) due to the reduced friction losses which increases brake power.
60 50 Brake Power (kW) 40 30 20 10 0 4000 Experiment (TC-PFI) Simulated Model

5000

6000

7000

8000

9000

10000

11000

Engine Speed (rev/min)

Figure 6.19: Comparison of experimental and predicted engine performance for the test engine operating in the TC mode.

178

The importance of using a validated simulation model in setting the engine specifications during the initial design phase (Chapter 3.3) is clearly demonstrated in Figure 6.19. Excellent agreement between actual and simulated performance is evident, with minor differences at 10,000 rev/min when the actual brake power tends to fall faster than simulated. This is partly explained by higher pumping losses than simulated (Figure 6.22) and increasing friction losses (Figure 7.11) associated with increasing component flexure at the high engine speeds. In addition to the performance comparisons of Figure 6.19, the development process used to improve TC performance through intake and exhaust geometries together with valve events further illustrates the agreement with simulation results.

6.6 Final Turbocharger Matched Operation

Figures 6.20 and 6.21 display the compressor and turbine maps for the GT-12 turbocharger, with the engine operating points overlaid. The airflow is limited to ~0.07 kg/s due to the flow restriction (Appendix D.1 and H.5), with engine operation at this limit highlighted by the WOT line between 6,000 and 10,000 rev/min on the compressor map of Figure 6.20. Hence, when the engine operates at the choked condition, the compressor runs near peak efficiency at the highest levels of boost and maintains relatively good efficiency (70-75%) as the power is flow limited to maximum speed. This information further strengthens and explains the relatively low intake temperatures displayed in Figure 6.15. However, the compressor operates very close to the surge line at 6,000 rev/min, with peak efficiency occurring outside the choked engine speed range (77% at 5,500 rev/min). This suggests that slight improvements could be achieved with a larger compressor (increased A/R ratio). Figure 6.21 displays the turbine operating conditions. Although the compressor maintains good efficiency over the wide range of engine operating conditions near maximum power (Figure 6.20), the turbine efficiency drops rapidly above 6,000 rev/min and continues to reduce up to maximum speed. At WOT over this operating speed range, the turbine efficiency has a 30% relative reduction when compared to 6,000 rev/min. The significant reduction in turbine efficiency causes the exhaust back pressure to increase which in turn increases pumping losses (Figure 6.22) and hence pumping mean effective pressure (PMEP).

179

Garrett GT-12 Compressor

Pressure ratio and corresponding mass flow to achieve choked flow at 6,000 rev/min

E N I L E G R U S
T O W W O T

Pressure ratio and corresponding mass flow to achieve choked flow at ~10,000 rev/min

in

in

in /m re v
10 00

/m

/m

re v

re v

00

00

Figure 6.20: Garrett GT-12 compressor map with engine operating points overlaid for varying engine speeds and MAP.

40

60

80

00

0r

ev

/m

in

180

Garrett GT-12 Turbine

65% 60% 55%

6,000 rev/min

50%

W O T

250000 230000 205000 185000 160000 136000

Figure 6.21: Garrett GT-12 turbine map with engine operating points overlaid for varying engine speeds and MAP.

181

0 -20 PMEP (kPa) -40 -60 -80 -100 -120 80 100 120 140 160 180 200 220 240 260 280 Intake MAP (kPa) 4000 rev/min 5000 rev/min 6000 rev/min 7000 rev/min 8000 rev/min 10000 rev/min
WOT

240
WOT

Exhaust MAP (kPa)

200

160

4000 rev/min 5000 rev/min 6000 rev/min

120

7000 rev/min 8000 rev/min 10000 rev/min 80 100 120 140 160 180 200 220 240 260

80 Intake MAP (kPa)

7.0 6.5 6.0 5.5 5.0 4.5 -0.2

6000 rev/min 240 kPa MAP 100 kPa MAP

log (Pa)

0.2

0.4

0.6

0.8

1.2

log (Clearance Volume/Actual Volume)

Figure 6.22: Intake and exhaust pressures and the effects on engine pumping work for varying engine speeds. (Upper): PMEP. (Middle): Exhaust pressure. (Lower): Log pressure-volume, highlighting the PMEP reduction for rising MAP at 6000 rev/min.

182

Furthermore, Figure 6.22 displays the intake and exhaust pressures and pumping work over the engine operating range, giving further insight into the turbocharger matched operation. Over all operating conditions, the pumping work in the upper diagram of Figure 6.22 increases for rising engine speeds at constant MAP. This is caused by an increase in throttling and valve flow work [98] due to higher losses associated with the increased air velocity during induction. However, for fixed engine speeds and rising intake MAP above atmospheric, the trends in pumping work switch with the PMEP increasing at constant speeds above 6,000 rev/min. Below 6,000 rev/min, the pumping work is shown to decrease as the intake MAP increases, as further illustrated in the log pressure-volume diagram (Figure 6.22 Lower). The reduced pumping work for rising intake MAP indicates good matching between the compressor and turbine as previously verified in Figures 6.20 and 6.21 with high compressor and turbine efficiencies. The good matching below 6,000 rev/min is further demonstrated by the intake and exhaust pressures (Figure 6.22 - Middle), with an intake/exhaust pressure ratio > 1 at these speeds. The greater than unity pressure ratio for rising intake MAP has enabled the pumping losses to reduce. The pumping work decreases for these conditions as the compressor provides work to force the piston down during induction, which offsets the losses taken to drive the turbine. Near WOT (240 kPa intake MAP) at 6,000 rev/min, the pumping work is reduced to -20 kPa, a 2.5 fold reduction when compared to engine operation at intake MAP equal to atmospheric conditions (100 kPa). Hence, if the intake MAP is further increased, it is feasible that the pumping losses could be eliminated or perhaps the compressor could produce positive pumping work and hence crankshaft power. losses to increase due to the rising exhaust back pressure. When analyzing the middle diagram of Figure 6.22 at 6,000 rev/min, it is perceived that the compressor should be producing positive pumping work due to the higher intake pressure when compared to the exhaust at WOT. However, this effect is not witnessed in the pumping loop shown in the lower diagram of Figure 6.22. Although there seems conflicting pressures between the intake and exhaust data at 6,000 rev/min from the middle and lower graphs of Figure 6.22, it is noted Above 6,000 rev/min, the less than unity intake/exhaust pressure ratio causes the pumping

183

that the bottom pressure-volume graphs are in-cylinder recorded data. losses through the valves and ports.

The

differences are related to the valve flow work which corresponds to pressure These losses increase as the airflow increases and the critical Mach index value approaches 0.6 [220] (Appendix H.5).

6.7 Summary
This chapter has described the successful implementation of a high pressure ratio turbocharging system to the odd firing two cylinder test engine, enabling near two bar boost pressure to be realized at mid-range speeds. valve events was also described. Turbocharger performance optimization involving intake and exhaust geometry together with In addition, this chapter highlighted the importance of using simulation as an aid in both accelerating engine development and performance optimization, with excellent agreement between experimental and simulated results. Significant turbocharger development involving internal modification was required due to the upstream throttle location and regulated intake restriction. The associated compressor throttling caused oil consumption issues that were compounded by the modern turbocharger design, which featured piston ring type oil seals on the compressor side. The research and development steps to overcome this oil control issue were outlined, together with a final novel solution. This turbocharger development has significant relevance, as it permits the continual use of compressor side piston ring seals, while also allowing the turbocharger to be placed downstream of the throttle. The compressor throttled layout has additional benefits in reducing turbo lag and increasing turbocharger rotational acceleration during engine transience, without the need for costly and sophisticated technologies. No turbocharger problems were evident after internal development, with the TC engine completing extensive periods of both static and transient testing. Experimental and simulation results have clearly shown that the engine performance is not limited by the turbocharger. This is further emphasized as experiments were completed with controllable intake temperatures, without the need for intake air intercooling or evaporative charge cooling in the compressor.

184

Analysis of the compressor performance showed that it was not the limiting factor, with combustion limits reached (Chapter 9) prior to the compressor reaching its flow and hence boost limit. Additionally, the turbocharger mechanical design proved adequate after seal development to suit the compressor throttled application, hence not limiting performance. Due to the odd firing two cylinder configuration and resultant uneven gas flow, the exhaust manifold volume was found to be the most important parameter to optimize, while the intake and valve events had minimal effects in comparison. In addition to the volume, the exhaust geometry effects with both pulse and constant pressure systems were investigated. Results from experiments and simulation have shown significant brake power improvements for this particular odd firing engine when employing constant pressure type turbocharging. This demonstrated that with exhaust geometry optimization, high pressure ratio turbocharging can be successfully applied to any odd firing configuration, overcoming the unequal firing spacing effects. Hence, the intake and exhaust gas flow abnormalities associated with the reduction in the number of cylinders can be overcome for any engine configuration with two or more cylinders, thus not limiting engine performance. Experimental and simulation results have shown that by successfully implementing turbocharging to this flow restricted engine, near constant power can be achieved over half the speed range. This is due to the intake restriction, which limits airflow and thus performance when operating at the choked condition. Furthermore, the test engines TC performance could be further improved beyond what was experimentally shown, with additional opportunities to improve peak power, low speed torque and transient response. Moreover, analysis from the turbocharger matched operation showed poor turbine efficiencies at high engine speeds highlighting potential gains with improved turbine or VGT technology.

185

186

CHAPTER 7
P O W E R E D B Y W A T T A R D

Operating Limits and Experimental Results

7.1 Introduction
This chapter defines the engine limitations (knock and airflow operating limits) associated with the NA and boosted test modes. Cross plots defining the operating limits have been constructed by analyzing the experimental data over the test domain data set. Small engine testing involved varying engine speed, MAP and CR parameters with performance, efficiency and emissions experimental results displayed over all test modes. Experimental results and trends due to parameter variation are also compared to other small engines and typical larger bore engines as found in passenger vehicles. The majority of data presented in this chapter is expressed in either contour or cross plot form. This allows trends to be established and visualized to quantify parameter variation effects. Moreover, plots also allow comparison to larger bore engines to quantify any performance, efficiency or emission effects attributed to the reduced cylinder capacity. As outlined in Table 7.1, data is expressed across two domains in plots, joined by isopleths in dimensional space. In addition, lines are drawn denoted as the performance limit (PL) corresponding to each modes WOT condition for a given CR. Hence plots express: Engine speed versus MAP (at the experimental test CR closest to the HUCR) CR versus engine speed (at the PL or WOT)

187

Table 7.1: Chapter 7 contour and cross plot data expressed over multi-domains with parameter variation. Contour and Cross Plots across all Test Modes Engine Speed CR MAP and Spark Timing Engine Speed vs. MAP plots at CR vs. Engine Speed plots at Varied (4,000 - 10,000 rev/min) Varied (9 - 13) Varied (55 kPa - PL) Varied (Table 4.9) Test CR closest to the HUCR MAP = PL (WOT)

This chapter presents only steady state experimental results, achieved while the engine was installed to the dynamometer test rig as outlined in Chapter 4. Many brake and indicated output parameters were observed or calculated, and as a result of this large data space, numerous contour and cross plots were generated. This chapter presents only the most relevant plots achieved with the test rig.

7.2 Operating Limits

Figures 7.1 to 7.3 display the knock and airflow limitations as functions of engine speed, MAP and CR. The KL and DL have previously been defined in Table 4.8. The cross plots have been constructed from multi-CR experimental data points, gathered by incrementally varying the CR to values dictated by the knock severity. These CR values included 9.6, 10, 11, and 13 (Chapter 5.4.3). displayed in Figure 5.8. The cross hatched areas in Figures 7.1 to 7.3 indicate domains where engine operation was KL but could be controlled via tuning strategies (Table 4.9) to avoid the DL. The shaded areas in plots indicate where engine operation was not Airflow possible due to airflow limitations or heavy knock exceeding the DL. The resulting piston crown and combustion chamber shapes associated with each CR are

limitations were only associated with the boosted modes, where compressor delivered airflow was insufficient to meet the desired operating requirements (Table 4.5) or when the intake restriction limited airflow. The PL is highlighted by the dashed blue line, with variations between modes due to the varying MAP conditions at WOT associated with NA and boosted operation.

188

With knock compensation tuning strategies (varying degrees of spark retard and/or fuel enrichment) together with CR reductions being the only forms of knock control used throughout experimentation, the cross plots displayed must first be put into context. It is noted that further knock reductions are expected with the implementation of various other strategies thus allowing increased MAP and/or CR before exceeding the DL. These strategies could include DI, EGR, intake charge cooling (intercooling and/or evaporative), VVT, combustion enhancement through chamber design and increases in fuel quality [42, 79, 98, 130, 198, 220]. However, these knock preventative strategies were not implemented during experiments due to their added complexity and as such were out of the scope of the project. Figure 7.1 displays the operating limits for both NA modes. No airflow limitations were associated with these modes as maximum MAP was dictated by the intake system (Chapter 4.5). Hence the PL line corresponds to near 100 kPa MAP at WOT for all NA modes. It can be seen that the KL associated with carburetion was not reached over the range of test CRs when compared to PFI. This was partly due to the limitation on the maximum achievable CR, which was limited by the piston design coupled with slightly richer fuel mixtures in the low speed WOT domain (Figure 7.7) when compared to the PFI mode. These mixtures were richer than those defined in the tuning strategy, however accurate control was difficult to achieve due to the carbureted system. Furthermore, the carbureted system also used an intake system with shorter intake runners, when compared to that of the PFI systems (Table 4.4). Thus peak BMEP values occurred at higher engine speeds, where the engine is less susceptible to knock [79, 171]. In the NA - PFI mode, knock levels above the DL were observed in the 6,000 rev/min region at the highest achievable CR of 13. Hence knock control was achieved with reductions in spark timing while keeping fuel mixtures constant.

189

(MODE A) NA - CARBS
100
CR = 13 Limited to 100kPa MAP, CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000 6000 5000 4000

Operating (A) at Performance Limits (WOT) HUCR at WOT

P L

90 MAP (kPa)

80

70

60 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

10

11 CR

12

13

(MODE B) NA - PFI
100
CR = 13 Limited to 100kPa MAP, CR = 13
PL

10000 9000 Speed (rev/min)

Operating at Performance Limits (WOT) (A) HUCR at WOT

P L

90 MAP (kPa)

KNOCK COMPENSATION
RETARD

8000 7000 6000 5000 4000

KNOCK COMPENSATION
RETARD

80

70

60 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

10

11 CR

12

13

Figure 7.1: NA mode knock limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard to compensate for knock. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): NA - Carburetion. (Lower): NA - PFI.

190

Figure 7.2 displays the operating limits for both boosted modes. due to varying boost levels.

The PL line

varies between SC and TC modes since the maximum MAP levels differed at WOT, The desired airflow limits, defined in Table 4.5 involved 150 kPa MAP for supercharging and fixed mass airflow (~0.075 kg/s Appendix D.1) for turbocharging due to the intake flow restriction. However, both these airflow limits could not be achieved over the entire speed range in experiments due to insufficient compressor delivery. In the SC mode, the Roots type supercharger could not supply adequate airflow to maintain 150 kPa MAP above 6,000 rev/min, resulting in the PL line decreasing as the engine speed is increased past this limit. Adequate airflow could not be maintained because of electric motor speed limitations, as the variable speed supercharger was externally driven via mains power supply and not the crankshaft (Chapter 4.5). The varying PL line has resulted in high knock intensities at low engine speeds due to the higher boost levels. Hence, when compared to NA modes, a CR reduction to 11 was needed to avoid the DL. CR reductions were also combined with mild levels of knock compensation to compensate for low knock intensities at these speeds. In the TC mode, compressor delivery effects were reversed, with inadequate supply at low engine speeds due to limitations on turbocharger speed associated with matching (Chapter 6.6). The compressor driven by the exhaust turbine failed to provide enough airflow to choke the flow restriction until 6,000 rev/min. As the speed increases past this limit, MAP levels and therefore the PL line decrease as the airflow is restricted. The highest knock intensities were recorded at 6,000 rev/min, corresponding to the highest MAP value of 270 kPa. High knock intensities were likely in this region due to the high levels of boost needed to achieve sonic flow through the intake restriction. Consequently, further CR reduction was needed when compared to the SC mode. Initially, a CR reduction to 9.6, coupled with mild levels of knock compensation were used to ensure the DL was avoided in this particular region.

191

(MODE C) SC - PFI
150Limited kPa MAP Limited,MAP CR = 11 to 150kPa Operating at Performance Limits (WOT) (A) HUCR at WOT
P L

150 140 130 120 MAP (kPa) 110 100 90 80 70 60

10000 9000 Speed (rev/min) 8000 7000 6000 5000 4000

RETARD AND OR ENRICH

KNOCK COMPENSATION

COMPRESSOR FLOW LIMIT

PL

RETARD AND OR ENRICH

KNOCK COMPENSATION

DAMAGE LIMIT

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

10

11 CR

12

13

(MODE D) TC - PFI
Intake (A) No Flow wastegate Restricted, control CR = 10
260 240 220 200 MAP (kPa) 180 160 140 120 100 80 60 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

10000 9000 Speed (rev/min) 8000 7000 6000 5000 4000

(A) HUCR at WOT Operating at Performance Limits (WOT)

P L
RETARD AND OR ENRICH

KNOCK COMPENSATION

RETARD AND OR ENRICH

KNOCK P COMPENSATION L

DAMAGE LIMIT

10

11 CR

12

13

Figure 7.2: Boosted mode knock and airflow limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard and/or fuel enrichment to compensate for knock. Shaded areas indicate non-operation due to knock levels above the DL or limited airflow. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): SC - PFI. (Lower): TC - PFI.

192

This CR reduction strategy compromised power and efficiency at lower MAP levels and therefore the CR was later increased to 10 to reduce these effects. This increased knock intensities which were counteracted with increasing levels of knock compensation, the highest being at 270 kPa MAP ( = 0.77, KL-ST = 10 BTDC). However, running on the edge of the DL was limited to this particular operating point. This ensured the choked flow FSAE objectives could be achieved over the widest possible speed range, with reduced compromises in power and efficiency at lower MAP levels. Table 7.2 summarizes CR results across the test modes. The estimated HUCR for each mode has been determined from the experimental data presented in the cross plots of Figures 7.1 and 7.2. Some safety margin was assumed in the estimation to ensure the DL was avoided. The estimated HUCR is compared to the experimental test CR, which was determined by analyzing pressure cycle data during engine development. It is noted that slight discrepancies exist between the two due to the inability to easily alter the CR, achieved only through piston crown modifications during experiments. However, these differences are assumed negligible on the performance, efficiency and emissions results presented in this chapter.

Table 7.2: Estimated HUCR and the experimental test CR which was closest to the estimated value. Mode
(A) NA - Carburetion (B) NA - PFI (C) SC - PFI (D) TC - PFI

Estimated HUCR
13.5 13 11.5 9.5

Test CR
13 (piston design limited) 13 (piston design limited) 11 10

193

TC - PFI: 200 kPa MAP wastegate limited

(B) 200 Wastegate kPa MAP limited Limited, to 200kPa CR = 10MAP 260 240 220 200 MAP (kPa) 180 160 140 120 100 80 60 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
TURBO FLOW WASTEGATE LIMIT LIMITED RESTRICTED FLOW LIMIT

10000 9000 Speed (rev/min) 8000 7000 6000 5000 4000

Operating at Performance Limits (WOT) (B) HUCR at WOT

P L KNOCK COMPENSATION

RETARD AND OR ENRICH

RETARD AND OR ENRICH

PL KNOCK COMPENSATION

DAMAGE LIMIT

10

11 CR

12

13

TC - PFI: 150 kPa MAP wastegate limited

150 kPa MAP Limited, CR = 10MAP (C) Wastegate limited to 150kPa
260 240 220 200 MAP (kPa) 180 160 140 120 100 80 60 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

10000 9000 Speed (rev/min) 8000 7000 6000 5000 4000

Operating at Performance Limits (WOT) (C) HUCR at WOT

P L

WASTEGATE LIMITED PL

RETARD AND OR ENRICH

KNOCK COMPENSATION

DAMAGE LIMIT

10

11 CR

12

13

Figure 7.3: TC - PFI mode (boost controlled) knock and airflow limitations versus engine speed, MAP and CR. Cross hatched areas indicate operation with spark retard and/or fuel enrichment to compensate for knock. Shaded areas indicate non-operation due to knock levels above the DL or limited airflow. PL is the performance limit line defined at each modes WOT for a given CR. (Left): Engine speed versus MAP domain. (Right): CR versus engine speed domain. (Upper): 200 kPa MAP limited. (Lower): 150 kPa MAP limited.

194

Figure 7.3 displays cross plots of the TC mode with wastegate limited MAP values. This produces a constant PL line when not limited by airflow. The purpose of displaying these plots is for the reader to conceptualize knock levels associated with varying CRs at constant MAP values. The effects of limiting MAP can be seen, with the possibility of optimizing the CR over a wide speed range at the PL instead of a narrow range as previously shown due to the flow restriction. Hence, it is advantageous to optimize the CR to a fixed value, resulting in incremental levels of rising MAP for engine speed increases to maximize power and efficiency, while continuously avoiding the KL. Wastegate limiting the MAP from 200 kPa to 150 kPa results in reduced knock intensities for a given CR. This demonstrates the potential to further increase CR before exceeding the DL, with specific MAP levels defining the HUCR. At 150 kPa MAP, the HUCR is shown to be 12 compared to 11 for the 200 kPa MAP case. These CR values are relatively high when compared to larger bore engines with similar boost levels [66, 143, 174, 249]. Trends from Figures 7.1 to 7.3 confirm that for a given MAP and CR condition, the engine is less susceptible to knock at higher speeds [12, 14-16]. The reduced knock likelihood is a consequence of the increased flame speeds within the combustion chamber (Chapter 9), which consume the unburnt mass in the endgas region more quickly. Increasing flame speeds decreases the likelihood of knock due to the reduced end-gas residence time within the combustion chamber. It is also interesting to note that spark retard and/or fuel enrichment can be used as methods of knock control for up to 1-2 CR points depending on the knock severity. CR. This can be a useful tool during boosted engine development, as it allows the calibrator some margin to avoid the DL without having to reduce the Moreover, this method of knock control allows the operating limit to be extended, thus enabling increased performance. However, the possible efficiency increases due to potential CR and/or MAP increases were found to be offset by the increased enrichment and less optimum spark timing. The heavy fuel enrichment near WOT could cause high piston and cylinder wear rates associated with cylinder bore wash [200], however this effect was not

195

witnessed in experiments. Emission reductions at WOT would also be difficult to achieve as the heavy enrichment would significantly reduce TWC efficiencies, limiting the exhaust aftertreatment effectiveness. However, there is no requirement for emission control near WOT in present standards (Appendix C.2) so concern is about avoiding catalyst overheating with mixtures just rich of stoichiometric. Thus the present control strategy should be satisfactory as only low boost levels are needed over the NEDC (Chapter 8.8), where engine operation is closer to stoichiometric conditions. The relationship between engine speed, MAP and CR parameters and their effect on the operating limits have been described for this small engine. Experiments highlighted that spark knock is the most dominant factor in limiting the performance of this small engine. However, reductions in end-gas knock severity at similar MAP and CR has enabled the performance operating limits to be extended for small engines when compared to larger units. This is demonstrated with a peak BMEP value of 2,490 kPa, recorded while running in the TC mode at 270 kPa MAP and a CR of 10. This exceeds the highest specific output in recent evaluation of TC engines, newly reported to be 2,200 kPa from General Motors (GM) 2.0 L TC Ecotec engine [16, 198]. The experiment BMEP value has greater significance in that the GM value was achieved with intercooling, DI, dual scroll turbocharging, VVT and sodium cooled valves, which are all advantageous in increasing performance, reliability and reducing knock severity. The increased performance of this particular small engine is attributed to the physical size reduction, particularly the reduced bore size (when compared to passenger vehicles) and the fast burn combustion chamber. This results in engine speed increases together with end-gas volume reductions around the periphery of the chamber, allowing CR and/or MAP values to be increased before exceeding the DL. This has major benefits in extending the operating limits in downsized applications.

7.3 Contour Plot Generation

Throughout the experimental period, it was unfeasible to maintain precise engine speed and MAP values during day to day operation. As a result, small variations in

196

prescribed operating conditions meant that the raw data needed to be interpolated to a rectangular grid. Hence, contour plots were created by applying a surface algorithm over the experimental data points in 3-dimensional space. This was achieved using Sigmaplot [218], a modern graphing software package. Out of the seven smoothing algorithms available in Sigmaplot, the Loess smoothing algorithm with tri-cube weighting and polynomial regression was found to be best suited to the wide range of experimental data points. A major concern over smoothing is the error which can arise between the actual and the generated surface. A sensitivity analysis found that the contour plots presented in this chapter predominantly hold an error of less than 5%. Appendix O further outlines the error analysis. However, values as high as 10% can exist in extreme conditions. These errors are considered insignificant for the qualitative investigation when analyzing trends, with care taken when directly comparing results between modes. Experimental data presented in this chapter is displayed in brake format. Brake data is considered to be more relevant for trends and comparative analysis as the frictional losses are considered, which can vary between modes depending on varying combustion and inertia loading. The possibility of errors which can exist with indicated data is also eliminated, with brake results being relatively easier to quantify. The multi-cylinder configuration and constraints limiting in-cylinder This was due to the performance pressure analysis to the first cylinder also made it unfeasible to use indicated data to describe the total engine performance. discrepancies between cylinders associated with the unequal firing (Chapter 8.3). Using brake data also reduces the assumptions made when correlating to legislative emission standards. Contour plots are displayed with shaded areas, indicating where engine operation was not possible due to the previously described airflow limitations. However, contours are extrapolated into these regions so the reader can deduce expected results if the limitations were removed. PL is the performance limit line defined at each modes WOT for a given CR and is highlighted by the dashed blue line.

197

7.4 Performance Contours

7.4.1 BMEP
Figure 7.4 displays BMEP contours for NA, SC and TC modes. When compared to larger bore engines, trends show matching BMEP effects for varying MAP and CR [11, 25, 98]. Increases in BMEP are shown to be directly proportional to increases in MAP for all engine speeds, primarily due to increases in air consumption and pumping loss reductions. For all PFI systems, for a given MAP unlimited by airflow, peak BMEP occurs in the 7,000 rev/min region, which corresponds to the tuned intake speed. For the carbureted system, peak BMEP occurs at a higher speed due to the shorter inlet tracts (Table 4.4). For all modes, the BMEP falls quickly above the tuned intake speed as the induced airflow and MECH decrease due to the increased frictional losses associated with the higher speeds. Increases in CR are shown to directly correlate to increases in BMEP for all speeds. The increases in BMEP are attributed to a combination of both increased air consumption and improved combustion. As the CR is increased, the residual gas fraction within the cylinder decreases. Reducing the hot products within the cylinder minimizes the warm-up of the incoming charge during the induction process, improving charge density and air consumption. A bulk increase in charge density also enhances the combustion process with improved burning rates. A BMEP improvement in the order of 10-13% was recorded across both carburetion and PFI with a CR increase from 10 to 13. These values closely correspond to experimental results recorded in larger bore engines together with fuel-air cycle analysis [67, 68, 98, 220]. crown geometry constraints. For the NA modes, a peak BMEP of 1,260 kPa was recorded at 8,000 rev/min with carburetion. This is a 9% peak improvement over the PFI system due to intake restrictor effects (Chapter 8.3). For boosted operation, peak BMEP values of 1,590 kPa and 2,490 kPa were recorded with the SC and TC modes. The performance discrepancies are attributed to the varying MAP and accommodating CR, needed to suppress knock below the DL to ensure engine reliability. Decreasing BMEP values are evident as the engine becomes flow restricted for rising engine speeds. However, further performance improvements associated with CR increases are primarily limited by knock together with piston

198

100

BMEP (kPa), CR = 13
1200 1100
PL

10000 9000 Speed (rev/min) 8000 7000

BMEP (kPa), MAP = 100kPa

1000 1050 1100 1050 1100 1150 1150 1200
P L

90 MAP (kPa)

1000 900

1250 1200 1150

80

800 700

70

600 500

6000 5000

1100 1050 950 1000

60

200

300

400

NA - CARBS

4000 9

NA - CARBS
11 CR 12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

10

100

BMEP (kPa), CR = 13
1100
PL

10000 9000 Speed (rev/min) 8000 7000 6000 5000

BMEP (kPa), MAP = 100kPa

800 850 900 950 1000
P L

90 MAP (kPa)

1050

1000 900
1000 1050 1100

1100

80

800 700

1150 1150 1050 1000 850 900 950 1100

70

600 500

60

200

300

400

NA - PFI

4000 9 10 11 CR

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140 130 120 MAP (kPa) 110 100 90 80 70 60
200
P L

BMEP (kPa), CR = 11
1600 1500 1400 1300 1200 1100 1000 900 800 700 600 400
COMPRESSOR FLOW LIMIT

BMEP (kPa), CR = 10 260 240 220 200 MAP (kPa) 180 160 140 120 100 80 60
TURBO FLOW LIMIT

2500

2350 2200 2000 1800 1600 1400 1200 1000 800 600 300
PL RESTRICTED FLOW LIMIT

SC - PFI

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.4: BMEP trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

199

7.4.2 Brake Power

Figure 7.5 displays brake power results for varying engine speed, MAP and CR parameters. Contour plots show identical trends to larger bore engines [25, 98, 120, 220]. For all speeds, brake power increases are a direct result of increasing MAP. This is expected as brake power is proportional to BMEP and engine speed. For all test modes at constant MAP, brake power increases are associated with increasing speed. However, at the higher speeds, the rate of power increase begins to level off as reductions in BMEP are offset by increasing engine speeds. Increases in CR are directly shown to improve brake power due to the increase in BMEP for fixed speeds. However, MAP is shown to have a greater influence on brake power when compared to CR effects. In the NA - PFI mode, a brake power reduction in the region of 50% is evident as a result of decreasing the MAP by 30% from the PL line. In contrast, a 30% reduction from the HUCR results in brake power losses on the order of 10%, several magnitudes lower. This indicates that it is advantageous to employ higher MAP (intake boosting) when compared to CR increases if significant brake power improvements are required. This is the case in downsized engine applications, enabling smaller engines to be comparable to their larger counterparts. For both NA modes, peak brake power near 35 - 40 kW is recorded at maximum engine speed. This is not the case in the boosted modes as maximum airflow is maintained over a varying speed range due to the airflow limitations. This results in near constant power contours over these regions at the PL line. This indicates that brake power is heavily dependent on airflow. In the TC mode, extrapolated contours show that 200 kW/litre could be achieved at maximum MAP and engine speed with the removal of the intake flow restriction. Higher power values could also be achieved with boost reductions at mid-range speeds, enabling increased MAP and/or CR at higher speeds where knock is less susceptible (Section 7.2). This high specific output suggests a 1.0 L TC version, achieved with a four cylinder configuration of the current engine design would exceed the engine performance quoted for the 3.5 L Holden Commodore [100] and 4.0 L Ford Falcon [72]. as described in Chapter 8.8. Hence, the potential exists to replace larger engines fitted to Australian family vehicles, with obvious advantages

200

100

Brake Power (kW), CR = 13

PL

41 38 35 32

10000 9000 Speed (rev/min) 8000 7000 6000 5000

Brake Power (kW), MAP = 100kPa

40 P L

90
29

36 34 32 30 28 26 24 22 20 18 16

38

MAP (kPa)

26

80
17 14 11 8 5

23 20

70

60

NA - CARBS

4000 9

14

NA - CARBS
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

10

11 CR

100

Brake Power (kW), CR = 13

PL

33

10000 9000 Speed (rev/min) 8000 7000 6000 5000

Brake Power (kW), MAP = 100kPa

34 P L 30 28 26 24 22 20 18 16 14 32 30 28 26 24 22 20 18 16 PFI

30

90 MAP (kPa)
24

27 21 18 15

80

70
9 6 3

12

60

NA - PFI

4000 9 10

NA
11 CR 12

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140 130 120 MAP (kPa) 110 100 90 80 70 60
1 4 7 34 31 28 25 22 19 16 13 10
P L

13

Brake Power (kW), CR = 11

42 40 COMPRESSOR FLOW LIMIT 37

Brake Power (kW), CR = 10 260 240 220 200 MAP (kPa) 180 160 140 120 100 80
TURBO FLOW LIMIT

65 74 RESTRICTOR 66 60 FLOW LIMIT 55 58 52 50 45 46 40 35 30 25 20 15 10

PL

SC - PFI

60

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.5: Brake Power trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

201

7.4.3 Spark Timing

All spark timing data points displayed in Figure 7.6 were recorded in accordance with the tuning strategy (Table 4.9) to ensure a compromise between performance and efficiency while maintaining engine reliability. Thus the spark timing was varied to achieve minimum spark advance for maximum brake torque (MBT-ST). MBT-ST is usually limited by the onset of knock at high MAP and/or high CR values and thus the spark timing was retarded (KL-ST) as combustion became knock limited. Trends from Figure 7.6 show that as the speed increases, for all non-knock limited conditions, the spark timing must be advanced to maintain MBT-ST since the combustion burn duration in CA degrees increases. In the NA modes, reductions in spark advance values as the MAP or CR increase are results of decreases in burn duration. Decreases in burn duration are caused by increased flame speeds, as turbulence levels increase (Figure 9.6) and charge density within the cylinder improves as a result of the reduced residual gas content (Chapter 9.4). Spark timings over the test domain for both NA modes are higher relative to larger bore engines [11, 25]. However, spark timing trends for parameter variation closely follow the literature [98, 220]. Spark timing variations between small and large cylinder engines are primarily caused by different operating speed ranges associated with the varying bore size and bore/stroke ratio. Furthermore, smaller bore NA engines generally operate with little, if any spark retard for knock compensation, due to the reduced propensity for knock to occur. This is highlighted by a modern larger bore engine requiring a KL-ST of 5 BTDC at low engine speeds (2,000 rev/min) to ensure engine reliability, which subsequently compromised efficiency [148]. In the boosted modes, further reductions in spark advance as the MAP increases are results of knock compensation (KL-ST) in order to avoid the DL and ensure engine reliability. The severity of knock compensation is evident at 6,000 rev/min in the TC mode, with extreme KL-ST values at 270 kPa MAP. The faster burning nature of the boosted modes also results in later spark timing (Chapter 9.5) when compared to the NA modes at a given operating condition.

202

100

MBT (deg. BTDC), CR = 13

27 28 29 30
PL

10000 9000 Speed (rev/min)

MBT (deg. BTDC), MAP = 100kPa

44 42 40 38 36 30
P L

90 MAP (kPa)

31

8000 7000 6000 5000

80
32

34

32

70

33 34

28

60

NA 36 - CARBS

35

4000 9 10 11 CR

NA - CARBS
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100

MBT (deg. BTDC), CR = 13

27 28
PL

10000 9000 Speed (rev/min)

MBT (deg. BTDC), MAP = 100kPa

44 35 42 34 40 33 38 32 36 31
P L

90 MAP (kPa)

80

29 30 31 32 33

8000 7000 6000 5000

34 29 30 32

28 30 27 28 26 27

70

34 35

60

36 NA PFI

4000 9 10 11 CR

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140 130 120 MAP (kPa) 110 100 90 80 70 60
30 32 35 34

MBT (deg. BTDC), CR = 11

20 22
P 24 L 26 COMPRESSOR FLOW LIMIT

MBT (deg. BTDC), CR = 10 260 240 220 200 MAP (kPa) 180 160 140 120 100
36 38
TURBO FLOW 12 LIMIT RESTRICTED FLOW LIMIT

14 16 18 20 22 24 26 28 30 32
PL

28

80 60

SC - PFI

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.6: Spark timing trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

203

7.5 Efficiency Contours

Experiments were completed with varying across the test matrix (Chapter 4.6.4), with desired fuel mixtures set in accordance with the tuning strategy (Table 4.9). Hence, for comparative purposes when noted, efficiency data from experiments has been corrected to stoichiometric conditions enabling trends to be examined. The methodology behind the correction procedure is outlined in Appendix M.6.

7.5.1 Lambda
Figure 7.7 displays experimental values for varying MAP and CR. When

comparing NA modes, which required no fuel enrichment for knock control, a higher spread away from the pre-defined tuning strategy is evident in the carbureted system. Accurate control was difficult to achieve in this system due to difficulties in tuning both Mikuni carburetors over all operating conditions. Nevertheless, near mixture strength was achieved with carburetor modification, involving re-jetting and metering needle modification. As the rate of fuel induced in the carbureted mode was dependent on the induced mass flow of air, no recalibration was required for different CRs. Although mixture control was difficult to achieve in the carbureted system, minimal variation existed between both cylinders when compared to the PFI modes. This was due to airflow discrepancies between cylinders in the PFI modes associated with the intake flow restriction and the odd firing interval (Chapter 8.3). Trends from Figure 7.7 and brake data from Figure 7.5 show that for this particular small engine, peak power occurs at ~ 0.9 in regions which are not knock limited. NA experimental results also found minor power variations (< 2%) when was varied between 0.86 - 0.93 (Figure M.2). Both boosted modes display high levels of enrichment near the PL, which was needed for knock compensation, with quantities defined in contour plots. Increasing levels of enrichment were needed for the rising MAP which significantly reduced TH. However, combustion stability could still be maintained at these rich fuel mixtures (Figure 9.9), due to the extension of the rich limit caused by boosted operation increasing in-cylinder pressures and temperatures.

204

100
0.88

Lambda, CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000 6000 5000

Lambda, MAP = 100kPa

0.93
P L

90 MAP (kPa)

0.94 0.91

0.93 0.90 0.86 0.88

80
0.97

70

1.00

60

NA - CARBS

1.05

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 Lambda, CR = 13
PL

4000 9 10 11 CR

NA - CARBS
12 13

10000 9000 Speed (rev/min) 8000

Lambda, MAP = 100kPa

P L

90 MAP (kPa)

0.90

0.93

80
0.96

0.90 0.89

7000 6000 5000

70

1.00

60

1.10

1.05

NA - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60
1.10 1.05 0.88 0.92 0.96 1.00
P L COMPRESSOR 0.82 FLOW LIMIT

4000 9 10 11 CR

NA - PFI
12 13

Lambda, CR = 11 260 240 220 200 MAP (kPa) 180 160 140 120 100 80

Lambda, CR = 10
TURBO 0.77 FLOW LIMIT RESTRICTED FLOW LIMIT

0.84

0.79 0.81 0.83 0.85 0.88 0.92 1.15 1.08 1.00

PL

SC - PFI

60

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.7: fueling requirements versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

205

7.5.2 BSFC and Brake Thermal Efficiency

Both the raw brake specific fuel consumption (BSFC) and TH contours for the varying modes are displayed in Figure 7.8. It is noted that the variations (Figure 7.7) make it difficult to compare results to previous published data. However, uncorrected results are shown for the reader to digest efficiency trends for this particular tuning strategy. For all modes, peak TH occurred in the 80-90 kPa MAP, 5,000-6,000 rev/min (10 m/s MPS) region as a result of high MECH, low pumping losses and fuel mixtures which were nearer to stoichiometric. In the NA modes, peak efficiency values are compared at similar with carburetion achieving 32% and PFI 34%. However, further improvements were expected from intake charge boosting, particularly turbocharging [12, 148, 174, 191, 248, 249]. These results did not eventuate, largely due to the excessive increases in fuel enrichment and spark retard needed to control knock for the rising MAP (Figures 7.6 and 7.7). At high load conditions, the NA modes have a BSFC advantage due to the higher CR, less enrichment and more optimum ignition timing [250]. However, if knock could be controlled via other methods previously defined (Section 7.2), TC improvements in BSFC and TH are expected. Consistently, across the speed range for all modes, BSFC and TH results were poor at low MAP due to the high residual gas fraction, low MECH and high pumping losses. Near the PL (WOT), efficiency results were also poor as peak power was targeted in the NA modes and knock suppression in the boosted modes. Furthermore for boosted operation, the higher cycle pressures and hence temperatures associated with the rising MAP cause higher levels of dissociation and heat losses, which also reduces TH. As the engine speed increases, efficiency results were also poor due to dissociation, friction and heat losses. BSFC trends due to CR changes show efficiency improvements for increasing CRs. In the NA modes, increases of the order of 13% in relative efficiency can be achieved by increasing the CR from 9 to 13 across all speeds at WOT. These results correlate to previous published data for larger bore engines [11, 25, 97, 161, 220, 223]. These results contribute to the majority of BMEP performance improvement associated with the CR increases of Figure 7.4 and 7.5.

206

100

BSFC (g/kWh) and TH (%), CR = 13

32.5%
PL

BSFC (g/kWh) and TH (%), MAP = 100kPa 10000

310

90 MAP (kPa)
245 250 265 280 300 320 350

30.7%

9000
300

300

290

280 270

P L

Speed (rev/min)

8000 7000

290

280

260 31.2%

80

27.1%

29%

260 270

70

6000
27.1%

23.2%

60
450

5000
400

290

300 310

280

NA - CARBS

4000 9

NA - CARBS
10 11 CR 12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 BSFC (g/kWh) and TH (%), CR = 13
P L 290

BSFC (g/kWh) and TH (%), MAP = 100kPa 10000 9000 Speed (rev/min)
315 310 300
27.1%

90 MAP (kPa)

33.8% 240 245 250 31.2% 275 290 28% 260

290

P L

280 270

8000
290 280 270
31.2% 29%

80

7000 6000 5000 4000 9 10 11 CR

280 290 270

260

70
320 350 400 23.2%

250 32.5% 260

60

NA - PFI

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110
29% 300 P 280

BSFC (g/kWh) and TH (%), CR = 11 260

360
COMPRESSOR 330 FLOW LIMIT L

BSFC (g/kWh) and TH (%), CR = 10 240 220 200 MAP (kPa) 180 160 140 120 100 80 60
300 29% 280 360 400 27.1% 300 340 320
TURBO FLOW LIMIT RESTRICTED FLOW LIMIT 360

360 340 23.9%

PL

100 280 90 80 70 60
300 330 360 400 31.2% 260 270 27.1%

SC - PFI

TC400 - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.8: BSFC trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

207

Figure 7.9 displays BSFC and TH contours corrected to stoichiometric fuel conditions to allow efficiency comparisons to other engines irrespective of the fuel mixture. Further correction details are given in Appendix M.6. The corrected results confirm previous fuel efficiency findings. Once again, the fuel efficiency benefits expected from turbocharging are not realized. This is due to the inability to correct for the significant reductions in spark timing needed to avoid the DL for the rising MAP. However, TC BSFC corrected results show the potential to match previous efficiency trends if other knock preventive strategies besides spark retard are used. In the NA modes, a corrected peak brake TH of 37% was calculated with PFI at the highest test CR of 13. A 3% lower TH is also observed for carburetion, likely due to the poor mixture homogeneity due to reduced atomization and fuel breakup. These efficiency results are slightly higher than those found in similar small engines [124, 162], with the improved efficiency associated with high MECH (Figure 7.11) and low parasitic losses. Other factors include enhanced combustion (Chapter 9.4) due to the rapid burn combustion chamber which extended the KL, thus allowing increases in CR when compared to other small engines (Chapter 8.6). These contributing factors allowed the engine package to achieve comparatively high BMEP values (Table 8.3). It is also recognized that further potential exists for improved BSFC and TH using well documented methods such as EGR, lean operation and DI [98, 220]. NA peak TH values are also compared to modern, larger bore automobile engines operating at stoichiometric conditions. Generally, typical peak efficiency values for modern automobile engines range between 32-34% [25, 31, 72, 98, 100, 220]. Smaller bore engines inherently suffer from increased heat losses when compared to larger engines due to the higher surface to volume ratio [95], which results in reducing small engine efficiencies. However, the test engines improved efficiency is attributed to the extension of the KL due to the reduced bore size and fast burn combustion chamber. This allowed for CR increases, which offset the higher dissociation, heat and friction losses associated with the reduced bore size.

208

100

BSFC corr to = 1 (g/kWh), CR = 13

36.1%
PL

BSFC corr to = 1 (g/kWh), MAP = 100kPa 10000

300 290 280 270 260

90 MAP (kPa)

220 225

9000 Speed (rev/min)

31.2%

P L

8000 7000 6000

240
33.8%

250 240

80

230 240

260 255 270 280

70

300 325 350 400 450

27.1% 23.2%

260
31.2%

250

230
35.3%

60

5000 260

250

240

NA - CARBS

4000 9 10 11 CR

NA - CARBS
12 13

230

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100

BSFC corr to = 1(g/kWh), CR = 13

36.9%
PL

BSFC corr to = 1(g/kWh), MAP = 100kPa 10000 9000 Speed (rev/min)

290 280 270 250 260 240 250
33.8%
P L

90 MAP (kPa)
230 240

220

33.8%

8000 270 260 7000 6000

260
31.2%

240 230

230 225

80

260

29%

280

36.1%

70

300 325 350 375 400 450

225 250 240 230

23.2%

60

5000

NA - PFI

270

4000 9 10 11 CR

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140

BSFC corr to = 1 (g/kWh), CR = 11

31.2%

BSFC corr to = 1 (g/kWh), CR = 10 260

280 290
RESTRICTED FLOW LIMIT

260
P L COMPRESSOR 275 FLOW LIMIT

240 200 MAP (kPa) 180 160 140 120 100

130 260 120 MAP (kPa) 110 100 90 80 70 60

250

TURBO 280 FLOW 220 LIMIT

34.6%

260

235 240 250 275 300 350 400 450

27.1%

29% 31.9%

PL

255

260

270

280 290

SC - PFI

80 60
450

320 350 400

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.9: Corrected BSFC trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

209

7.5.3 Volumetric Efficiency

As can be seen from Figure 7.10, VOL increases are caused primarily by the rising MAP for a given engine speed. For the boosted modes, the VOL contour lines begin to diverge as the MAP is increased past atmospheric conditions. This is associated with decreasing air charge density, which is caused by the rising MAT associated with compressor delivery. This effect could be minimized with the use of intercooling and/or fuel evaporative charge cooling, which would allow further increases in MAP and/or CR. However, these strategies were not implemented during testing (Chapter 3.7). Trends from CR variation in NA modes show minor increases in VOL as the CR is increased from 9 to 13. The improved cylinder filling is caused by the reduced heating effect from the lower residual gas fraction, which increases the trapped charge density. range. Trends from all modes suggest that for a MAP above 80 kPa, air consumption rates are highly dependent on intake manifold resonant tuning. For the NA modes, the pressure waves in the inlet manifold begin to form, increasing in intensity to become most dominant at WOT. This resonant wave, when correctly timed in the induction process, improves air consumption and reduces the residual gas fraction within the cylinder [246] (Chapter 3.7). documented in NA systems (Figure 3.29). Resonant wave instantaneous pressures in the order of 180 kPa absolute have previously been In boosted applications, resonant tuning is generally less important as the engine airflow can be increased using the compressor, with the higher intake pressures naturally causing through scavenging during valve overlap to reduce the residual gas content. Results from experiments show identical air consumption trends to larger bore engines [25, 98, 220]. With resonant tuning, peak VOL values near 100% have been achieved in both NA modes. Sufficient air trapping is also evident across all modes at mid to high range speeds due to the narrow valve overlap camshafts (Chapter 6.5), which provided good scavenging but avoided excessive flow losses. This change in VOL is only minor, with a 1-3% variation corresponding to a 50% decrease in the residual gas fraction over the test CR

210

100

VOL (%), CR = 13
PL

105 108 100 103 90

10000 9000 Speed (rev/min) 8000 7000 6000 5000

VOL (%), MAP = 100kPa

104 104 107 107 104 101 98 104 95 100 92 98 89 86 95 83 88

90 MAP (kPa)
80

109
P L

80
70

70
60

60
40

50

NA - CARBS

4000 9 10 11 CR

NA - CARBS
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100

VOL (%), CR = 13
95 96 90
PL

10000
96

VOL (%), MAP = 100kPa

96 98 92 95 100 102 97 102 100 98 95 96 92 98 92 P 100 95 L 102 97 98 104 102 97 100 98 95 96 92 93 88

90 MAP (kPa)
80

9000 Speed (rev/min) 8000 7000 6000 5000

85

80
70

70

60 50

93 88 90 86 80 90

60

40

NA - PFI

4000 9

NA -80 PFI 86
10 11 CR 12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60
40
P L 130 120 110

VOL (%), CR = 11
150 140
COMPRESSOR FLOW LIMIT

VOL (%), CR = 10
260 240 220 200 MAP (kPa) 180 160 140 120 100 80
TURBO FLOW LIMIT

240 220 200 180 160 140 120 100 80 60

P L RESTRICTED FLOW LIMIT

100 90 80 70 60 50

SC - PFI

60

40

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.10: VOL trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

211

7.5.4 Mechanical Efficiency

Figure 7.11 displays MECH and exhaust pressure trends for the TC - PFI mode. The MECH results show similar trends to other small and large engines [120, 259261], with improvements recorded for increasing MAP due to the pumping loss reduction (Figure 6.22). Moreover, MECH reductions are recorded for increasing speeds due to the increased friction and pumping losses, further magnified by the increasing exhaust pressure at high speeds due to poor turbine matching (Chapter 6.6).
MECH (%)
260 240 220 MAP (kPa) 200 180 160 140 120
P

Exhaust Pressure (kPa) 260

300 275 RESTRICTED FLOW 250 LIMIT 225 200 175
P L

93 TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140

TURBO FLOW LIMIT

93 90

87 84 81 78

150 125

120

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100 100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.11: MECH and exhaust pressure versus engine speed and MAP. Shaded areas indicate airflow limited regions. PL is the performance limit line. TC - PFI, CR = 10.

7.6 Emissions Contours

Raw engine-out emissions were determined from individual cylinder exhaust gas products, with Appendix M outlining sampling and analysis techniques. It is noted that governments regulate emission control in terms of vehicle measured tailpipe emissions relative to distance traveled (g/km or g/mi) [43]. Hence, small engines fitted to small vehicles require reduced aftertreatment clean-up to satisfy emission standards (Appendix C.2). Nevertheless, engine-out brake specific emission results are presented to allow comparisons between modes and other engines, irrespective of vehicle and emission treatment strategies.

212

0.16 Mole fractions CO, CO 2 2, O 2 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0.7

RICH

LEAN

CO CO2 O2

0.8

0.9 Lambda ( )

1.0

1.1

1.2

550 HC concentrations (ppm C6)6)

RICH
450 350 250 0-20 kW 150 50 0.7 0.8 0.9 Lambda ( ) 1.0 1.1 20-40 kW 40-60 kW

LEAN

1.2

3000

RICH
NO X concentrations (ppm) 2500 2000 1500 1000 500 0 0.7 0.8 0.9 Lambda ( ) 1.0 1.1 0-20 kW 20-40 kW 40-60 kW

LEAN

1.2

Figure 7.12: Engine-out raw emission concentrations for varying and power outputs, achieved with engine speed and MAP variation. TC - PFI, CR = 10.

213

The documented variability over the test matrix due to the tuning strategy and individual cylinder variation increases the complexity of the emission analysis due to the emissions formation high dependence on (Figure 7.12). However, raw results are shown together with corrected emissions when noted. For comparative purposes, individual cylinder emissions data from experiments has been corrected to alternate constant values ( = 0.9,1) to observe the engine speed, MAP and CR effects. Appendix M.6. Varying correction processes depending on the emission type have been used to compensate for the variation, as outlined in

7.6.1 BSHC
Figure 7.13 displays raw brake specific hydrocarbon (BSHC) data across the range of test modes. It is difficult to analyze trends and make comparisons, due to large variations over the domain caused by combining and BSFC variations. However, effects at constant power contours for this particular engine show increasing HC fractions as the fuel is enriched, as confirmed by the middle diagram of Figure 7.12. In the NA modes, HC comparison between PFI and carburetion is made for this particular tuning strategy due to relatively similar BSFC and over the test domain. Trends show that the PFI system has clear advantages in reducing HC emissions when compared to carburetion. This is due to the improved fuel breakup and atomization which helps vaporize the liquid fuel. This minimizes the wall wetting and pooling along the length of the inlet tract [11, 95, 98, 211]. When comparing NA to boosted modes, raw BSHC trends provide conflicting results, with opposing trends due to the high levels of enrichment needed for knock control in the boosted modes. efficiency near the PL. HC correction to stoichiometric conditions along constant power contours allows engine speed, MAP and CR trends to be visualized without the varying effect. Figure 7.14 displays corrected results, with the correction procedure outlined in Appendix M.6. In the NA modes, corrected HC results from steady state testing confirm previous findings, with HC reductions in the order of 20% at WOT when implementing PFI over carburetion. This inherently increased HC fractions and decreased

214

100

BSHC (g/kWh), CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000

4.5

BSHC (g/kWh), MAP = 100kPa

P L

90 MAP (kPa)
5.7 5.8 6.0 6.5

4.8

5.1

5.4

80

5.1

5.4

5.7

70

7.0 9.0 7.5 8.0

6000 5000
5.4 5.7 6.0

60

NA - CARBS

4000 9 10 11 CR

NA - CARBS
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 BSHC (g/kWh), CR = 13
5.5 4.5
PL

10000 9000 Speed (rev/min) 8000 7000 6000 5000

3.4

BSHC (g/kWh), MAP = 100kPa

P L

90 MAP (kPa)

80
5.5 4.5

3.6 3.8 4.0 4.2 4.4

70

4.8 5.0

4.5 5.0

60

4.0 5.5
6.0 3.3

NA - PFI

4000 9 10 11 CR

NA - PFI 4.8
12 13

4.6

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140 130 120 MAP (kPa) 110 100 90 80 70 60
3.5 4.0 4.0 4.0
P L

BSHC (g/kWh), CR = 11
5.0 5.0
COMPRESSOR 4.5 LIMIT FLOW

BSHC (g/kWh), CR = 10
5.5

260 240 220 200 MAP (kPa) 180 160 140 120 100
3.5 3 2.5 2 4.5 5
PL

TURBO FLOW LIMIT6

RESTRICTED FLOW LIMIT

SC - PFI

5.0

80 60

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.13: BSHC trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

215

Moreover, PFI HC reductions are more obvious at part load due to the reduced airflow and absence of resonant waves in the inlet tract, which tend to break up the larger fuel droplets associated with carburetion. This validates published data [211] and is one of several reasons for the universal adoption of PFI in larger engines. PFI use in small engines is not as prominent due to less stringent emission regulations among other factors. When comparing PFI between NA and TC modes at constant MAP in Figure 7.14, HC emissions are reduced with the implementation of turbocharging, with 30% reductions at low MAP. Hence, TC engines produce lower HC emissions due to the improved exhaust burn-up. The improved burn up is a result of higher exhaust pressures and temperatures associated with the turbine, assuming there is sufficient excess oxygen in the exhaust to complete the oxidation. Trends from raw and corrected results in Figures 7.13 and 7.14 show increases in BSHC emissions with CR increases across all speeds. This is due to varying surface to volume ratios and the increased importance of crevice volume effects at high CR. These increasing HC effects are difficult to avoid due to the premixed charge being forced into the crevices as the CR increases. fractions are offset by reducing BSFC for CR increases (Figure 7.8). As the engine speed increases for fixed MAP values, the corrected contours show decreasing HC trends due to increasing wall temperatures and changes in the incylinder and exhaust oxidation rate. hence reduces HC emissions. engine speeds. Increasing wall temperatures due to increasing combustion temperatures at high MPS increases the heat flux and This offsets the reduced burn up times at high The effectiveness of unburned hydrocarbon oxidation is Engine-out HC fractions are higher than those depicted in the BSHC contours, as increasing raw

significantly enhanced with speed increases as the expansion stroke and exhaust gas temperatures substantially increase [98]. As load or MAP increases for constant speeds, HC concentrations also decrease for similar reasons involving increased burn up due to higher gas temperatures. Therefore, specific HC emissions are shown to decrease as power is increased for all conditions across all modes.

216

100

BSHC corr to = 1 (g/kWh), CR = 13

PL

BSHC corr to = 1 (g/kWh), MAP = 100kPa 10000 9000

P L

90 MAP (kPa)

4.6 4.8

Speed (rev/min)

5.0

8000 7000 6000 5000

3.6 3.8 4.0 4.2 4.4

80
6.0

70

7.0 8.0

60

9.0

NA - CARBS

4000 9 10 11 CR

NA - CARBS
12 13

4.6

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100 BSHC corr to = 1 (g/kWh), CR = 13
PL

BSHC corr to = 1 (g/kWh), MAP = 100kPa 10000 9000 Speed (rev/min)

P L

90 MAP (kPa)

3.5

2.1

8000 7000 6000 5000

2.3 2.5 2.8 3.0 3.2

80

4.0 4.5

70
5.5

60
7.5

6.5

NA - PFI

4000 9 10 11 CR

NA - PFI
12 13

3.5

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60 BSHC corr to = 1 (g/kWh), CR = 11
2.6
P L COMPRESSOR FLOW LIMIT

BSHC corr to = 1 (g/kWh), CR = 10 260 240 220 200 MAP (kPa) 180 160 140 120 100 80 60
5.0 2.2 2.4 2.6 3.0 4.0
TURBO FLOW LIMIT

2.0
RESTRICTED FLOW LIMIT

2.8 3.0 3.5 4.0 4.5 5.0 5.5

2.6

2.0 PL

SC - PFI

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.14: Corrected BSHC trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

217

In the boosted modes where spark retard was used as a method of knock control, the increased exhaust temperature increases the post-combustion oxidized fraction during the expansion and exhaust strokes. However, the higher exhaust temperature is offset by reduced efficiency and heat loss to the walls which results in only mild decreases in BSHC levels in these regions. When comparing the BSHC trends to larger bore engines, experimental results show higher HC levels when compared to modern engines. HC emissions of small bore engines are generally higher due to the increased surface to volume ratio and higher CR [95, 220]. When compared to older larger bore engines [25, 92], the increased surface to volume ratio and higher CR of the test engine are offset by the significantly reduced crevice volume, mainly attributed to the piston top land thickness. Although HC concentrations vary between engines, trend comparisons show that the HC effects due to experimental parameter variation closely follow the literature [98]. Furthermore, comparisons to previously published HC data for modern small engines show that the test engine has lower engine-out BSHC levels [199]. The differences between other small engines with similar surface to volume ratios are explained by a number of factors, but mainly attributed to the engine design which further enhanced performance. The engine design featured minimal piston ring crevice volume together with a gasketless interface [16] (Chapter 3.6 and 5.2). The gasketless interface had two positive HC reducing effects. The first involved a more uniform temperature distribution across the cylinder block and head, which reduced localized cool spots and hence reduced HC levels. Furthermore, the elimination of any crevice volume around the bore periphery at the interface reduces crevice sourced wall quenched HC emissions. These effects together with other factors resulted in the test engine recording actual best engine-out HC emission concentrations as low as 57 ppm C6 in the TC mode. Concentrations at these levels negate the need to use any post engine HC emission control strategy at this condition to satisfy Euro 3 HC emission standards [17] (Appendix C.2). Furthermore, these low HC values are comparable to levels of a hydrogen fuelled CFR engine in the same research laboratory, attributed to the oil consumption of the hydrogen test engine [92].

218

7.6.2 BSNOx
The literature documents temperature, oxygen concentration and residence time as factors contributing to the emission formation of NOX in engines. Thermal NOX emissions generally form above 2,000 K, with Zeldovich the first to discover the exponential increase as a function of temperature [32, 92]. Figure 7.15 displays raw engine-out brake specific oxides of nitrogen (BSNOX) results across NA and boosted modes. It is noted that the varying fuel mixtures, as previously described, heavily affect NOX concentrations and care must be taken when analyzing the data.
100
5 4

BSNOX (g/kWh), CR = 13
PL

10000 9000 Speed (rev/min)

BSNOX (g/kWh), MAP = 100kPa

P L

90 MAP (kPa)

8 7 6 5

6 7

8000 7000 6000 5000

80
10

70

4 3

60

12

NA - CARBS

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 9 10 11 CR

NA - CARBS 2
12 13

150 140 130 120 MAP (kPa) 110 100 90 80 70 60

BSNOx (g/kWh), CR = 11
260
3
P L 4 COMPRESSOR FLOW LIMIT

BSNOx (g/kWh), CR = 10 240 220 200 MAP (kPa) 180 160 140 120 100 80 60
12.0 10.0 2.0 TURBO FLOW 2.5 LIMIT 3.0 3.5 4.0 5.0 6.0 7.0 8.5
PL RESTRICTED FLOW LIMIT

5 6 8 10 12

SC - PFI

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.15: BSNOX trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

219

Figures 7.12 and 7.16 highlight the effects of varying fuel mixtures on NOX emissions across NA and boosted modes. NOX concentrations highlight that peak formation occurs slightly lean of stoichiometric. This is due to the excess availability of oxygen in slightly lean mixtures, although peak flame temperature occurs in slightly rich mixtures. Results show that for lean and rich mixtures on either side of the peak, NOX emissions rapidly decrease. Hence, trends highlight that NOX formation is highly dependent on combustion temperatures, which are heavily influenced by fuel mixtures and power levels. Thus, NOX concentrations are shown to increase for rising engine speed and/or MAP as documented in the literature for larger bore engines [98]. Figure 7.15 displays BSNOX trends specific to this particular tuning strategy, where fuel mixtures are slightly lean of stoichiometric at light load ( = 1.1) and become richer (up to ~0.77 ) as the load increases. Raw results for MAP variation show highest BSNOX levels at low loads, associated with poor BSFC (Figure 7.8) and high NOX concentrations due to the slightly lean fuel mixtures. Hence BSNOX levels reduce as the MAP increases due to the lack of oxygen associated with rich mixtures, decreasing further with very rich mixtures due to the reduced combustion temperatures. The upper right diagram in Figure 7.15 shows the BSNOX emissions with varying CR. It is possible to analyze the CR effects in the carbureted mode as mixtures were self-controlling and dependent on airflow, thus little mixture variation existed at constant speeds (Figure 7.7). BSNOX results show negligible changes with CR variation at low power levels (associated with low engine speeds), as confirmed in the raw emission concentrations of Figure 7.16 - Upper. However, experimental results show BSNOX and raw engine-out NOX concentrations increasing as the CR increases at high power levels, near peak performance, due to the higher combustion temperatures. The CR NOX trends follow those in larger bore engines. Extrapolating to stoichiometric conditions at peak power across all modes, NOX concentrations are shown to be highest for the NA modes (3,200 ppm) at a CR of 13. However, concentrations are lowest in the TC mode (3,000 ppm) despite higher intake temperatures and a 1.5 fold increase in peak power. These results are related to the retarded spark timing (Figure 7.6) and associated combustion effects discussed in Chapter 9.7.1.

220

3000

RICH
NO X concentrations (ppm) 2500 CR = 10 2000 1500 1000 500 0 0.75 CR = 11 CR = 13

Increasing Power
0.80 0.85 Lambda ( ) 0.90

NA - CARBS
0.95 1.00

3000

RICH
NO X concentrations (ppm) 2500 2000 1500 1000 500 0 0.7

0-15 kW

LEAN

15-30 0-15 kW kW 30-45 kW 15-30 kW

30-45 kW

Increasing Power
0.8 0.9 Lambda ( ) 1.0

NA - CARBS
1.1 1.2

3000

RICH
NO X concentrations (ppm) 2500 2000 1500 1000 500 0 0.7 0.8 0.9 Lambda ( ) 1.0 1.1 0-15 kW 0-15 kW 15-30 kW 15-30 kW 30-45 kW 30-45 kW

LEAN

Increasing Power

SC - PFI
1.2

Figure 7.16: Engine-out NOX emission concentrations for varying and power outputs, achieved with engine speed, MAP and CR variation. (Upper): NA - Carburetion: MAP = 100 kPa. (Middle): NA - Carburetion: CR = 13. (Lower): SC - PFI: CR = 11.

221

7.6.3 BSCO2
Figure 7.17 displays raw engine-out brake specific carbon dioxide (BSCO2) emissions. Trends are difficult to visualize due to the varying and resulting Hence, BSCO2 contours are proportional variations in CO and CO2 (Figure 7.12).

displayed at corrected fuel mixtures thus allowing fair comparisons at constant values. Figures 7.18 and 7.19 display corrected contours at stoichiometric ( = 1) and rich ( = 0.9) conditions. Correction from the test AFR is made via calculation using a carbon balance, initiating from the mole compositions of the dry exhaust products (Appendix M.6.2). BSCO2 results directly correspond to previous BSFC patterns across all modes. The BSCO2 effects resulting from MAP, engine speed and CR variations directly correlate to the efficiency trends previously outlined in Figure 7.8. Hence, the merging contour lines show efficiency to be directly proportional to BSCO2, with improvements in fuel efficiency corresponding to reduced CO2 emissions. In both NA modes at = 1 (Figure 7.18), a CO2 reduction in the order of 13% is achieved by increasing the CR from 9 to 13, which also corresponds to previous TH improvements. When compared to larger bore automobile engines [25, 120], parameter variations give identical BSCO2 emission trends, with the test engine having reduced BSCO2 levels due to improved efficiencies previously highlighted in Section 7.5.2. Figure 7.19 compares BSCO2 emissions at the rich case, highlighting identical contours but varying amplitudes when compared to other CO2 data. BSCO2 levels are reduced in the rich case even though more fuel is consumed. This is caused by increasing levels of CO (Figure 7.21) which offsets the BSCO2 reductions. Hence, experimental results show the CO2 production is dependent on the amount of fuel consumed together with the CO contribution. At stoichiometric conditions, the majority of fuel consumed is converted into CO2 as the CO contribution is small (~0.75%) with increasing contributions at richer fuel mixtures (Figure 7.12 Upper). At these conditions, a minimum BSCO2 value of 730 g/kWh was recorded in the NA - PFI mode, corresponding to the highest TH of 36.9%. This is compared to a 15% increase in CO2 with the implementation of turbocharging, caused by high levels of fuel enrichment and ignition retard which prevented higher efficiencies but allowed knock to be suppressed to enable increased BMEP.

222

100

BSCO2 (g/kWh), CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000 6000

BSCO2 (g/kWh), MAP = 100kPa

950 900 850 800 850 700 700 750 800 675 750
P L

90 MAP (kPa)

610 650 690 700

80

750 800 900

70

1000 1200

60

5000

700

1400

NA - CARBS

4000 9 10 11 CR

NA - CARBS
12 13

675

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 BSCO2 (g/kWh), CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000 6000

BSCO2 (g/kWh), MAP = 100kPa

900 850 790 790 735 700 675 790 735 700
P L

90 MAP (kPa)

650 700

850

80

800 900

70

1000 1100

675 735 700

60

1200 1300

5000

NA - PFI

4000

850

NA - PFI
10 11 CR 12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60
700 800 900 1000 1200 1400 750 700
P L COMPRESSOR FLOW LIMIT

BSCO2 (g/kWh), CR = 11
800

BSCO2 (g/kWh), CR = 10 260 240 220 200 MAP (kPa) 180 160 140 120 100 80
750 800 900 1000 1200
TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

750

650

650 700

PL

800

SC - PFI

60

1400

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.17: BSCO2 trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

223

100

BSCO2 corr to = 1 (g/kWh), CR = 13

PL

BSCO2 corr to = 1 (g/kWh), MAP=100kPa 10000

1050

90 MAP (kPa)

800

9000 Speed (rev/min) 8000 7000 6000 5000

950

1000

950

900

850

P L

80

750 800 850 900 1000 1200

950 900 800 850 800 770 770 900 800 850 - CARBS NA

70

60

1400

NA - CARBS

4000

1000

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 BSCO2 corr to = 1 (g/kWh), CR = 13
PL

10

11 CR

12

13

BSCO2 corr to = 1 (g/kWh), MAP=100kPa 10000 9000 Speed (rev/min) 8000 1000 7000 6000
950 1000 950 900
P L

90 MAP (kPa)

730 740

850

80

800 900 1000

950 900

850 800

800 770

70

1100 1200 1400 1300

60

5000

900

850

770 800

NA - PFI

1000

4000 9 10 11 CR

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60
1000 1200 1400 850 825 800 825 850 900
P L

BSCO2 corr to = 1 (g/kWh), CR = 11 260

1000 COMPRESSOR FLOW LIMIT

BSCO2 corr to = 1 (g/kWh), CR = 10 240 220 200 MAP (kPa) 180 160 140 120 100 80
900 950 1000 1100 1400
TURBO FLOW LIMIT 1000

1200 RESTRICTED FLOW LIMIT 1100

PL

SC - PFI

60

1200

TC PFI 1400

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.18: Corrected BSCO2 trends ( = 1) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

224

100

BSCO2 corr to = 0.9 (g/kWh), CR = 13

700
PL

700

BSCO2 corr to = 0.9 (g/kWh), MAP=100kPa 10000 900

850

90

9000
850

800

750 700

Speed (rev/min)

MAP (kPa)

650

8000 7000 6000

800

800

80

750 700 675 675 750

700 750 850 950 1050

70

60

5000

1400

1200

NA - CARBS

4000

850

NA - CARBS
10 11 CR 12 13

700

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 100 BSCO2 corr to = 0.9 (g/kWh), CR = 13
PL

BSCO2 corr to = 0.9 (g/kWh), MAP=100kPa 10000

900 850

90 MAP (kPa)

640 630

9000 Speed (rev/min) 8000 850 800 7000 6000 5000

800 850 750

800 750

750 715 700

700 715

80

700

800

685 675

70

900 1000

675 685 700 715

60

1100 1200

NA 1100

PFI

4000

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) 150 140 130 120 MAP (kPa) 110 100 90 80 70 60
1200 900 1000 1100 700 750 800 750
P L

10

11 CR

BSCO2 corr to = 0.9 (g/kWh), CR = 11 260

900 COMPRESSOR FLOW LIMIT

BSCO2 corr to = 0.9 (g/kWh), CR = 10 240 220 200 MAP (kPa) 180 160 140 120 100 80 60
1200 800 900 950 1000 1050 1100 850
TURBO FLOW LIMIT 900 950 1000 RESTRICTED 1100 FLOW LIMIT

1050
PL

SC 1200 - PFI

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.19: Corrected BSCO2 trends ( = 0.9) versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

225

7.6.4 BSCO
Figure 7.20 displays CO mole concentrations for and brake power (engine speed and MAP variation) variation. For both boosted modes with alternative CRs, it is shown that CO emission formation is strongly influenced by the AFR, but largely independent of brake power. CO concentrations leaner of stoichiometric are shown to be negligible, with rising levels for increasingly rich mixtures. For rich mixtures, CO levels are high because complete oxidation of the fuel carbon to CO2 is not possible due to insufficient oxygen. Hence, increasing levels of fuel richness reduces combustion efficiency which produces increasing amounts of CO. Figure 7.21 displays raw engine brake specific carbon monoxide (BSCO), with large variations over the domain caused by and BSFC variations, with variations being the dominant contributor due to high fold changes resulting from the mixture irregularities. BSCO variations range from 20-550 g/kWh, the highest recorded in the TC mode due to the excessive enrichment and resulting poor BSFC. Across all modes at constant fuel mixture contours (Figure 7.7), results show BSCO variations across the speed range due to the changing efficiency. When BSCO levels are corrected to constant values, contour trends mimic BSCO2 due to the common BSFC and the relationship between CO and CO2 emission formation at constant .
CO (%), 150kPa MAP (%), Limited, = 11 CO Concentration CR CR = 11
UNEXPLORED REGION

40

60 50 Brake Power (kW) 40 30 20 10

CO (%), 150kPa MAP (%), Limited, = 10 CO Concentration CR CR = 10

15
UNEXPLORED REGION

Brake Power (kW)

30

20

15 12 11 9 9 7 5

13 11 9

10
UNEXPLORED REGION

0 0.7 0.8 0.9 1.0 Lambda

SC - PFI
1.1 1.2

UNEXPLORED REGION

0 0.7 0.8 0.9 1.0 Lambda

TC - PFI
1.1 1.2

Figure 7.20: Mole fraction CO trends for varied and brake power (engine speed and MAP variations). Shaded areas indicate unexplored regions. (Left): SC - PFI. (Right): TC - PFI.

226

100

BSCO (g/kWh), CR = 13
300 250
PL

10000
120

BSCO (g/kWh), MAP = 100kPa

120 110 120 160 200 250 300 270 120 110 110 100
P L

90 MAP (kPa)

9000
170 120

Speed (rev/min)

8000 7000 6000 5000 4000 9 10

80
80 50 80

70

40

60

NA - CARBS

350

NA - CARBS
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

11 CR

100

BSCO (g/kWh), CR = 13
PL

10000 9000 Speed (rev/min) 8000 7000 6000 5000

BSCO (g/kWh), MAP = 100kPa

240 200 180 160 140
P L

200

150

MAP (kPa)

125

80
100 75 50

60
20

NA - PFI

4000 9 10 11 CR

NA - PFI
12 13

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

150 140 130 120 MAP (kPa) 110 100 90 80 70 60
20 50 100
P L

BSCO (g/kWh), CR = 11 260

350 COMPRESSOR FLOW LIMIT 300 250 200 150

BSCO (g/kWh), CR = 10 240 220 200 MAP (kPa) 180 160 140 120 100 80
300 200 100 25 400
TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

500

PL

SC - PFI

60

TC - PFI

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 7.21: BSCO trends versus engine speed, MAP and CR. Shaded areas indicate airflow limited regions. PL is the performance limit line. (Upper): NA - Carburetion. (Middle): NA - PFI. (Lower Left): SC - PFI. (Lower Right): TC - PFI.

227

7.7 Summary
This chapter presented the performance, efficiency and emissions experimental results from the test engine, which has been operated in a variety of NA and boosted modes. NA modes included carburetion and PFI fuel delivery, while boosted running included SC and TC intake systems. Results are displayed in the engine speed, MAP, CR and domains, with engine speeds exceeding 10,000 rev/min, MAP reaching 270 kPa and CRs ranging from 9.6 to 13. When altering the engine speed, MAP and CR parameters, experimental results showed similar order effects on performance, efficiency and emissions when compared to other small engines and typical larger bore engines found in passenger vehicles. Experimental results show minimum BSFC or maximum TH values in the order of 220 g/kWh or 37% could be achieved. Furthermore, a maximum NA BMEP of 13 bar was achieved, while 25 bar was reached in the TC mode. This TC value is believed to be the highest recorded for small engines while operating on pump gasoline, with the value especially significant as it was achieved using no complex knock preventive strategies. The increased performance of this particular small engine is attributed to the physical size reduction, particularly the reduced bore size and fast burn combustion chamber. The relationship between engine speed, MAP and CR parameters and their effect on the operating limits has been described for this small engine across all four test modes. Experiments highlighted that end-gas knock is the dominant factor limiting the performance of this small engine. Testing demonstrated that spark retard and/or fuel enrichment can be used as methods of knock control for up to 1-2 CR points depending on the knock severity. This allows the operating limit to be extended, enabling increased performance at the cost of reducing efficiencies. Knock severity was also found to decrease for reducing end-gas volumes. Furthermore, trends from the knock limits also found that for a given CR and MAP condition, knock is less susceptible at higher engine speeds. The reduced knock likelihood is a consequence of the increased flame speeds within the combustion chamber (Chapter 9), which consume the unburnt mass in the end-gas region more quickly. Increasing flame speeds decreases the knock likelihood due to the reduced end-gas residence time. This has major benefits in extending the operating limits in downsized applications.

228

CHAPTER 8
P O W E R E D B Y W A T T A R D

Performance, Efficiency and Emission Comparisons

8.1 Introduction
This chapter compares the test engines performance, efficiency and emission experimental results between test modes (Chapter 7) and with other OEM based production engines. The effects of varying fuel delivery systems for the test engine are compared between carburetion and PFI, including the CR and intake flow restriction effects. Further results comparisons are made between NA, SC and TC modes at the test CR closest to the HUCR to quantify the effects of the varying intake systems. Comparisons are also made between the test engine data and several modern production engines with similar and larger capacities. The purpose of the This is comparisons is to highlight the potential of this research in giving direction to the future development of small engines (~0.5 L) for passenger vehicles. in popularity. Downsizing potential is exemplified, with a case study performed to determine the feasibility of replacing a modern 1.25 L NA engine found in a compact sized regular passenger vehicle with the TC test engine. especially important as engine downsizing to improve energy consumption grows

229

8.2 Carburetion and PFI Fuel Delivery

The two systems of fuel delivery involving carburetion and PFI are compared for the NA modes at steady state conditions, with specifications previously outlined (Chapter 4.5). It is noted that both fuel delivery systems require differing intake geometry (Table 4.4), resulting in minor changes to engine air consumption rates as later discussed (Section 8.3). Figure 8.1 displays fuel delivery comparisons at the PL (WOT) to evaluate the performance, efficiency and emission effects. The effects of the fuel delivery system with varying CR are also evaluated. Only three data sets are compared across the test range. However, results provide the reader with clear findings from the experiments regarding fuel delivery effects. The performance comparisons (Figure 8.1 - Upper) for equal CRs suggest similar BMEP potential between PFI and carburetion when accounting for the airflow discrepancies, which could be eliminated with intake system optimization [14, 16]. The reduced air consumption of the PFI system is associated with fitting the mandatory FSAE intake restriction. Although engine air consumption rates were not high enough to cause choked flow, the intermittent pulsing attributed to the uneven firing interval limited instantaneous peak airflow through the nozzle, resulting in reduced VOL and hence BMEP (Section 8.3). The fuel consumption comparisons (Figure 8.1 - Middle) indicate clear gains when implementing PFI over carburetion at WOT. The improved fuel break-up and atomization of the PFI system aids in vaporization, resulting in improved mixing together with reduced wall wetting and pooling. These effects improve efficiency and have more influence at low engine speed and load conditions (Figure 7.8). It is also reiterated that these results are attained from steady state testing, with transient testing likely to demonstrate significantly different results due to fuel control and pooling issues associated with the high fuel entry position of carburetion. Hence, continuous stoichiometric mixture delivery for efficient TWC operation during engine transience is difficult to achieve with this system as previously recorded [95, 128].

230

1300 1200 BMEP (kPa) 1100 1000 NA - CARBS (CR = 13) 900 800 4000 NA - CARBS (CR = 10) NA - PFI (CR = 10) 5000 6000 7000 Engine Speed (rev/min) 8000 9000 10000

340 320 BSFC (g/kWh) 300 280 260 240 4000

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

600 HC Mole Concentrations (ppm C6)

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

500

400

300

200 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

Figure 8.1: Performance, efficiency and emission effects for alternative CRs and fuel delivery (carburetion and air restricted PFI). NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): BMEP. (Middle): BSFC. (Lower): Engine-out HC mole concentrations.

231

When comparing emissions between carburetion and PFI (Figure 8.1 - Lower), HC emissions standout as being the major difference between the two modes of fuel delivery. Reductions in the order of 20% are shown at WOT when implementing PFI during steady state testing, with HC differences between both modes expected to increase during transient operation. The differing TH between carburetion and PFI is shown to have insignificant effects on NA combustion (Chapter 9.4), consistent with the similar performance potential findings discussed. However, the increased levels of engine-out HC emissions have contributed to the poorer TH of the carbureted system. These performance, efficiency and emission results confirm previous findings from larger bore engines, where PFI was adopted over carburetion due to improved fuel and emission control [95, 211].

8.3 Intake Flow Restrictor Effects

Restrictor effects have been previously documented for NA SI engines [45, 54, 63, 187]. However, the effects documented are generally for four cylinder configurations which feature even fire spacing, with no information found for the flow restriction effects associated with an odd firing inline twin. Figure 8.2 displays the experimental results at WOT for odd and even fire engine configurations, with further details in Table 8.1. Both engines have been fitted with and without a 20 mm diameter intake flow restriction and an accompanying shift from carburetion to PFI. Both configurations also feature similar plenum The even fire flow restriction effects volumes (~4 L) for PFI intake systems.

shown were found to be representative of four cylinder configurations (Figure 8.2 - Lower), irrespective of the manufacturer [75, 138, 147, 221].
Table 8.1: Specifications for odd and even fire engines compared in Figure 8.2. Odd fire
Type Configuration Engine Capacity Firing Interval CR Mode Test Engine Inline Twin 0.43 L 0, 180 CA 13 Unrestricted (NA - CARBS) Restricted (NA - PFI)

Even fire
Suzuki GSX-R600 Inline Four 0.6 L 0, 180, 360, 540 CA 12 Unrestricted (NA - CARBS) Restricted (NA - PFI)

232

60 50

Brake Power (kW)

40 30 20 10 0 3000 Test Engine Unrestricted (NA - CARBS) Test Engine Restricted (NA - PFI) Suzuki GSX-R600 Unrestricted (NA - CARBS) Suzuki GSX-R600 Restricted (NA - PFI) 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000

Engine Speed (rev/min)

1200 1000

BMEP (kPa)

800 600 400 200 0 3000 Test Engine Unrestricted (NA - CARBS) Test Engine Restricted (NA - PFI) Suzuki GSX-R600 Unrestricted (NA - CARBS) Suzuki GSX-R600 Restricted (NA - PFI) 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000

Engine Speed (rev/min)

Perfromance Variation (%) [(Rest - Unrest) / Rest] [(Restricted Unrestricted) / Restricted]

20

10

Test Engine (2 cyl) Suzuki GSX-R600 (4 cyl) Honda CBR-600RR (4 cyl) Yamaha YZF-R6 (4 cyl)

0 3000 -10

4000

5000

6000

7000

8000

9000

10000

11000

12000

13000

-20

Engine Speed (rev/min)

Figure 8.2: Performance effects for odd (0.43 L, inline twin, CR = 13) and even (0.6 L, inline four, CR = ~12) fire configurations in restricted (NA - PFI) and unrestricted (NA CARBS) modes. WOT, MBT spark timing.

233

The performance comparisons of Figure 8.2 show significant performance changes between odd and even fire configurations when implementing the intake flow restriction. Both engine configurations show matching trends, with reducing However, performance at elevated engine speeds due to the flow restriction.

experimental results show large differences between the engine speeds at which the restrictor begins to reduce performance. For the four cylinder configuration, the restrictor is shown to reduce performance at speeds above 9,000 rev/min compared to the test engine at 7,000 rev/min. The test engines results also follow the data trend of Figure 8.1, which displayed the restrictor effects at a reduced CR of 10. The reader is reminded that the two cylinder engine has two thirds the capacity of the four cylinder, so it would be anticipated that the engine would run to a higher speed than 9,000 rev/min before a turning point in the difference curve occurred. This suggests the reduced air consumption of the second cylinder was restricted, as its cylinder capacity is larger when compared to the four cylinder. Consequently, test engine airflow comparisons are made between carburetion and PFI to see the effects of implementing the flow restriction. Before experimentation, it was anticipated that the flow restriction would have little or no effect on engine air consumption, as average rates were not high enough to cause choked flow (Appendix D). This assumption was based on previous simulation results of Chapter 3.3.2 and 6.5, which documented the intake flow restriction effects for TC operation, where engine air consumption rates are several magnitudes higher. Figure 8.3 displays the test engines air consumption for restricted (NA - PFI) and unrestricted (NA - CARBS) modes. The upper diagram of Figure 8.3 displays the individual cylinder VOL (as measured in Chapter 4.3.1) versus engine speed for the NA test engine, with varying fuel delivery and induction systems. It is evident that the carbureted mode has very small VOL variation between cylinders, as the intake system featured no flow restriction or plenum chamber, as each cylinder featured individual carburetors and open air trumpets (i.e. equivalent to an infinite size effective plenum volume). This configuration allowed each cylinder to operate independently from each other, with no air consumption limitations or cylinder interactions. Consequently, it was assumed little or no performance differences existed between cylinders, which was later confirmed experimentally by a Morse test [211].

234

Individual Cylinder VOL (%)

110

Cyl Cyl Cyl Cyl

1 (NA 2 (NA 1 (NA 2 (NA

- CARBS) - CARBS) - PFI) - PFI)

100

90

80

70 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

0.08
RESTRICTED STEADY FLOW LIM IT

Mass Airflow AIR (kg/s)

0.06

Unrestricted (NA - CARBS) Restricted (NA - PFI)

0.04

0.02

0 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

15 10 Airflow Variation (%) [(a-b)/a] 5 0 4000 -5 -10 -15

NA - CARBS [Cylinder (1-2)/1] Cylinder (NA - CARBS), [(1-2)/1] NA -- PFI), PFI [Cylinder Cylinder (NA [(1-2)/1] (1-2)/1] [Mode (PFI-CARBS)/PFI] Mode (PFI vs CARBS), [(PFI-CARBS)/PFI]

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 8.3: Air consumption effects for restricted (NA - PFI) and unrestricted (NA - CARBS) modes. WOT, CR = 13, MBT spark timing. (Upper): Individual cylinder VOL. (Middle): Engine air consumption and limitations. (Lower): Cylinder and mode airflow variations.

235

However, PFI results from Figure 8.3 show an overall reduction in air consumption when compared to carburetion over all speeds. For the PFI mode, the first cylinders airflow decreases over all speeds when compared to carburetion, amplifying in effect as the engine speed increases. Also evident is the second cylinders VOL reduction, with recorded differences reaching 10%. These PFI air consumption differences are primarily caused by implementing the mandatory intake flow restriction and concomitant plenum chamber, with the effects shown in Figure 8.4. These airflow results follow the trends of the performance differences previously highlighted. Hence, the majority of performance difference associated with the change in modes is attributed to the intake manifold geometry, specifically the flow restriction effects rather than the fuel delivery system. PFI air consumption rates in the NA mode were not high enough to cause restrictor choking, with peak induced mass flow approximately half of the steady flow limit (Figure 8.3 - Middle). However, the intermittent pulsing caused by the unequal induction spacing between cylinders limited instantaneous peak airflow through the nozzle, resulting in reduced air consumption (Figure 8.4). This effect increases with engine speed as engine air consumption rates increase, resulting in further reductions in performance as seen in Figures 8.2 and 8.3.

0.14 Mass Airflow AIR at x (kg/s) 0.12 0.10 0.08 0.06 0.04 0.02 0.00 0 120 240 360 Crank Angle (deg) `

Unrestricted - Odd Fire Twin Restricted - Odd Fire Twin Unrestricted - Even Fire Four Restricted - Even Fire Four

CHOKED FLOW

480

600

720

Figure 8.4: Predicted air consumption effects versus time at a given upstream condition (x) for odd (0.43 L, inline twin) and even (0.6 L, inline four) fire configurations in restricted and unrestricted modes.

236

To smooth out the large flow fluctuations through the restrictor, larger plenum volumes are needed, which may compromise transient response [14, 16]. This effect is significantly reduced in multi-cylinder configurations with even firing intervals (> three cylinders) due to the smaller induced cylinder volume per cycle and more frequent induction pulses through the restrictor [54, 245]. Hence, reducing the cylinder volume and evenly spacing the induction pulses maintains sonic flow through the restrictor for longer periods of time (for equal engine capacities) when operating at the choked condition, as demonstrated in Figure 8.4. For the odd fire two cylinder configuration, the close induction spacing between cylinders also causes higher rates of cylinder airflow variation (Figure 8.3 - Lower), with reduced second cylinder filling due to the plenum chamber being partially evacuated by the first cylinders previous induction (Figure 8.4). With the induction between the first and second cylinders spaced by 180 CA, there was insufficient time for plenum recovery resulting in an uneven air distribution between cylinders. This resulted in cylinder variations in VOL and trapped residual gas fraction, as confirmed by each cylinders exhaust gas analysis. Thus, individual cylinder fueling requirements were needed to equalize AFR, with up to a 10% trim needed for the second cylinder at WOT in the NA - PFI mode. It is possible that the charge robbing effect could be minimized with improvements in intake manifold design. optimized for TC operation. Plenum chamber scavenging variations due to cylinder induction are eliminated in even fire configurations, significantly reducing airflow variations between cylinders when compared to odd fire configurations. However, depending on the intake geometry, cylinder airflow variations can still be prevalent in even fire configurations due to cylinder flow biasing [54, 138]. geometry into the plenum chamber. Experimental results and analysis have shown that even fire engine configurations have the potential to induce higher air consumption rates in NA modes for similar engine and intake manifold designs. However, the NA odd firing airflow This effect generally increases as the numbers of cylinders increase, which can compromise the duct However, the plenum chamber geometry had been

237

reductions attributed to the intermittent pulsing through the intake restriction are shown to be significantly reduced when turbocharging is employed (Chapter 6.5). In the TC mode, effects on total engine air consumption due to the flow restriction are significantly reduced as the engine is force fed air via the compressor. This minimizes the pulsing effect through the flow restriction caused by the piston motion during induction, as the compressor induces mass flow through the orifice over all CA degrees. Hence, installing a turbocharger is shown to be very However, the uneven beneficial when applied to an odd firing engine (Chapter 6.5), greatly reducing the intermittent pulsing effects through the flow restriction. induction spacing in either NA or boosted odd fire engines results in higher intermittent levels of airflow and hence power output variation between cylinders, which can only be reduced with increases in intake plenum volume.

8.4 NA, SC and TC Mode Comparisons

Performance, efficiency and emission comparisons are presented for all test modes across varying engine speeds. Experimental results are compiled from tests completed at the CR closest to the HUCR, as previously defined in Table 7.2. For each engine speed across all modes, comparisons are made at the maximum power which corresponds to the PL. The PL is defined at each modes WOT for a given CR as MAP values vary in the boosted modes due to the airflow delivery limitations previously described (Chapter 7.2).
300 250 MAP (kPa) 200 150 100 50 4000 NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10)

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

Figure 8.5: Maximum MAP values achieved at the PL (WOT) across all test modes at the experimental test CR closest to the HUCR.

238

Figure 8.5 defines the MAP operating characteristics corresponding to each modes PL and hence maximum power for varying engine speeds. As can be seen, MAP decreases in the TC mode above 6,000 rev/min, as the boost pressure ratio is reduced from ~3 to maintain sonic flow conditions through the intake restrictor. The removal of the mandatory flow restriction, would most likely enable the delivery of sustained MAP values above this speed if turbocharger choking limits could be avoided. The SC mode also shows a similar trait with decreasing MAP. However, this is due to the externally driven Rootes blower being incapable of delivering the required airflow to maintain the desired MAP of 150 kPa at the higher engine speeds (Chapter 7.2). Performance characteristics are displayed in Figure 8.6, with BMEP, brake power and spark timing compared for the various modes. Performance levels correspond to each modes air consumption rates. Consequently, both boosted modes have improved performance when compared to NA configurations, with turbocharging delivering greater output than supercharging due to the higher achievable MAP. However, despite the cam timing and exhaust system being optimized for TC operation (Chapter 6.5), the high BMEP of the NA modes (~13 bar) indicates excellent port design [246] as confirmed by the flow bench results in Appendix H. Furthermore, the engine and combustion chamber design (Chapter 3) has enabled high MECH and TH, resulting in higher BMEP values when compared to other engines found in todays passenger vehicle marketplace (Section 8.7). This was identified to be an important factor in obtaining the best possible torque before the turbocharger produces wastegate limited boost [16, 191, 248]. For NA operation, the BMEP and resultant brake power for the PFI system is reduced when compared to carburetion over all speeds due to the airflow differences previously detailed (Section 8.3). The upper diagram of Figure 8.6 displays a comparison of BMEP between modes, with a peak value of 25 bar recorded in the TC mode. This is believed to be the highest recorded value for small engines operating on pump gasoline [14-17, 35, 198] (Chapter 7.2). The test engine values have greater significance when compared to other results [35, 198] in that this was achieved using no complex knock preventive strategies, relying on traditional methods of tuning strategies and CR reduction.

239

2400

NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10)

BMEP (kPa)

2000

1600

1200

800 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

60 50 Brake Power (kW) 40 30 20 10 0 4000 NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10) 5000 6000 7000 Engine Speed (rev/min) 8000 9000 10000

40 Spark Timing (deg. BTDC)

30

20 NA - CARBS (CR = 13) NA - PFI (CR = 13)

= KL-ST

10

SC - PFI (CR = 11) TC - PFI (CR = 10) 6000 7000 Engine Speed (rev/min) 8000 9000 10000

0 4000

5000

Figure 8.6: Performance comparisons at the PL (WOT) for all modes at the experimental test CR closest to the HUCR. (Upper): BMEP. (Middle): Brake Power. (Lower): Spark Timing.

240

The increased performance from the test engine is due to the physical size reduction, particularly the reduced bore diameter and fast burn combustion chamber. This results in engine speed increases together with end-gas volume reductions around the periphery of the chamber, allowing CR and/or MAP values to be increased before exceeding the DL (Chapter 7.2). The middle diagram of Figure 8.6 demonstrates the test engines ability to deliver almost constant power over the 6,000-10,000 rev/min speed range in the TC mode. The concept of achieving near constant power over a wide speed range is documented to improve vehicle drivability with a reduction in the required gearshifts [13]. At 6,000 rev/min with a boost pressure just under two bar, the engines peak torque is 2.6 times that of the NA - PFI version. At higher engine speeds, the torque falls as the power is limited by sonic flow through the restrictor. The spark timing shown in the middle diagram of Figure 8.6 highlights areas where spark retard was used to control high knock intensities. The non-knock limited boosted areas display significant reductions in spark timing when compared to the NA modes due to charge density increases. burning rates (Chapter 9). A peak VOL of 250% was achieved in the TC mode as shown in the upper diagram of Figure 8.7. The airflow reductions observed in the NA - PFI mode due to the flow restriction are significantly reduced in the boosted modes, as previously described (Section 8.3). Hence, in the boosted modes, the compressor has the capability of inducing the maximum mass flow through the orifice over all CA degrees. In the unrestricted carbureted mode, resonant wave tuning has enabled This increased charge density significantly enhances the combustion process due to the increased

VOL values to exceed 100% between 7,000 and 9,500 rev/min, resulting in a peak
NA VOL of 108% at 8,000 rev/min. The middle and lower diagrams of Figure 8.7 demonstrates the engines fuel efficiency at the PL. Raw results are shown together with results corrected to stoichiometric conditions (Appendix M.6). Although results are corrected to = 1, no compensation can be made for the lost efficiency due to the retarded spark timing in regions where this strategy was used to control high knock intensities.

241

300 250 200 VOL (%) 150 100 50 0 4000

NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10)

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

420

370 BSFC (g/kWh)

NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10)

320

270

220 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

350 BSFC corr to = 1 (g/kWh)

NA - CARBS (CR = 13) NA - PFI (CR = 13) SC - PFI (CR = 11) TC - PFI (CR = 10)

300

250

200 4000

5000

6000

7000 Engine Speed (rev/min)

8000

9000

10000

Figure 8.7: Efficiency comparisons at the PL (WOT) for all modes at the experimental test CR closest to the HUCR. (Upper): VOL. (Middle): BSFC. (Lower): BSFC and TH corrected to = 1.

242

Across all modes, a best BSFC and TH value of 220 g/kWh or 37% was achieved for the NA - PFI mode, corresponding to the highest CR used in testing. As previously described, it is clear that the PFI system has advantages in fuel atomization and vaporization over the carbureted system, resulting in improved mixing and thus improving BSFC and TH. However, the expected improvements in efficiency from turbocharging have not been realized at the PL as previously detailed in Chapter 7.5.2. This is mainly due to the varying MAP, resulting in the CR being significantly reduced to avoid the DL at the highest pressure ratio. Consequently, this has compromised the efficiency at lower boosted operating conditions.

8.5 Comparison to FSAE Engines

The original intent of this project was to achieve success in FSAE competition with an innovative new engine design. As previously described (Chapter 1.1.1), the initial project objectives surrounded improved vehicle performance and fuel efficiency, which could be achieved with an engine design specifically targeted for the competition. In contrast, the majority of the engines used in the competition are adapted from OEM motorcycle engines, which compromise performance, vehicle mass and packaging in order to comply with FSAE regulations. Table 8.2 compares the final TC test engine to typical four cylinder, 0.6 L motorcycle engines adapted for use in competition, clearly demonstrating that the initial project objectives regarding FSAE (Appendix B. 6) have all been achieved. Figure 8.8 displays the performance of a range of FSAE engines tested on the same dynamometer as the test engine. Although engines were fitted with Consequently, all engines had different flow restrictions, flow bench results confirm near identical peak mass flow through the restrictor (Appendix H.5.4). comparable air consumption when operating in the choked condition. Intake flow restricted results show that the test engine achieved a 7% improvement in maximum power over previous FSAE engines, despite being only two thirds the capacity. Due to the limited airflow, this improvement is due to the improved TH associated with the higher charge densities and fast burn combustion chamber. This was achievable due to the engine design, which permitted both an extension

243

of the operating limits and frictional improvements.

Further power gains are

attributed to frictional improvements associated with the reduced engine speeds and smaller capacity. Turbocharging has also enabled a 2.5 fold increase in peak BMEP over typical FSAE engines. Furthermore, turbocharging has also allowed sonic flow to be achieved through the flow restriction, resulting in near constant power over half the speed range as previously described. Operating near constant power has resulted in improved vehicle drivability due to the avoidance of gear shifting for much of the FSAE competition [13].

Table 8.2: Engine specifications and performance data for a range of in-house developed FSAE engines across NA and boosted modes. Typical FSAE Engine
Type Configuration Engine Capacity Maximum Engine Speed Induction Maximum Power Constant Power Range Maximum Torque Power/Capacity Maximum BMEP Maximum VOL Maximum Restricted AIR Engine Mass (bare) Volume Size COG Height Engine Packaging Motorcycle Inline Four 0.6 L ~14,000 rev/min NA (restricted) 53 kW ~25% 52 Nm 88 kW/L 1,100 kPa ~110% ~ 0.07 ~65 kg dry ~60 L ~300 mm Poor

Test Engine (TC Mode)

Prototype Inline Twin 0.43 L 10,500 rev/min TC (restricted) 57 kW ~50% 83 Nm 132 kW/L 2,490 kPa 250% ~ 0.07 47 kg dry ~30 L ~230 mm Designed to suit

Improvement 7% 100% 37% 50% 126% ~127% 0% ~27% ~50% ~22%

244

60 50 Brake Power (kW) 40 30 20 10 0 3000 Test Engine (TC - PFI) Test Engine (NA - PFI) Honda CBR-600RR (NA) Suzuki GSX-R600 (NA) Kawasaki ER5 (NA) 4000 5000 6000 7000 8000 9000 10000 11000 12000 13000 14000

Engine Speed (rev/min)

2500 2000 BMEP (kPa) 1500 1000 500 0 3000

Test Engine (TC - PFI) Test Engine (NA - PFI) Honda CBR-600RR (NA) Suzuki GSX-R600 (NA) Kawasaki ER5 (NA)

4000

5000

6000

7000

8000

9000

10000 11000 12000 13000 14000

Engine Speed (rev/min)

60 50 Brake Power (kW) 40 30 20 10 0 4

Test Engine (TC - PFI) Test Engine (TC - PFI) Honda CBR-600RR (NA) Suzuki GSX-R600 (NA) Kawasaki ER5 (NA)

10

12

14

16

18

20

Mean Piston Speed (m/s)

Figure 8.8: WOT performance comparisons for a variety of engines tested for FSAE application. Tests conducted on the same dynamometer with engines fitted with a 20 mm intake flow restriction.

245

8.6 Comparison to Small Engines

The test engine results are compared to modern small engines (~0.5 L) with similar bore diameters (65 - 70 mm) across both NA and boosted modes. Comparisons allow the test engine to be gauged against alternative but comparably sized powertrain. Additionally, comparisons are intended to highlight the performance potential of small engines, in particular the test engine, as a potential replacement for larger engines. Currently, small engines are not employed in regular passenger vehicles, hence test engine results are compared to engines found in current motorcycle and Kei microcar classes. NA results are compared to a Suzuki GSX-R600 [187, 215] and TC results to a Smart Fortwo [205], with engine specifications and comparisons given in Table 8.3 and Figure 8.9. When examining the operating limits of the small engines for both NA and TC modes, it is evident that the test engines HUCR is higher for equal MAP to the comparison OEM engines. Moreover, these CRs are higher than those found in typical larger bore engines (Section 8.7). The elevated CR of the small engines is due to the reduced bore size, which permits the extension of the operating limits as also found from experiments (Chapter 7).
Table 8.3: Engine specifications and performance data for a range of small engines across NA and boosted modes. NA - Carburetion
Type Configuration Engine Capacity Cylinder Capacity Bore x Stroke Engine Speed CR Maximum Power Maximum Torque Maximum BMEP Maximum VOL Test Engine Inline Twin 0.43 L 217 mL 69 x 58 mm 10,500 rev/min 13 Suzuki GSX-R600 Motorcycle Inline Four 0.6 L 150 mL 66 x 44 mm 13,000 rev/min 12

TC - PFI
Test Engine (Restricted) Inline Twin 0.43 L 217 mL 69 x 58 mm 10,500 rev/min 10 Smart Fortwo Microcar Inline Three 0.7 L 233 mL 66.5 x 67 mm 6,200 rev/min 10 45 kW @ 5,250 rev/min 95 Nm @ 2,000 rev/min 1,710 kPa ~170%

40 kW 61 kW ~ 57 kW @ 6,000@ 10,500 rev/min @ 12,500 rev/min 10,000 rev/min 43 Nm @ 8,000 rev/min 1,260 kPa 108% 49 Nm @ 10,700 rev/min 1,100 kPa ~110% 88 Nm @ 6,000 rev/min 2,490 kPa 250%

246

70 60

Brake Power (kW)

50 40 30 20 10 0 1000

Test Engine (NA - CARBS) Suzuki GSX-R (NA - CARBS) Test Engine (TC - PFI) Smart Fortwo (TC - PFI)

3000

5000

7000 Engine Speed (rev/min)

9000

11000

13000

2500

2000

Test Engine (NA - CARBS) Suzuki GSX-R (NA - CARBS) Test Engine (TC - PFI) Smart Fortwo (TC - PFI)

BMEP (kPa)

1500

1000

500 1000

3000

5000

7000 Engine Speed (rev/min)

9000

11000

13000

2500

2000

Test Engine (NA - CARBS) Suzuki GSX-R (NA - CARBS) Test Engine (TC - PFI) Smart Fortwo (TC - PFI)

BMEP (kPa)

1500

1000

500 0 10 20 30 Brake Power (kW ) 40 50 60

Figure 8.9: WOT performance comparisons for a variety of small engines tested and compared to the test engine across NA and TC modes. Engine specifications outlined in Table 8.3.

247

Although the test engine CRs displayed in Table 8.3 were similar to the comparison OEM small engines, it is evident from Figure 8.9 that the test engines performance is superior over both NA and TC modes. This is especially evident in the TC comparisons, with the test engine recording a 50% higher BMEP than the Smart Fortwo. Although CRs are identical at 10, the test engine can tolerate over twice the amount of boost pressure before exceeding the DL. This is a due to a number of factors related to the engine, combustion chamber and turbocharger system design (Chapter 3, 6 and 9), which have enabled the operating limits to be extended. The higher levels of boost result in improved VOL, which in turn increase performance. Consequently, the performance results for the test engine show the potential to replace larger sized passenger vehicle engines with downsized turbocharged units. This is not possible with the Smart engine found in the Kei class vehicle range, which features similar capacity but reduced performance.

8.7 Comparison to Larger Bore Engines

Engines found in passenger vehicles in todays marketplace are usually more than twice the cylinder capacity of the test engine. Hence, the test engines Several and experimental results are compared to previously found performance, efficiency and emission quantitative results from larger, lower speed engines. important differences involving in-cylinder flow, combustion

frictional/temperature effects may be expected as a result of the reduced bore size and increased engine speed. Hence, larger engines are compared to the test engine (~0.5 L) across NA and boosted modes to document their effects. Table 8.4 summarizes the performance, efficiency and emission effects attributed to the reduced cylinder capacity. Experimental results are compared to published data for modern, larger bore engines found in passenger vehicles [11, 31, 72, 98, 100, 174, 220].

248

Table 8.4: Comparing the best performance, efficiency and emissions of the test engine with typical larger bore engines found in passenger vehicles. Test Engine
Cylinder Capacity Bore Diameter Maximum Speed Mode Fuel Delivery CR Maximum BMEP Maximum TH Maximum VOL Maximum MECH Minimum BSCO2 NA Carburetion / PFI 13 1,260 kPa 37% 108% 91% 730 g/kWh 217 mL 69 mm 10,500 rev/min SC PFI 11 1,600 kPa 35% 150% 780 g/kWh TC PFI 10 2,490 kPa 32% 250% 93% 850 g/kWh

Typical Larger Bore Engine [11, 72, 98, 100, 174, 220]
400 - 650 mL 85 - 95 mm 5,000 - 7,000 rev/min NA PFI 9.5 - 10.5 TC PFI 8-9

1,000 - 1,100 kPa 1,400 - 1,600 kPa 30 - 34% 95 - 105% 90 - 93% 30 - 32% 140 - 160% -

800 - 900 g/kWh 850 - 900 g/kWh

Table 8.4 displays the NA and boosted operating CRs for both the test engine and the typical larger bore engine. Across all modes, it is evident that the test engine can operate with considerably higher CR for a given MAP, with results showing potential for NA engine operation at a CR exceeding 13 [11]. The high CR achieved for this particular small engine is attributed to the physical size reduction, particularly the reduced bore diameter and fast burn combustion chamber. This results in engine speed increases together with end-gas volume reductions around the periphery of the chamber, allowing CR and/or MAP values to be increased before exceeding the DL. This enabled the test engine to achieve relatively high efficiencies and BMEP when compared to larger bore engines. Consequently, the test engine recorded a peak of 37% TH and 13 bar BMEP in NA form with turbocharging achieving a peak of 32% TH and 25 bar BMEP. Table 8.4 highlights the maximum BMEP values achievable for the test engine and larger bore engines over varying modes. In NA form, the test engine recorded a peak BMEP of 1,260 kPa at mid-range speeds, compared to values in the order of 1,000 - 1,100 kPa for larger bore engines [11, 25, 72, 100, 120]. The BMEP performance discrepancies are largely associated with the CR differences previously discussed, further highlighting the performance potential for small cylinder engines.

249

The test engines spark timing over the experimental domain is generally more advanced, relative to larger bore engines [25, 148]. Spark timing variations between small and large cylinder engines are primarily caused by different operating speed ranges associated with the varying bore size and bore/stroke ratio. Furthermore, smaller bore engines generally operate with little, if any spark retard for knock compensation, due to the reduced propensity for knock to occur [11, 14, 15] as previously discussed in Chapter 7.4.3. Peak TH values are also compared to modern larger bore engines operating at stoichiometric conditions, as outlined in Table 8.4. Generally, typical peak efficiencies for modern automobile engines occur at low engine speeds (2,000 3,000 rev/min) and range between 30 - 34%. Smaller bore engines inherently suffer from increased heat losses when compared to larger engines due to the higher surface to volume ratio [95], which results in reducing small engine efficiencies [220]. However, the test engines improved efficiency is attributed to the extension of the KL due to the reduced bore size and fast burn combustion chamber. This allowed for CR increases, which offset the higher heat losses. It is also noted, that the test engine recorded peak efficiencies at higher engine speeds (4,000 - 5,000 rev/min) but similar MPS, thus highlighting the engines high MECH (Chapter 7.5.4). Hence, as demonstrated by the test engine, it is possible for reduced bore size engines to match or exceed the efficiency of larger bore engines. However, this was only possible after CR optimization to compensate for the higher levels of dissociation, friction and heat losses associated with the smaller cylinder capacity. When comparing NA air consumption rates in Table 8.4, peak VOL values near 100% occur for both the test engine and the typical larger bore engine [11, 25]. However, test engine values are attained at higher engine speeds where the engine is less susceptible to knock (Chapter 7.2). It is also noted that the test engine values were achieved with narrow valve overlap (35), similar to production based passenger vehicle engines [39]. In the boosted modes, the potential exists for small bore engines to have higher VOL for a given CR due to the possible extension of the operating limits as previously described.

250

When comparing MECH in Table 8.4, similar peak values are recorded between the test engine and typical larger bore engines for similar engine designs (three ring piston). In a larger bore engine, Khan [120] recorded a 93% peak MECH at 1,250 rev/min (5 m/s MPS), which is comparable to values attained from the NA test engine at similar MPS (3,000 rev/min engine speed). The test engine also shows slight improvements in MECH when turbocharging is employed due to further reductions in pumping losses as the MAP is increased (Figure 7.11). The high

MECH results of the test engine differ to previous findings from non-current smaller
engines [220], which document that smaller engines generally have reduced MECH when compared to larger bore engines due to higher friction and pumping losses associated with the higher engine speeds. The peak MECH of the larger bore engine can be matched at similar MPS due to the smaller test engines modern design components, which reduce the energy lost to friction.

8.8 Extension to a Future Application - Feasibility for a 1.25 Liter Replacement

To illustrate the potential of downsized engines in passenger vehicles, a preliminary feasibility study of replacing a 1.25 L NA engine with a downsized boosted engine with similar brake power was completed [12]. The downsized option used in comparisons is the test engine operating in the TC mode (optimized for the flow restricted condition, CR = 10), with full and part load test data (Chapter 7) used in comparisons. The purpose of the comparisons is to determine if the performance of downsized engines can match larger counterparts and to find what fuel efficiency benefits and CO2 emission reductions are probable. The objective of these comparisons is to show that there are opportunities to improve engines found in the compact sized regular passenger vehicle class, thus highlighting the potential to downsize what are already considered small engines in todays marketplace.

251

The example taken for this comparison is a Ford 1.25 L Duratec engine fitted to the 2007 Fiesta Mark VII series, with engine specifications given in Table 8.5 [73]. A summary of the Fiesta Duratec performance, efficiency and emissions data is given in Table 8.6 [73, 193]. It is noted that the supplied data is for tests conducted using 95-RON pump gasoline compared to the 98-RON used in the test engine. The Ford Fiesta was chosen for comparison, being a leader in the small vehicle class in Europe with three decades of past and current dominance in the UK market [74]. The Duratec engine also shares design similarities when compared to the test engine. If the smaller engine were to be installed into the Fiesta passenger vehicle, there are several foreseeable obstacles that would require attention. The main obstacle is the NVH due to the higher engine speeds and odd fire inline twin configuration. The uneven torque pulses due to the engine configuration may also reduce the driving feel. However, the possibility of an even fire twin configuration with balance shaft development may produce a similar driving feel to a lower speed four cylinder engine, due to the increased firing frequency at the elevated speeds. Hence, the torque pulsations for an even fire two cylinder engine are identical to a four cylinder engine operating at half the engine speed. Considering emissions, control would most likely not be more difficult with the test engine as both engines share design similarities including matching fuel supply systems. The Fiesta exhaust aftertreatment (TWC) could be implemented on the test engine enabling both engines to meet Euro IV emission standards. However, exhaust aftertreatment testing was out of the scope of the project. Furthermore, the higher exhaust backpressures due to the TWC may also reduce turbocharger performance. Additionally, the heavy fuel enrichment used by the test engine near WOT would not pose major emission problems as there is no requirement for emission control near WOT in present standards (Chapter 7.2). Thus the present enrichment knock control strategy should be satisfactory as only low boost levels are needed over the NEDC (Figure 8.11).

252

Table 8.5: Specifications for the Ford Duratec engine fitted to the 2007 Fiesta Mark VII series, which is compared against the TC - PFI test engine. 2007 Ford Fiesta - Mark VII Series ENGINE TYPE CAPACITY BORE x STROKE COMPRESSION RATIO COMBUSTION CHAMBER VALVE ACTUATION INDUCTION FUEL EMISSION CONTROL Ford Duratec In-line 4 cylinder, 4 stroke SI, Liquid-cooled, Aluminum head/ block 1.242 L 71.9 x 76.5 mm 10:1 Pent roof central spark plug 16-valve DOHC NA sequential PFI 95-RON pump gasoline Closed-loop TWC, Euro IV compliant

8.8.1 Performance
Figure 8.10 compares the WOT performance of both engines, achieved via normalizing the engine speeds to accommodate for the different speed ranges. Normalizing allows comparisons to be made, which could realistically be achieved through transmission or driveline ratios to obtain matching vehicle speeds. Three WOT performance curves are shown in Figure 8.10; a baseline for the Ford Fiesta and two possible WOT operating conditions for the smaller test engine. It is noted that the test engine remained unchanged between both operating conditions with equal CR and turbocharger systems. include: (1) Limited MAP by a wastegate (Maximum MAP = 170 kPa) (2) Limited airflow by an intake restriction (Intake restriction = 20 mm in diameter) As previously described, the test engine was optimized for the limited airflow (2) condition, resulting in a CR of 10 and turbocharger matching to suit this condition. Hence, results presented for the limited MAP (1) condition could be further improved. However, the purpose of displaying two operating performance curves is to highlight the performance opportunities downsized engines can achieve. The varying operating conditions for the test engine are a result of different intake flow conditions, which

253

300 250 MAP (kPa) @ WOT 200 150 100 50 0.2

1.25L NA FORD FIESTA 0.43L TC MAP LIMITED (1) 0.43L TC FLOW LIMITED (2)

0.4

0.6 Engine Speed (Normalised)

0.8

2600 2200 BMEP (kPa) @ WOT 1800 1400 1000 600 0.2

1.25L NA FORD FIESTA 0.43L TC MAP LIMITED (1) 0.43L TC FLOW LIMITED (2)

0.4

0.6 Engine Speed (Normalised)

0.8

60 Brake Power (kW) @ WOT 50 40 30 20 10 0 0.2 0.4 0.6 Engine Speed (Normalised) 0.8 1 1.25L NA FORD FIESTA 0.43L TC MAP LIMITED (1) 0.43L TC FLOW LIMITED (2)

Figure 8.10: WOT performance comparison between the Ford Fiesta 1.25 L NA engine and the smaller TC test engine used in experiments. (Upper): MAP. (Middle): BMEP. (Lower): Brake Power. Two performance curves are shown for the smaller test engine; (1) Limited MAP by wastegate (170 kPa), (2) Limited airflow by a restriction (20 mm).

254

As can be seen from the middle diagram of Figure 8.10, the test engine recorded peak BMEP values of 25 bar. This performance was achieved at mid-range engine speeds using 270 kPa MAP for the limited airflow (2) condition. This equates to a 2.5 fold increase in peak BMEP when compared to the Fiesta engine. Consequently, the peak power of the Fiesta engine is matched or exceeded over both intake flow conditions, with a 66% reduction in swept capacity. WOT performance comparisons are also made across the normalized speed range, as peak power is seldom used in general driving patterns as demonstrated by the drive cycle time frequency distribution of Figure 8.11 [61, 121]. The MAP limited (1) smaller test engine, with 0.7 bar boost is shown to match or exceed Fiesta brake power over half the normalized speed range. However, Fiesta power is not met at low engine speeds due to the turbochargers inability to supply the required boost to increase performance. The smaller engines lack of low speed performance could be overcome with improved turbocharger matching, as the system was optimized for the limited airflow (2) intake condition. Dual scroll or VTG turbocharger technology together with valve overlap through scavenging techniques could also be implemented to improve low speed performance [9, 123, 136, 143, 250, 252]. Furthermore, the vehicle transmission could be designed with changes to match the reduced engine inertia. This could include higher first gear ratios in manual transmissions or higher rates of clutch/torque converter slip for direct shift gearboxes (DSG) and automatic transmissions to accommodate the reduced low speed performance. Whichever strategy is used, some compromises are needed to give good driving feel in the low speed WOT domain, which is commonly used in initial vehicle acceleration.

8.8.2 Fuel Consumption and CO2 Emissions

Fuel consumption and resulting CO2 emissions are also compared between both engines. It is noted that CO2 levels are calculated via fuel consumption, which assumes all fuel carbon consumed by the vehicle is eventually converted into CO2. Comparisons between both engines for equal power outputs near WOT show that the larger displacement NA engine has a fuel consumption advantage for equal CRs due to the reduced levels of fuel enrichment and spark retard needed to

255

control knock. At reduced load conditions, the smaller engine shows a reduction in BSFC due to the improved MECH associated with the greatly reduced pumping losses. The effects at idle conditions and over the NEDC are also investigated, with comparisons assuming the test engine is fitted to the Fiesta chassis as outlined in Table 8.6. At idle speeds, a 62% reduction in fuel consumption, and hence CO2 emissions, is recorded for the smaller test engine. Vehicle mass out emissions are also reduced by a similar magnitude due to the engine size, prior to catalyst lightoff. Start up emissions contribute a significant minority of emissions in heavily populated areas, thus highlighting the potential for this type of powertrain in these areas. Idle benefits are caused by the smaller engines ability to run at lower MPS. A minimum MPS of 1.9 m/s corresponding to 1,000 rev/min was achieved while still maintaining adequate idle stability [98, 196].
Table 8.6: Possibilities of adapting the smaller test engine into the Fiesta chassis and the effects on vehicle performance, fuel consumption and CO2 emissions against the standard OEM vehicle. ENGINE COMPARISONS (2007 FORD FIESTA CHASSIS) Duratec (Standard OEM)
Capacity Configuration Induction Mass Vehicle (kerb) Powertrain Performance Max. Power Max. Torque Fuel Consumption* Idle Urban Extra Urban Combined (NEDC) CO2 Emissions* Idle Combined (NEDC) * EEC Directive 1999/100/EC 1.25 L Inline 4 cylinder NA - PFI

Test Engine-TC Mode (retro-fitted)

0.43 L Inline 2 cylinder TC - PFI MAP limited to 170 kPa (1) See Table 8.7 84 kg (98-RON) 53 kW - 9,000 rev/min 63 Nm - 7,000 rev/min 0.3 L/hr - 1,000 rev/min Table 8.7 0.7 kg/hr Table 8.7

1,096 kg 180 kg (95-RON) 55 kW - 5,200 rev/min 110 Nm - 4,000 rev/min 0.8 L/hr - 800 rev/min 8.2 L/100 km 4.7 L/100 km 6.0 L/100 km 1.9 kg/hr 142 g/km

256

Lower idle speeds are achievable with further development, which would further reduce fuel consumption and emissions. However, this is dependent on fuel injectors with an improved turn-down ratio to improve combustion stability [24, 121]. Other factors include the crankshaft velocity and vibration effects due to the unequal firing spacing of the inline twin configuration. Hence, it is doubtful that the larger engines idle speed of 800 rev/min could be matched, without balancing improvements attainable with further development.

Combined (NEDC)
120 100 Velocity (km/h) 80 60 40 20 0 0 200 400 600 Time (s)
1.0
Total Length Maximum Speed Total Duration Average Speed Idle Fraction 11.007 km 120 km/h 1220 s 32.5 km/h 27%

800

1000

1200

NEDC
0.8

Engine Torque (Normalized)

0.6
0.8 0.4 5 2. 1 .5 5 . 2

4 0. 0.8 1.5

0.4

2. 5

0.2

1. 5 .8 0

0.4 0.4

0.0

0.2

0.4

0.6

0.8

1.0

Engine Speed (Normalized)

Figure 8.11: NEDC operating points for the Ford Fiesta chassis, used to compare fuel consumption and CO2 emissions for the OEM 1.25 L engine and the downsized 0.43 L test engine used in experiments. (Upper): Combined Urban and Extra Urban drive cycle forming the NEDC [61]. (Lower): Generated time frequency distribution for the NEDC [121].

257

Quasi-steady analysis is now reported for the Euro NEDC for both engines, with the test cycle characterized by an urban/extra-urban driving mix. Vehicle and corresponding engine operating points are given in Figure 8.11 [61, 121], which displays the torque-speed, time frequency distribution for the NEDC. The vehicle transmission is assumed adjusted so that both engines produce matching vehicle speeds, thus allowing engine speed normalization over the drive cycle frequency matrix. Consequently, a downside not clearly seen due to the speed normalization is the smaller engines increased speeds over the drive cycle. The increased engine speeds increase friction losses and hence increase fuel consumption, highlighting the potential for further improvements if engine speeds can be reduced. A fuel consumption advantage as a consequence of installing the smaller engine into the Fiesta chassis is caused by the reduced vehicle mass, with the effects analyzed in Table 8.7. Three configurations are analyzed, as listed: (A) (B) (C) No change in vehicle mass Reduced vehicle mass due to the smaller engines mass Reduced vehicle mass due to engine and chassis mass changes

Vehicle mass effects due to the engine (Configuration B) are based on values listed in Table 8.6, resulting in an approximate 100 kg reduction. This is primarily caused by halving the number of cylinders. However, further vehicle mass reductions are also achievable (Configuration C) due to the possible chassis weight reduction to support the smaller engine after redesign. This involves repackaging the front of the vehicle to suit the smaller engine. [29]. A consequence of reducing the vehicle mass is the effect on road power required to maintain the correct vehicle speed over the NEDC. These effects are based on previous data [98, 220] and documented in Table 8.7. It is noted that reducing the vehicle mass by 10% correlates to only a 5% reduction in the required road power. Furthermore, equal reductions in road power do not correspond to equal fuel savings. This is due to the higher throttling needed to produce less road power for a fixed vehicle speed. The lower MAP causes higher pumping losses However, accurately quantifying the reduction is difficult, with estimations based on empirical data

258

and hence engine operation at a reduced efficiency point. This effect is not as prevalent in TC engines because engine operation is at points where pumping losses are already low, resulting in only minor changes in engine efficiency over the drive cycle. Further fuel saving potential also exists as experiments were conducted with varying (Chapter 7.5.1) due to the intended FSAE application. Hence, more accurate calibration to stoichiometric conditions would produce overall fuel and emission benefits, as shown in Table 8.7 (Configuration C=1). Recalibration is also required for efficient TWC operation over the drive cycle.

Table 8.7: Vehicle comparisons for several configurations involving the Fiesta chassis fitted with both OEM and smaller engines. The effects of vehicle mass on the power required to drive the vehicle at equal speeds over the NEDC and the resulting effects on fuel consumption and CO2 emissions. Vehicle Configuration Baseline
OEM production Fiesta

A
Fiesta chassis with smaller engine

C=1
Mod. Fiesta chassis with recalibrated engine

Fiesta Mod. Fiesta chassis with chassis with smaller smaller engine engine

(NA 1.25 L)

(TC 0.43 L)
No vehicle mass reduction

(TC 0.43 L)

(TC 0.43 L)

(TC 0.43 L)
Vehicle mass reduction due to engine and chassis, =1 900 50 kg 196 50 kg 0.89 y

Vehicle mass Vehicle mass reduction due reduction due to engine only to engine and chassis 1,000 20 kg 96 20 kg 0.95 y 900 50 kg 196 50 kg 0.89 y

Vehicle Mass Vehicle Mass Reduction Road Power Required [29] Fuel Consumption (NEDC) CO2 Emissions (NEDC) Fuel Consumption & CO2 Benefit

1,096 kg NA y 6.0 L/100 km 142 g/km Baseline

1,096 kg 0 y

*5.4 L/100 km *5.2 L/100 km *4.9 L/100 km **4.7 L/100 km 128 g/km 10% 123 g/km 14% 116 g/km 19% 111 g/km 22%

y = power required to maintain equal vehicle speeds over the NEDC * Fuel consumption from raw experimental results (Chapter 7.5.2) ** More accurate calibration to stoichiometric conditions (calculated)

259

Furthermore, it is also noted that the engine was optimized for the limited airflow (2) intake condition with MAP values reaching 270 kPa, resulting in a CR of 10. Hence, the potential exists to increase the CR as MAP values would not need to exceed 170 kPa in order to match the OEM Fiesta performance for the intended application. An increase in CR would improve engine efficiency, as documented for this particular engine [11]. As further improvements are possible, Table 8.7 serves only as a guide to determine the feasibility of replacing the larger engine with the smaller option. It can be concluded from Table 8.8 that a 22% reduction in fuel consumption and CO2 over the NEDC may be achievable by implementing the smaller engine into the Fiesta chassis. The result is attributed to a combination of factors, including engine operation at higher efficiencies and vehicle mass reductions. The efficiency benefits are associated with operating the smaller TC engine at higher MAP when compared to the larger engine, which reduces pumping losses and improves MECH [98]. Hence, the smaller TC engine operates closer to peak efficiency over the NEDC which results in reduced fuel consumption, even though both engines produce similar peak efficiencies (~30%).
Table 8.8: Summary of fuel consumption and CO2 emission benefits arising from replacing the larger 1.25 L NA engine found in the Ford Fiesta with the smaller 0.43 L TC test engine. Compiled from Tables 8.6 and 8.7. Relative fuel consumption and CO2 reduction
At Idle Over the NEDC 62% 22%

8.9 Summary
In this chapter, performance, efficiency and emission experimental results are compared with those from a variety of different engines and engine configurations. When comparing the test engines fuel delivery systems, results show equal performance potential between carburetion and PFI. However, a 3% relative improvement in peak TH was observed with PFI due to the improved

260

mixture homogeneity as indicated by the reduced HC emissions. However, HC emissions were shown to be the most significant difference between both fuel delivery systems, with a reduction in the order of 20% at WOT when implementing PFI. These performance, efficiency and emission results confirm previous findings from larger engines and is one of the factors contributing to PFIs universal adoption over carburetion. The effects of the intake restrictor have been outlined and compared for odd and even fire engine configurations. In the NA mode, it was shown that flow restriction begins to limit airflow and thus power at higher engine speeds for an even fire configuration. This is due to the odd fire engines piston motion causing intermittent sonic flow pulsing through the intake restriction, which consequently reduces airflow. This effect was shown to be significantly reduced with turbocharging as the engine is fed via the compressor, which induces airflow through the nozzle over all CA degrees. Furthermore, the odd fire configuration was shown to have greater airflow variation between cylinders, due to the charge robbing effects associated with the uneven induction spacing. This variation was evident in both NA and TC modes. The test engine was compared to other engines used in FSAE vehicles as the original intent of this project was to achieve success in FSAE competition. Results and comparisons show that the initial project objectives have been achieved, with the TC mode recording a 7% improvement in maximum power over previous FSAE engines. Due to the limited airflow, power gains are attributed to improved TH associated with enhanced combustion (shorter burn duration and hence smaller MBT-ST) and pumping loss frictional improvements associated with the reduced engine speeds and smaller capacity. Turbocharging has also enabled a 2.5 fold increase in peak BMEP, with near constant power achieved over half of the speed range due to the airflow limit. for much of the FSAE competition. The test engines superior performance characteristics are further demonstrated when compared to other small and large engines. Across all modes, it is evident that the test engine can operate with considerably higher CR for a given MAP, with These benefits had vehicle advantages in maximizing acceleration, reducing fuel consumption and minimizing gear shifting

261

results showing potential for engine operation at a CR exceeding 13.

The

extension of the operating limits has enabled the test engine to achieve high efficiencies and BMEP. The test engine achieved 25 bar BMEP, believed to be the highest recorded value for small engines using pump gasoline. Across the varying modes, the test engines performance, efficiency and CO2 benefits were found to have the potential to match or exceed typical larger bore engines found in passenger vehicles. However, this was only possible after CR optimization, which compensated for the higher levels of dissociation, friction and heat losses associated with the smaller cylinder capacity. The feasibility of replacing larger engines found in compact sized regular passenger automobiles with small engines (~0.5 L) has also been investigated in this chapter. A case study was performed to determine the feasibility of replacing a larger 1.25 L NA engine found in the 2007 Ford Fiesta. Results show that the performance of the larger engine could be readily matched with the smaller TC unit, with a 66% reduction in engine capacity while using no complex knock preventative methods. This indicates the potential for the swept capacity of all NA engines fitted to automobiles to be halved with no loss in performance. Analysis performed when assuming the downsized test engine is fitted to the Fiesta chassis shows a 22% best reduction in fuel consumption and CO2 emissions over the NEDC, including the reduction to 62% at idle conditions. These benefits over the NEDC are shown to be a consequence of operating the test engine closer to peak efficiency, together with engine and chassis mass reductions. The reduction in CO2 would shift the vehicle under the 2012 Euro target of 120 g/km.

262

CHAPTER 9
P O W E R E D B Y W A T T A R D

Combustion Analysis

9.1 Introduction
This chapter comprises the test engine combustion analysis results for both NA and boosted modes. The effects of varying CR, engine speed and fuel delivery on combustion parameters are investigated for NA operation. Additionally, the TC results are used to document the MAP effects. Results give valuable insight into small cylinder combustion, relevant to modern engine designs following the trend of decreasing capacity. The combustion results for the small cylinder bore engine are also compared to typical larger bore engines, with several important differences which affect combustion highlighted. To achieve these analyses, post processing of the raw cylinder pressure data was completed, with the technique described. to the test engine and its experimental data. The effects of knock on combustion and engine components are also discussed, together with implemented and proposed strategies used to avoid this uncontrolled phenomenon. Combustion analysis was completed using a validated two-zone quasi dimensional model, which was modified to apply

263

9.2 Combustion in Small Engines

There has been little documented in the literature concerning combustion in small bore (<70 mm) SI engines. Furthermore, most documented is generally for larger bore engines (~85-95 mm) similar to those found in passenger vehicles [27, 92, 98, 116]. However, combustion has been found to be the dominant limitation in extending the operating limits for downsized applications [12, 14, 15] (Chapter 7). Hence, understanding the combustion effects in small cylinder engines will assist in their future development. Several important differences may be expected as a result of the smaller capacity when compared to larger bore engines found in passenger vehicles. include: Increase in wall effects on tumble and swirl dissipation Reduction in squish velocities due to the need to maintain a fixed minimum squish thickness (squish height) Reduction in the turbulent length scale Increased heat losses from the increased surface to volume ratio Increased flame quenching area resulting from smaller combustion chamber clearance distances This research, for the first time, presents combustion results covering a wide range of varying parameters in a single small test engine. include varying engine speed, MAP, CR and fuel delivery. These parameters When compared to These

larger engines, the main parameter differences include the very large engine speed (2,000 - 10,000 rev/min), CR (9 - 13) and MAP (55 - 270 kPa) ranges.

9.3 Data Post Processing

The pre-processed data was exported from WaveView as outlined in Chapter 4.3.5, which allowed data post processing for the combustion analysis. The first stage involved conversion from the time to CA domain. One obstacle encountered in this stage involved the limited crank position recordings and the varying crankshaft velocity throughout the cycle, as shown in Figure 9.1.

264

6 5 Pressure (MPa) 4 3 2

10200

10000

9800 1 0 0 60 120 180 240 300 360 420 480 540 600 660 720 Crank Angle (deg.) 6 5 800
CYL 1 CYL 2

9600

Pressure - Cyl 1 Pressure - Cyl 2 Instantaneous Torque Average Torque

600 400 Torque (Nm) Peak Stress (MPa) 200 0 -200 -400

Pressure (MPa)

4 3 2 1 0 0 60 120 180 240 300 360 420 480 540 600 660 Crank Angle (deg.) 0.8 0.7

-600 -800 720

Twist Angle Peak Stress

900 750 600 450 300 150 0 1000

Twist Angle (deg.)

0.6 0.5 0.4 0.3 0.2 0.1 0 0 200 400 600 Torque (Nm) 800

Figure 9.1: Crankshaft effects at high engine speeds throughout one cycle. NA Carburetion, 10000 rev/min, WOT, CR = 10. (Upper): Velocity fluctuations. (Middle): Torque fluctuations. (Lower): FEM analysis crankshaft elastic deformation.

265

Crankshaft Velocity (rev/min)

Pressure - Cyl 1 Pressure - Cyl 2 Instantaneous Velocity Average Velocity

10400
CYL 1 CYL 2

The crankshaft position recordings were limited due to the small diameter encoder wheel which was required for high speed operation as it exhibited low rotational inertia. Furthermore, the crankshaft velocity variations throughout the cycle were exacerbated by the unequal firing interval. To determine the crankshaft velocity profile throughout the cycle, a cubic spline was fitted to the known velocities and corresponding CA positions [126]. The cubic spline fit was more accurate than linearly interpolating between encoder points as it compensated for the speed variations throughout the cycle. This technique was then applied to each cycle, enabling the reference cylinder pressure to be accurately pegged to the crankshaft position. At high engine speeds, little work has been published in the area of combustion analysis. Thus, some validation of component flexure under high combustion and inertia loading was completed to ensure accurate results. 3.5.1). This work was completed due to the deflections anticipated from the load analysis (Chapter Initially, the piston position was analyzed, with results showing that although significant deflection occurred during valve overlap, the deflection during the combustion event was significantly reduced as the combustion and inertia forces somewhat cancelled (Figure 3.6). Consequently, piston movement due to component deflection had negligible effect on indicated and combustion analysis data. At peak loading, the crankshaft angular elastic deformation was analyzed as shown in the middle and lower diagrams of Figure 9.1. Analysis was completed to determine the crankshaft twist angle between the encoder wheel and first cylinder, where cylinder pressure was measured. Results show a peak angular displacement of 0.25 along the entire length of the shaft at peak loading, equating to a twist angle of approximately 0.06 between the encoder and the crankpin ( of the shaft length). (Appendix J.8). Hence, crankshaft angular deformation was neglected as the twist angle was lower than the defined CA phasing accuracy

9.3.1 E-CoBRA
Combustion analysis was completed using E-CoBRA [91], a two-zone quasi dimensional model developed and validated by Hamori [92]. E-CoBRA was

266

adapted to incorporate the modern four valve, pent roof geometry of the test engine. This involved the calculation of the flame area burnt gas volume amongst other parameters as a function of the flame radius in relation to the spark plug location. The software was then applied to the experimental cylinder pressure versus CA data, enabling combustion parameters to be estimated, as outlined in Appendix N. E-CoBRA outputs included: Indicated power, torque, thermal efficiency and specific fuel consumption Net/gross indicated mean effective pressure (IMEP), friction and pumping loss data Pressure (P ) and temperature (T ) Mass fraction burned (MFB) and mass fraction burned rate (MFBR) Laminar (SL) and turbulent flame speed (ST,a) and flame speed ratio (FSR) Turbulence intensity (u ) estimation Knock initiation, duration, peak amplitude and intensity Combustion data versus CA, at 50% MFB and peak values Following conversion of the pressure trace to the CA domain, data was sampled at 0.5 CA increments by interpolating between data points using the double parabolic algorithm [126]. This interpolation was required for the E-CoBRA input. Figure 9.2 displays the effects of the re-sampling for the worst case heavy knocking cycle.
12 10 Pressure (MPa) 8 6 4 2 0 -180 Raw Resampled

-140

-100

-60

-20

20

60

100

140

180

Crank Angle (deg.)

Figure 9.2: The effects of re-sampling the raw pressure trace to 0.5 CA increments, required for the E-CoBRA software input. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11.

267

Forty consecutive cycles has previously been shown to be satisfactorily representative of the combustion process for SI engine analysis [92]. However, as the test engines combustion variability was inherently low over the majority of the domain (< 2% CoV of IMEP) due to near stoichiometric operation [98], a 10 cycle analysis was deemed adequate. Figure 9.3 highlights the low combustion variability over 40 consecutive cycles at high engine speed. The reduced cycle analysis enabled a large number of data points to be investigated, even with limited computer resources. A 10 cycle analysis was also representative of the burn characteristics due to the low variability. In addition, when using so few cycles, each cycle has a greater impact on the CoV of IMEP. combustion data presented has been computed over 10 cycles.
7 6 Pressure (MPa) 5 4 3 2 1 0 0 50 100 Volume (cm 3) 150 200 250

Hence, all

CoV of IMEP = 1.1%

Figure 9.3: Pressure-volume diagram for 40 consecutive cycles highlighting the low combustion variability. NA - Carburetion, 10500 rev/min, WOT, CR = 13.

9.4 NA Combustion
The NA combustion effects for varying CR, engine speed and fuel delivery are investigated at WOT (Figures 9.4 to 9.8). Figure 9.4 presents varying MFB durations in CA and time domains versus engine speed. The 0-10% MFB duration indicates the initial kernel growth period to a fully developed flame, while the 1090% is indicative of the main energy release phase and usually comprises the last 30% of flame travel. MFB results are computed using the Rassweiler and Withrow method [40, 182, 183], which assumes all fuel is consumed (MFB = 1).

268

60 55 0-10% MFB (CA degrees) 50 45 40 35 30 25 20 4000 5000 6000 7000 8000

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 0-10% MFB (ms) 0-90% MFB (ms) 10-90% MFB (ms)

9000

10000

Engine Speed (rev/min) 60 55 10-90% MFB (CA degrees) 50 45 40 35 30 25 20 4000 5000 6000 7000 8000 9000 10000 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 Engine Speed (rev/min) 85 80 0-90% MFB (CA degrees) 75 70 65 60 55 50 45 4000 5000 6000 7000 8000 9000 10000 2.2 1.9 1.5 1.2 0.8 0.5 0.1 -0.3 -0.6

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

Engine Speed (rev/min)

Figure 9.4: Combustion burn duration effects for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): 0-10% MFB. (Middle): 10-90% MFB. (Lower): 0-90% MFB. Solid lines associated with CA axis, dashed lines with time axis.

269

Figures 9.4 to 9.7 display the combustion effects for varying CR. The results for the initial burn fraction (0-10%) in the upper diagram of Figure 9.4 show similar combustion durations for equal CR over the engine speed and fuel delivery range. At the higher CR, the initial burn rate decreases by 3-5 CA corresponding to spark timing reductions (Chapter 7.4.3) across all engine speeds for any given mode. The faster initial burn rates for the higher CR correspond to a temperature increase at the point of ignition, resulting from the higher pressure at ignition. This results in faster laminar flame speeds, which leads to faster flame kernel development. Conversely for the main burning phase (10-90%) in the middle diagram of Figure 9.4, the higher CR results in reduced burning rates. The reduced burning rates are caused by a combination of factors, which include: The higher surface to volume ratio, which leads to higher heat losses to the more proximate surfaces and hence reduced flame speed The reduced turbulent length scale in the more constrained cylinder volume around TDC The retarded spark timing, which causes the last part of the burn to occur in a larger cylinder volume. This leads to reduced late-in-cycle pressures and temperatures, which slow the flame propagation There are also secondary effects which slow the main 10-90% burning phase. The higher peak pressures due to the increased CR cause dissociation effects, corresponding to the higher peak cycle temperatures. Consequently, the amount of dissociated CO2 (largely to CO) and H2O (largely to H2) increases. This reduces the temperature from that expected without dissociation, reducing the laminar flame speed and consequently the turbulent flame speed, which increases the burn duration. A further consequence of this is that combustion goes to The late-in-cycle energy release results in less completion later in the cycle.

piston work or IMEP than could have been achieved if the mixture was immediately consumed by the flame. Considering the various effects present during the initial and main phases of combustion, the total burn times (0-90%) show almost no variation with CR as displayed in the lower diagram of Figure 9.4.

270

1 0.8 0.6 MFB 0.4 0.2 0 -40

NA-CARBS (CR = 13) [30 BTDC] NA-CARBS (CR = 10) [35 BTDC] NA-PFI (CR = 10) [35 BTDC]

-30

-20

-10

10

20

30

40

50

Crank Angle (deg.)

0.06 0.05 MFBR (/deg.) 0.04 0.03 0.02 0.01 0 0

NA-CARBS (CR = 13) NA-CARBS (CR=10) NA-PFI (CR=10)

0.1

0.2

0.3

0.4

0.5 MFB

0.6

0.7

0.8

0.9

0.06 0.05 MFBR (/deg.) 0.04 0.03 0.02 0.01 0 -40

NA-CARBS (CR = 13) [30 BTDC] NA-CARBS (CR=10) [35 BTDC] NA-PFI (CR=10) [35 BTDC]

-30

-20

-10

10

20

30

40

50

Crank Angle (deg.)

Figure 9.5: Combustion burn effects for alternative CRs and fuel delivery. NA, 10000 rev/min, WOT, = 0.9 0.02, MBT spark timing [ ]. (Upper): MFB versus CA. (Middle): MFB versus MFBR. (Lower): MFBR versus CA.

271

Figure 9.5 graphically displays the effects on combustion burn rates due to alternative CRs and fuel delivery systems. The effect of the faster initial burn and slower later burn phases for the increased CR can be seen in the in the MFB and MFBR data. The implied turbulence effects due to the CR variation are also analyzed in Figure 9.6. The higher CR has led to increased turbulence intensities due to increased squish velocities associated with the reduced turbulent length scale. The increased turbulence levels have contributed to faster initial flame velocities, which has resulted in faster flame kernel development as verified by the 0-10% MFB data of Figure 9.4. Figure 9.7 displays the effect of the increased flame velocities for the higher CR. The increased velocities due to the CR increase are not visible at 50% MFB in the upper and middle graphs of Figure 9.7. This is due to the differing burning profile of the high CR with higher initial burning rates rapidly consuming the charge, hence the burn rates begin to decrease prior to consuming 50% of the mass, as seen in the middle graph of Figure 9.5. Consequently, as the CR increases, the peak actual flame speed and peak FSR also increase as confirmed in the lower diagram of Figure 9.7. Furthermore, the effects on laminar flame speeds can be deduced from Figure 9.7, with results showing little variation at 50% MFB due to the CR change. Hence, the laminar flame speed is not largely affected by CR variation over the range of test CRs.

25 NA - CARBS (CR = 13) Turbulence Intensity u' (ms) 20 15 10 5 0 4000 NA - CARBS (CR = 10) NA - PFI (CR = 10)

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 9.6: Estimated turbulence intensity u for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing.

272

30 Actual Flame Speed ST,a (ms) @ 50%MFB NA - CARBS (CR = 13) 25 20 15 10 5 0 4000 NA - CARBS (CR = 10) NA - PFI (CR = 10)

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

50 Actual Flame Speed Ratio FSR a @ 50%MFB NA - CARBS (CR = 13) 40 30 20 10 0 4000 NA - CARBS (CR = 10) NA - PFI (CR = 10)

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

40

Actual Flame Speed ST,a (ms) Peak

35 30 25 20 15 10 5 0 4000

NA - CARBS (CR = 13) NA - CARBS (CR = 10) NA - PFI (CR = 10)

5000

6000

7000

8000

9000

10000

Engine Speed (rev/min)

Figure 9.7: Combustion flame speed effects for alternative CRs and fuel delivery. NA, WOT, = 0.9 0.02, MBT spark timing. (Upper): Actual flame speed @ 50% MFB. (Middle): Actual flame speed ratio @ 50% MFB. (Lower): Actual peak flame speed.

273

The effect of fuel delivery on NA combustion at WOT is displayed in Figures 9.4 to 9.7. When comparing carburetion and PFI, the initial and total burn durations (Figure 9.4) and combustion burn profiles (Figure 9.5) show little variation between both modes. This corresponds to similar MBT spark timing results (Chapter 7.4.3) across both modes of fuel delivery, supporting the concept of a similar mixture distribution and hence AFR gradient within the combustion chamber. The turbulence intensity and flame speeds also show little variation between the two modes of fuel delivery. Hence, combustion variations between the two modes are negligible at WOT, as verified by similar performance and efficiency trends in Chapter 7.

The effect of engine speed on combustion is also examined in Figures 9.4 to 9.8. Increasing engine speed results in a decrease in the time taken to consume a given mass fraction. Hence, doubling the engine speed does not double the burn duration in CA degrees (Figure 9.4). This effect is caused by an increase in turbulence levels (Figure 9.6), which promote faster flame speeds (Figure 9.7) and hence faster burning rates (Figure 9.8) [98]. Hence the turbulent to laminar FSR increases for rising engine speed, with calculated peak flame speeds reaching 40 m/s at maximum engine speeds (Figure 9.7). Note: this is not the apparent turbulent flame speed, but the turbulent flame speed with respect to the unburned gas velocity at the flame front. Although the flame speed increases, the MFBR decreases for rising engine speeds as seen in Figure 9.8, due to the reduced combustion efficiency associated with higher heat losses and dissociation at the elevated speeds.

274

1 0.8 0.6 MFB 0.4 0.2 0 0

4000 rev/min [28 BTDC] 6000 rev/min [31 BTDC] 8000 rev/min [33 BTDC] 10000 rev/min [36 BTDC]

10

20

30

40

50

60

70

80

Crank Angle (deg.) after Spark 0.07 0.06 0.05 MFBR (/deg.) 0.04 0.03 0.02 0.01 0 0 0.1 0.2 0.3 0.4 0.5 MFB
0.06 0.05

4000 rev/min 6000 rev/min 8000 rev/min 10000 rev/min

0.6

0.7

0.8

0.9

4000 rev/min [28 BTDC] 6000 rev/min [31 BTDC] 8000 rev/min [33 BTDC] 10000 rev/min [36 BTDC]

MFBR (/deg.)

0.04 0.03 0.02 0.01 0 -40

-30

-20

-10

0 Crank Angle (deg.)

10

20

30

40

Figure 9.8: Combustion burn effects for varying engine speeds. NA - Carburetion, WOT, CR = 10, = 0.9 0.02, MBT spark timing [ ]. (Upper): MFB versus CA after Spark. (Middle): MFB versus MFBR. (Lower): MFBR versus CA.

275

9.5 TC Combustion
Figures 9.9 to 9.13 display TC combustion results for varying engine speed and MAP operating conditions calculated from the in-cylinder pressure data. The IMEP and CoV of IMEP are displayed in the upper diagrams of Figure 9.9. IMEP results correspond to previous BMEP trends, allowing the MECH to be calculated as displayed in Chapter 7.5.4. The CoV of IMEP trends display the combustion variability over the test domain. Trends indicate that the combustion variability begins to increase as the engine speed surpasses 6,000 rev/min. The increased CoV of IMEP in the non-knock limited regions is caused by increases in retained residuals or internal EGR due to the higher exhaust back pressure associated with the poor turbine efficiencies, as described in Chapter 6.6. In some circumstances, small amounts of retained residuals can improve combustion stability, especially at the lower combustion temperatures [98, 227] associated with reduced loads. However, the reverse of this effect is seen here due to the already elevated temperatures associated with boost and high speed operation, which increase the levels of dissociation. Nevertheless, only a minor change in combustion variability is recorded with the increased retained residuals, with stability continuing to lie well within stable limits [92]. The CoV of IMEP also changes for rising MAP, increasing from 1-3% as the engine is operated in the knock limited regions. The increase is caused by two effects. The first involves the knock compensation strategies, consisting of increasing levels of spark retard and fuel enrichment. The spark retard and fuel enrichment affect the initial burn and flame kernel growth, which increases combustion variability [98] and decreases combustion efficiency (Figure 9.10). The second effect is the actual occurrence of knocking cycles as the knock compensation strategies only succeed in suppressing the knock below the DL at the elevated MAP, as shown in Figure 9.12. Nevertheless, the variation in combustion variability could be eliminated if knock could be controlled using other well documented proven techniques (Chapter 7.2).

276

IMEP (kPa) 260 240 220 MAP (kPa) 200 180 160 140 120
1000 1200 TURBO FLOW LIMIT 2600 2400 2200 2000 1800 1600 1400 1200
P L

CoV of IMEP (%) 260

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

TURBO FLOW LIMIT

2.8 5.0

RESTRICTED FLOW LIMIT

3.0 2.0

1.5 2.2

1.5 1.1

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
.) 0-10% MFB (deg. 260 240 14 TURBO FLOW LIMIT 220 16 MAP (kPa) 200 180 160 140 120
30 18 20 22 24
P 26 L

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) .) 10-90% MFB (deg. 260
34 TURBO FLOW LIMIT 32 RESTRICTED FLOW LIMIT 30 28 26 24 22

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

PL

28

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) .) 0-90% MFB (deg.

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Peak MFBR (/deg.) 260

TURBO FLOW LIMIT 0.030 RESTRICTED 0.035 FLOW LIMIT

260 240 220 MAP (kPa) 200 180 160 140 120
52 TURBO FLOW LIMIT 40 44 48 52 RESTRICTED FLOW LIMIT60

240 220 MAP (kPa) 200 180 160 140 120

52
P L 56

0.040

0.045

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

0.050

Figure 9.9: Combustion variability and burn effects for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.

277

The MFB and MFBR data is displayed in Figure 9.9 and 9.10. The 0-10% MFB duration (in CA degrees) decreases with rising MAP across all regions, even those which are heavily knock compensated. Initial burn effects associated with the The faster rising boost pressure are also displayed graphically in Figure 9.10.

initial burn is caused by the higher mixture density, which improves the flame initiation and hence kernel growth. Furthermore, as the engine speed increases, the 0-10% MFB data shows identical trends when compared to the NA data with increased CA durations over reduced time intervals. It is interesting to note that the increased retained residuals above 6,000 rev/min, due to the higher backpressure previously described, slows the initial burn duration at high speeds as seen in the middle diagram of Figure 9.9. extend the 0-10% burn duration. The 10-90% MFB (Figure 9.9) trends show increasing CA burn durations for rising MAP and engine speeds over the majority of the domain. This is unexpected as the increased charge density should promote faster burning rates due to laminar and hence turbulent flame speed increases. Reasons for the increased burn durations are the retarded spark timing as the boost pressure increases, resulting in combustion deteriorations associated with slow burning cycles (Figure 9.10 Lower). The reduced combustion efficiency associated with the retarded spark timing corresponds to the poor thermal efficiencies recorded as the boost pressure increases (Chapter 7.5.2), despite the pumping losses reducing (Figure 6.22). Furthermore, there is a clear switch in the 10-90% MFB trend line directions in regions where there are high retained residuals or knock compensation. This is due to further reductions in combustion efficiency as the retained residuals decrease laminar flame speeds and hence slow the flame propagation. The retarded spark timing promotes slow burning during the main energy release phase as indicated by the 10-90% burn duration data of Figure 9.9. However, this effect is offset by the faster initial burn (0-10% MFB), which results in no significant overall burn duration changes (0-90% MFB) over the majority of the test domain for rising MAP, as shown in Figure 9.9. The retained residuals result in higher rates of dilution which increase the ignition delay [98, 141, 180] and hence

278

1 0.8 0.6 MFB 0.4 0.2 0 0

TC-PFI (100 kPa MAP) [27 BTDC] TC-PFI (140 kPa MAP) [21 BTDC] TC-PFI (180 kPa MAP) [12 BTDC]

10

20

30

40

50

60

70

80

Crank Angle (deg.) after Spark

0.06 0.05 MFBR (/deg.) 0.04 0.03 0.02 0.01 0 0

TC-PFI (100 kPa MAP) TC-PFI (140 kPa MAP) TC-PFI (180 kPa MAP)

0.06 0.05 MBR (g/deg.) 0.04 0.03 0.02 0.01 0.00

0.1

0.2

0.3

0.4

0.5 MFB

0.6

0.7

0.8

0.9

0.06 0.05 MFBR (/deg.) 0.04 0.03 0.02 0.01 0 -30

TC-PFI (100 kPa MAP) [27 BTDC] TC-PFI (140 kPa MAP) [21 BTDC] TC-PFI (180 kPa MAP) [12 BTDC]

-20

-10

10

20

30

40

50

60

Crank Angle (deg.)

Figure 9.10: Combustion burn effects for varying MAP. TC - PFI, 7000 rev/min, CR = 10, spark timing [ ]. (Upper): MFB versus CA after Spark. (Middle): MFB versus MFBR (solid lines - left y axis) and MBR (dashed lines - right y axis). (Lower): MFBR versus CA.

279

Furthermore, minor 0-90% MFB changes are evident for MAP variation over the domain in regions which have higher rates of retained residuals. The dilution slows the initial and main phase energy release, resulting in an overall increase in burn durations over these regions. The 10-90% and 0-90% MFB data for speed variation follows identical NA trends, with CA burn duration increases as previously described (Section 9.4). Hence, for boosted operation, faster overall burning which improves combustion and leads to TH benefits can only be achieved if knock compensation strategies (non-optimal spark timing and fuel enrichment) are not heavily used. However, although these strategies reduce combustion and hence TH, experimental results (Chapter 7) have shown that the implementation of these knock control strategies have allowed for brake power increases via extending the MAP operating limits. Hence there is a tradeoff between peak power and good efficiency. Figures 9.9 and 9.10 display the peak MFBR. It is noted that the MFBR is a

normalized number allowing burn rate comparison over all operating conditions. However, actual mass burned rates (MBR) are dependent on the mass of charge consumed and thus increase with air consumption. The lower right diagram of Figure 9.9 shows that as the MAP increases, the peak MFBR decreases for reasons previously outlined describing the 10-90% MFB duration increases. This effect is clearly displayed at 7,000 rev/min in the lower diagram of Figure 9.10, with the retarded spark timing as the MAP increases from 100-180 kPa causing peak MFBR reductions, although the actual peak MBR increases by 60%. The MFBR engine speed effects closely follow the NA modes as previously shown in Figure 9.8 and described in Section 9.4. Figure 9.11 displays the peak pressure, temperature and flame speed effects. This data confirms the reduced combustion efficiencies when the MAP increases as previously described and shown in the MFB and MFBR data of Figures 9.9 and 9.10. The poorer main energy release burning profiles associated with the rising MAP due to the retarded spark timing and increased retained residuals causes the peak temperature and hence flame speed to reduce, however the peak pressure continues to rise due to higher charge density. This causes the peak pressure, temperature and flame speed locations to be further delayed in CA degrees.

280

Location of Peak Pressure (deg. ATDC) 260 240 220 MAP (kPa) 200 180 160 140 120
16 24 20 28
P L

Peak Pressure (kPa) 260 8000

TURBO FLOW LIMIT 32

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

TURBO 7500 FLOW LIMIT 7000 6500

RESTRICTED FLOW LIMIT

6000

PL

5500

5000

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) Location of Peak Temperature (deg. BTDC) 260 240 220 MAP (kPa) 200 180 160 140 120
7 9 11 13 5 TURBO FLOW LIMIT 3
P L

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) Peak Temperature (K) 260
2380 TURBO FLOW LIMIT 2400 2420
P

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

RESTRICTED FLOW LIMIT

2440 2460

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) Peak Flame Speed (m/s) 260
12 TURBO FLOW LIMIT

Location of Peak Flame Speed (deg. ATDC) 260 240 220 MAP (kPa) 200 180 160 140 120
5 0 10 15
PL

TURBO FLOW LIMIT

RESTRICTED FLOW LIMIT 20

240 220 MAP (kPa) 200 180 160 140 120

RESTRICTED FLOW LIMIT

15

18

P 21 L

24

27 30

5 100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 9.11: Combustion pressure, temperature and flame speed effects for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.

281

Figure 9.12 presents maximum and average peak knock amplitudes over the 10 cycle analysis. Both plots give identical trends, with the knock amplitude increasing in intensity as the boost pressure increases. The average peak knock amplitude indicates that combustion can be adequately controlled to avoid the upper KL of 4 bar (Table 4.8). However, the maximum knock amplitude data shows that spark knock is not eliminated at high boost pressures, with the occasional cycle displaying some knocking tendency, although the pressure oscillations are minor and still below the DL. These results indicate that at high boost pressures, knock can not be completely eliminated, however it can be controlled to avoid engine damage. Figure 9.12 also indicates that the knock intensity decreases with rising engine speeds due to flame speed increases (Figure 9.11 - Middle) which decrease the burn time in seconds, as previously outlined in Chapter 7.2. It is also noted that the retained residuals are higher at high engine speeds due to the increasing back pressure associated with poor turbine efficiencies (Chapter 6.6), which may play a minor role in knock propensity reduction at elevated engine speeds (Chapter 7.2).

Maximum Peak Knock Amplitude (kPa) 260 240 220 MAP (kPa) 200 180 160 140 120 100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
100 200 150 400 TURBO 350 FLOW LIMIT300 250 200

Peak Knock Amplitude (kPa) Average Peak Knock Amplitude (kPa) 260
100 TURBO FLOW 90 LIMIT 80 70 60
PL

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

RESTRICTED FLOW LIMIT

150
P L 100

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 9.12: Combustion knock amplitude effects for varying engine speed and MAP. TC PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.

282

@50%MFB - Crank Angle (deg. ATDC) 260 240 220 MAP (kPa) 200 180 160 140 120
5 15 10 20 TURBO FLOW LIMIT 25
P L

@50%MFB - MFBR (/deg.) 260

0.030 RESTRICTED FLOW LIMIT 0.035

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

TURBO FLOW LIMIT

0.040

0.045 0.050

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) @50%MFB - Burned Temperature (K) 260 240 220 MAP (kPa) 200 180 160 140 120
2300 2000 2100 2200 TURBO FLOW LIMIT 1900
P L

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
@50%MFB - SL (m/s) 260
0.3 TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

RESTRICTED FLOW LIMIT

240 220 MAP (kPa) 200 180 160 140 120

0.4

0.5

0.6

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min) @50%MFB - ST (m/s) 260 240 220 MAP (kPa) 200 180 160 140 120
21 24 27 TURBO 9 FLOW LIMIT RESTRICTED FLOW LIMIT

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

@50%MFB - FSR 260 240 220 MAP (kPa) 200 180 160 140 120 100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)
20 22 25 27 30 32 35 37
P L

40 42 TURBO FLOW LIMIT RESTRICTED FLOW LIMIT

12

15

18 P

100 4000 5000 6000 7000 8000 9000 10000 Speed (rev/min)

Figure 9.13: Combustion effects @50% MFB for varying engine speed and MAP. TC - PFI, CR = 10. Shaded areas = airflow limited regions. Cross hatched areas = knock compensated regions. PL = performance limit line at WOT.

283

Figure 9.13 displays TC combustion effects at 50% MFB as the flame is fully developed allowing examination of the laminar and turbulent flame speeds. The upper diagrams display the 50% MFB CA position and MFBR. The 50% MFB position is shown to correspond to the previous temperature, pressure and flame speed location results displayed in Figure 9.11. These results all indicate later burning as the MAP is increased due to the retarded spark timing. The retarded spark timing also causes the MFBR to decrease at 50% MFB, although the actual MBR increases due to the increasing charge density. The decreasing MFBR for engine speed or MAP increases also corresponds to a reduction in burned gas temperatures at 50% MFB, which highlights the potential to reduce thermal NOX emissions at high BMEP. Figure 9.13 also displays the laminar and turbulent flame speeds together with the FSR at 50% MFB. The laminar flame speed for gasoline decreases with rising pressures and decreasing temperatures (Appendix N.8.1). As previously shown in the NA combustion analysis (Figure 9.7), the laminar flame speed is largely unaffected by CR variation, also confirmed by Hamori [92]. However, as the MAP increases, the laminar flame speed at 50% MFB reduces due to the later combustion process (extension in the 50% MFB CA location) which reduces cycle temperatures. Although the cycle temperatures at 50% MFB decrease for rising MAP, the pressure continues to increase due to the improved charge density which enables improved power. The late-in-cycle main energy release phase due to the retarded spark timing causes the 10-90% burn duration to increase, which also results in the turbulent flame speed at 50% MFB to decrease for rising MAP. Although both the laminar and turbulent flame speeds reduce for rising MAP at 50% MFB, the FSR increases as shown in the lower diagram of Figure 9.13. For the rising MAP, the FSR increases as the laminar flame speed falls more rapidly than the turbulent flame speed, which is offset by increasing turbulence intensities that cause increased flame wrinkling and hence promote faster burning.

9.5.1 Ignition Energy

Determining the minimum required ignition energy to provide stable combustion is an important parameter as the energy significantly affects the flame kernel

284

initiation and hence growth.

The initiation and growth significantly affect the

combustion variability as the flame kernel needs to reach a radius of 10 mm before it can become fully independent of the spark energy [34, 80, 134]. At this stage, the flame kernel size is significant enough so that both large and small scale turbulence may distort the surface by wrinkling, thereby increasing the flame speed as it develops into a turbulent flame. When compared to NA engines (even with high CRs), boosted engines require higher minimum spark energies due to the relatively higher charge densities at the point of ignition. The primary impact on combustion from the increased ignition energy is the reduced flame development time due to the more rapid initial flame kernel growth. Hence, the higher ignition energy results in a larger flame kernel during the inflammation process, which thereby modestly reduces the MBT spark timing [4, 98, 207, 211]. Consequently, the ignition energy was an important parameter in the development of the high BMEP engine. Due to ignition coil reliability problems, the ignition system was de-rated to deliver 50 mJ of energy per ignition event as described in Chapter 5.3.1. This system was adequate for NA testing, however for boosted operation, combustion stability worsened with increasing levels of boost, with the effects at 6,000 rev/min displayed in Figure 9.14. At this speed with 160 kPa MAP, MBT spark timing with no knock mitigation strategy was 23o BTDC. However, for these conditions, at the optimum power mixture, combustion variability was large with 5.1% CoV of IMEP as shown in the upper diagram of Figure 9.14. It is well known that as the charge density increases, more ignition energy is required [98, 244]. The flame kernel growth is influenced by the local flow including the turbulence intensity, the mixture composition (including residuals) and the gas temperature (all of which increase with increasing charge density). Under these high load conditions, the cyclic turbulence variations are the most probable cause of the combustion variability that occurs with the marginal ignition energy. The influence of the ignition delay is evidence of this, as represented by the 0-10% burn duration of 28 CA at these operating conditions.

285

7 6 Pressure (MPa) 5 4 3 2 1 0 -180

50 mJ Spark Energy Spark timing = 23 BTDC IMEP = 1,702 kPa CoV IMEP = 5.1%

-140

-100

-60

-20

20

60

100

140

180

220

Crank Angle (deg.)

7 6 Pressure (MPa) 5 4 3 2 1 0 -180

50 mJ Spark Energy Spark timing = 17 BTDC IMEP = 1,614 kPa CoV IMEP = 7.6%

-140

-100

-60

-20

20

60

100

140

180

220

Crank Angle (deg.)

7 6 Pressure (MPa) 5 4 3 2 1 0 -180

100 mJ Spark Energy Spark timing = 18 BTDC IMEP = 1,746 kPa CoV IMEP = 1.7%

-140

-100

-60

-20

20

60

100

140

180

220

Crank Angle (deg.)

Figure 9.14: Effect of varying ignition energies on boosted combustion over consecutive tests. TC - PFI, 6000 rev/min, 160 kPa MAP, CR = 10, = 0.85 +/- 0.02.

286

Thus when the energy is doubled (Figure 9.14 - Lower), MBT spark timing reduces to 18o BTDC, the CoV of IMEP becomes more normal (1.7%) and the 0-10% burn duration reduces to 24 CA. The reduced combustion variability and increased initial burning rates are due to the more rapid flame kernel growth, with the rapid growth associated with the changing distorted wrinkled flame surface area. The middle diagram of Figure 9.14 displays the engine performance at 50 mJ of ignition energy with similar spark timing to the higher energy case shown in the lower diagram. Under these operating conditions with effectively retarded spark timing, the CoV of IMEP deteriorates (7.6%) and the 0-10% burn duration increases to 32 CA. Additionally, the IMEP falls by 7% from the MBT spark timing condition with this energy, whereas the IMEP increases by 4% for the increased ignition energy case at MBT spark timing. The flame growth for both ignition energies at similar spark timing is also compared in Figure 9.15, which highlights the improved kernel growth for the higher ignition energy case. The higher ignition energy has caused faster flame development which improves the turbulent flame speed as the flame radius expands to consume the charge. The overall burn duration (0-90% MFB) for the high energy ignition system under the conditions of Figure 9.14 is 48o CA (Figure 9.9). This is unusually small for engines of this capacity and even at 10,000 rev/min only increases to 59o CA, which is indicative of the high efficiency of the compact fast burn combustion chamber.
16 Actual Flame Speed ST,a (m/s) 14 12 10 8 6 4 2 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Flame Radius (normalised) 50 mJ Ignition Energy 100 mJ Ignition Energy

Figure 9.15: Effect of varying ignition energies on flame development over consecutive tests. TC - PFI, 6000 rev/min, 160 kPa MAP, CR = 10, 17 - 18 BTDC spark timing.

287

9.6 Knock
Chapter 2.7 outlines the background to combustion in SI engines, with spark knock (auto ignition) and surface ignition characterized as types of abnormal combustion. Engine development initially commenced without any knock detection due to a lack of expertise involving knock sensor implementation [111] and difficulty in physically locating a cylinder pressure transducer in the pent roof combustion chamber (Appendix J.7). As an alternative, the engine operator relied on audibly discerning the knock severity. This proved to be a difficult task with several engine failures occurring in the early stages of development (Chapter 5.4). Consequently, a Kistler 601-B1 cylinder pressure transducer was installed as outlined in Chapter 4.3.4. This significantly reduced development time, as any incipient of knock could be detected and avoided before engine failure. It is also noteworthy that the audible noise associated with the heavy knock condition was closer to a high frequency crackle at the elevated engine speeds, quite different to knocking resonance the author had become accustomed to in larger bore, lower speed engines. Plainly, the frequency is a function of the bore dimension to which the ear needs to become attuned. Figure 9.16 displays the cylinder pressure for a motoring and heavy knocking cycle, with the knock amplitude well above the DL previously defined (Table 4.8). The knock amplitude has been determined by filtering the raw pressure using a Butterworth filter [39, 92] as described in Appendix N.9. Running above the DL has many documented problems, ranging from component surface erosion to catastrophic failure, depending on the knock location and severity [15, 16, 98, 211]. In the experiments, it was found that high knock intensities above the DL caused significant in-cylinder component damage leading to engine failure. The pistons proved to be the most vulnerable component, with knock causing several piston land failures as outlined in Chapter 5.4 and depicted in Figure 9.17. These piston failures in turn caused further damage, sometimes irreparable, to the liners, piston rings, spark plugs and valves. Although cylinder head gaskets are commonly prone to failure from high knock intensities [66], this problem was not experienced due to the engines gasketless design (Chapter 3.6). Instead, knock above the DL would quickly result in piston land failure due to the absence of any gasket weakspot [17].

288

10 8 Raw Pressure (MPa) 6 4

9 6 3

2 0 -180

0 -3 180

-140

-100

-60

-20

20

60

100

140

Crank Angle (deg.)

Figure 9.16: Heavy knocking in the test engine well above the DL. TC - PFI, 7000 rev/min, 220 kPa MAP, CR = 11, 12 BTDC spark timing. Firing, motoring and high pass filtered pressure traces.

9.6.1 Knock Location

From examination of the failures, it was concluded that high knock intensities occurred in the end-gas on the intake side of the four valve, pent roof, combustion chamber. This was deduced from visual inspection of engine components (Figure 9.17), such as the piston land and cylinder head surface, which showed signs of pitting and erosion. This observation was further validated by the fact that the piston land failures occurred on the intake side, despite negligible differences in piston land structural integrity between the inlet and exhaust sides. Following initial inlet side piston land failure, it was theorized that the low injector location of the PFI system might be causing fuel biasing towards the exhaust side, resulting in cooler surface temperatures and hence reducing the knock propensity. This theory was dismissed after inlet side piston failure continued to occur, despite injector relocation to the inlet runner entry (Figure 5.1.4), which ensured a more homogenous in-cylinder mixture. Examination of carbon and fuel deposits on the combustion chamber surfaces also indicated higher exhaust side surface temperatures (Figure 9.18). Previous turbocharger experiments from various two valve combustion chambers with over head ignition [66, 248] have shown abnormal combustion in regions near the exhaust valve. In these designs, determining the type of abnormal combustion is difficult, as spark knock or surface ignition occurs in similar regions with similar audible noise. This is due to the spark plug location and combustion

289

Knock Amplitude (MPa)

Firing Motoring Filtered Knock

15 12

chamber geometry, which result in the last part of the burn occurring near the exhaust valve, causing spark knock [57]. The exhaust valve is also a surface Whichever the case, prone to hot spot formation leading to surface ignition.

abnormal combustion in two valve designs generally occurs on the exhaust side. The existence of exhaust side abnormal combustion, whichever the type, has been deduced from failed components including pistons and cylinder head gaskets. However, caution should be taken when deducing knock locations from gasket failures. In conventional gasket design, failures and blowouts due to abnormal combustion indicate that the failed gasket region is prone to the most severe knock intensities. However, it is feasible that these failures are a result of higher thermal stresses locally reducing the surface pressure and thus exposing the gasket as a potential weakness [17]. This is possible as the failed gasket regions generally occur between cylinders or near the exhaust side, which are prone to higher temperatures [55]. As the test engine featured a gasketless design [17], knock caused component wear and failure rather than gasket blowout, allowing more valid insight into actual knock locations. The intake side piston land failures experienced in experiments were contrary to those anticipated. It was assumed that knock would be more prominent on the exhaust side, based on the fact that exhaust side surface temperatures are higher in comparison to the intake side [55, 98] and from previous experiments in two valve engines [66, 248]. NA experiments had also shown that the intake side was relatively cooler when compared to the exhaust side, as shown in Figure 9.18. However, the higher rates of abnormal combustion experienced on the cooler intake side indicate that true spark knock and not surface ignition occurred in experiments. Consequently, with pent roof combustion chambers, knock severity is highest on the intake side due to the knock onset locations dependence on the flame front. As the flame front travels over the cooler surfaces of the intake valves, the flame speed significantly reduces resulting in the last burning end-gas occurring on the intake side. This is validated with simulation results by Teraji et al. [222], as Hence, surface temperature distribution within the shown in Figure 9.19.

combustion chamber plays a secondary role in knock location and severity [192].

290

IN

IN

INTAKE SIDE PISTON LAND FAILURE

IN

IN

CYLINDER DECK EROSION

PISTON CROWN AND LAND EROSION

INTAKE

INTAKE

Gasketless interface surface imprint

Figure 9.17: Adverse effects of knock on engine components, with highest knock intensities identified to occur in the end-gas on the intake side of the pent roof combustion chamber. (Upper): Inlet side piston land failures. (Lower): Cylinder head deck and piston erosion near the bore periphery on the inlet side.

291

Cooler Intake Side (Rich carbon deposits)

Hotter Exhaust Side (Carbon deposits)

CYLINDER HEAD

PISTON

Hotter Exhaust Side (White carbon deposits baked on valves)

Cooler Intake Side (No carbon buildup)

Gasketless interface surface imprint

Figure 9.18: Cylinder head and piston crown surfaces after engine operation, highlighting that the exhaust side is relatively hotter when compared to the intake in four valve, pent roof combustion chambers.

Temperature (K)

Integral value c - knock predictor

Figure 9.19: Simulation results from Teraji et al. [222] validating experimental findings of the intake side being the most prone to high knock intensities. Integral value is the knock predictor.

292

9.6.2 Implemented Knock Control Strategies

With spark knock highlighted as the major performance limitation (Chapter 7.2), numerous knock control strategies were employed in the attempt to extend the operating limits of the test engine. A summary of these is shown in Table 9.1, which outlines the benefits and drawbacks of each method together with where further information can be found throughout the thesis.
Table 9.1: Implemented knock control strategies and their affects. Knock Control Method Spark retard and/or fuel enrichment CR reduction Benefit 1-2 CR points Downside Expense of power and efficiency, TWC operation Compromise power and efficiency at other operating points Availability, cost Increased thermal and friction losses Increased parasitic losses Dynamic effects at high piston speeds, friction losses Reduced block structural integrity Bore distortion, no gasket weak spot Section Chapter 4.6.4 Chapter 7.2 Chapter 5.4.3 Chapter 7.2 Chapter 5.4 Appendix C.1 Chapter 3.3 Chapter 7.2 Chapter 5.4.4

Higher MAP 1-2 CR points with toluene blends (98 112 RON) Higher speeds, endgas volume reductions Unsuccessful Successful, MAP increases (180 to 270 kPa) Increased heat flux into cooling system Successful [17]

Fuel quality Reduced bore diameter Piston oil cooling

Piston ring design Open deck barrel design Head gasket design (gasketless interface)

Chapter 5.4

Chapter 3.6 Chapter 3.5.1 Chapter 5.2

9.6.3 Observations of Knocking Combustion

Figure 9.20 displays a typical 10 cycle pressure trace sequence. Notable in any sequence is a tendency for interaction between the cylinder pressure and heat transfer. The knocking sequences observed for most of the data show that after one or two consecutive knocking cycles, there are several low knock cycles.

293

12 10 1 10 Raw Pressure (MPa) 8 6 4 2 0 -100 46 22 57 93 68 14 89 35 7 10

-75

-50

-25

25

50

75

100

125

150

Crank Angle (deg.) 6 1 4 Knock Amplitude (MPa) 2 0 -2 -4 -6 20 25 30 35 40 Crank Angle (deg.) 45 50 55 60 6 2 7 3 8 4 9 5 10

7.0 6.5 6.0 5.5 5.0 4.5 -0.2

1 6

2 7

3 8

4 9

5 10

log (Pa)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

log (Clearance Volume/Actual Volume)

Figure 9.20: The effects of knock on in-cylinder pressures across 10 consecutive cycles. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Raw pressure. (Middle): Knock amplitude. (Lower): Smoothed log pressure-volume.

294

12 10 Extreme knock (60 bar) Heavy knock (30 (40 bar) Light knock (10 bar) No knock

Raw Pressure (MPa)

8 6 4 2 0 -100

-75

-50

-25

25

50

75

100

125

150

Crank Angle (deg.)

6 4 2 0 -2 -4 20 30 40 50 60 Crank Angle (deg.) 70 80 90 100 Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock

Knock Amplitude (MPa)

8
Smoothed Pressure (MPa)

Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock

0 -100

-75

-50

-25

25

50

75

100

125

150

Crank Angle (deg.)

Figure 9.21: Varying knock intensities extracted from Figure 9.20. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Raw pressure. (Middle): Knock amplitude - high pass filtered pressure. (Lower): Smoothed pressure.

295

1.0 0.8 0.6

MFB

Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock

0.4 0.2 0.0 -20 -10 0 10 20 Crank Angle (deg.) 30 40 50 60

0.10 0.08

MFBR (/deg.)

Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock

0.06 0.04 0.02 0.00 -20 -10 0 10 20 Crank Angle (deg.) 0.12 0.10 Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock 30 40 50 60

MFBR (/deg.)

0.08 0.06 0.04 0.02 0.00 0 0.1

0.2

0.3

0.4

0.5 MFB

0.6

0.7

0.8

0.9

Figure 9.22: Combustion burn effects for varying knock intensity cycles. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): MFB versus CA. (Middle): MFB versus MFBR. (Lower): MFBR versus CA.

296

This process is explained by knock occurring on the inlet valve side and the corresponding increase in heat transfer to the inlet valves. The subsequent reflecting pressure waves momentarily increase local gas and hence wall temperatures at the reflected surfaces, particularly the inlet valves. The higher inlet valve temperature further increases the boosted charge temperature, reducing the VOL and thus the trapped charge energy. Without the effects of VOL runaway knock would occur, which was not observed in experiments. Although there may be some residual gas fraction effects from the lower residual temperature of a knocking cycle, the mass fraction of the residual gas is very small (< 3%) at this high compression ratio. The pressure reduction at exhaust valve open (blow-down) can be identified in the bottom log pressure volume diagram of Figure 9.20. The variability in the ignition time is caused by the variability in the turbulent in-cylinder flow, which also influences knock propensity. The faster initiating cycles tend to be the first knocking cycle in a sequence or that of highest knock intensity. With these concepts in mind, the upper diagram of Figure 9.21 displays four typical knocking cycles of varying knock intensities extracted from Figure 9.20. The knock amplitude recorded from the high pass filtered raw pressure trace is also shown in the middle diagram of Figure 9.21. The peak knock amplitudes closely follow the peak pressure sequence, as can be observed in the bottom diagram of Figure 9.21. Figure 9.22 displays MFB and MFBR data for the varying knock intensity cycles displayed in Figure 9.21. It is noted from Figure 9.22 that the no knock (lower density/lower initiation) cycle has the lowest MFBR from flame initiation. Progressively increasing initial rates lead to higher levels of peak burn rates associated with knocking. The MFB data in the upper diagram of Figure 9.22 is shown for completeness and complies with the Rassweiler and Withrow [40, 182, 183] assumption that all fuel is consumed (MFB = 1). However, this assumption is not entirely valid in high speed engines or in engines where the fuel mixture is richer than stoichiometric. Nevertheless, the method avoids the instabilities that can occur in the end of burn analysis.

297

The turbulent to laminar FSR plots shown in Figure 9.23 indicate peak pre-knock rates that are typical of high charge density, high CR engines associated with turbulent entrainment rather than reaction sheet flames [50]. The knocking process is analyzed but needs cautionary interpretation due to the limitations of the 0.5 CA pressure data sampling interval and the double parabolic smoothing algorithm (Section 9.3.1). These effects can cause aliasing and limit the observed apparent peak flame rates. Despite this, results indicate up to a twofold increase in burning rate for the extreme knock condition.

50 Actual Flame Speed ST,a (m/s) 40 30 20 10 0 -10 0 10 20 Crank Angle (deg.) 30 40 50 Extreme knock (60 bar) Heavy knock (25 bar) Light knock (10 bar) No knock

80

Actual Flame Speed Ratio FSR a

60

Extreme knock (60 bar) Heavy knock (30 bar) Light knock (10 bar) No knock

40

20

0 -10 0 10 20 Crank Angle (deg.) 30 40 50

Figure 9.23: Combustion burn effects for varying intensity knocking cycles. TC - PFI, 7000 rev/min, 220 kPa MAP, 12 BTDC spark timing. (Upper): Actual flame speed versus CA. (Lower): Actual flame speed ratio versus CA.

298

9.7 Discussion
9.7.1 Comparison of Small versus Large Bore Combustion
Table 9.2 displays a comparison of small versus large bore combustion at two extreme operating conditions. These operating conditions include maximum engine speed and MAP together with peak torque. The larger bore engine used in comparisons is a modern 4.0 liter six cylinder Ford Falcon engine found in the Australian marketplace [25, 120]. NA comparisons are made between both bore size engines, with the TC test engine data also displayed in Table 9.2 to analyze the boosted combustion effects. The initial justification for completing the comparison between both NA engines was to compare the turbulent flame speed and FSR across the differing bore sizes as it was presumed both engines would have similar laminar flame speeds. Laminar flame speeds were assumed to be similar as both NA engines operate at similar MAP and have relatively high CRs, although the CR was shown to have only minor effect on laminar flame speeds (Section 9.4). However, the salient differences highlighted in Table 9.2 are the laminar flame speeds across both NA and TC modes for both small and large bore engines. At maximum engine speed, the turbulent flame speed across both engines is similar, ranging from 20-30 m/s. The larger bore engine and hence increased flame travel path gives similar actual flame speeds to the reduced bore engine due to the reduced maximum engine speed. Although the engine speed is reduced due to the longer stroke, both engines have comparable bore/stroke ratios which are almost square. However, the reduction in laminar flame speed for the reduced bore engine due to the higher pressures (due to the increased BMEP) and lower temperatures (associated with the engine speed and spark timing differences delaying the combustion process in CA degrees) causes significant increases in the FSR. Thus, the FSR for the smaller bore engine is doubled at maximum engine speed when compared to the larger bore engine.

299

Comparisons are also made at peak torque, with results highlighting similar differences in laminar flame speeds as previously shown. Furthermore, the turbulent flame speed also varies between both bore size engines. However, this is plainly a function of the mean piston speed, as both engines across NA and TC modes record peak torque at varying engine speeds. The larger bore engine records peak torque at half the mean piston speed of the smaller bore engine, which corresponds to approximately half the turbulent flame speed. Although the flame speeds vary between both bore size engines, comparisons show very similar 0-90% CA burn durations and MFBR at these conditions. This highlights that the reduced bore size engine can match the combustion efficiency of the larger bore engine at higher engine speeds due to the flame speed improvements. However, this is only comparable after CR optimization to compensate for the increased heat, friction and dissociation losses previously described. The increased turbulent flame speed of the reduced bore engine (associated with the higher speed) and the reduced laminar flame speed at 50% MFB has resulted in a 3-4 fold FSR increase when compared to the larger bore engine. The reduction in the laminar flame speed for the smaller bore engine (due to the delayed combustion process) has additional benefits. Although the smaller bore engines performance exceeds the larger bore engine, the lower laminar flame speed is a direct result of lower combustion temperatures. Hence, the potential exists for reduced bore size engines to match or exceed larger bore engines, however the reduced temperatures lead to the possibility of significant reductions in thermal NOX emissions. Although temperature data is unavailable for the larger bore engine, this trend is seen in the test engine across the NA and boosted modes. Furthermore, the engine-out NOX emissions shown in Table 9.2 further illustrate the significant reductions in NOX concentrations for the reduced bore size engine.

300

Table 9.2: Combustion comparison between small bore (test engine) and large bore (Ford Falcon [25, 120]) engines at stoichiometric conditions. Modern Larger Bore Engine [25, 120]
Cylinder Capacity Bore Diameter Stroke Length Bore/Stroke Ratio Mode CR 664 mL 92 mm 99 mm 0.93 mm NA 10 NA 13

Test Engine
217 mL 69 mm 58 mm 1.19 mm TC 10

AT MAXIMUM ENGINE SPEED AND MAP

Engine Speed MAP BMEP Actual Flame Speed ST @ 50% MFB Laminar Flame Speed SL @ 50% MFB Actual FSR @ 50% MFB 5,250 rev/min ~100 kPa 850 kPa 21 m/s 1.1 m/s 20 10,500 rev/min ~100 kPa 1,100 kPa 29 m/s 0.6 m/s 46 10,000 rev/min 170 kPa 1,380 kPa 23 m/s 0.6 m/s 39

AT MAXIMUM TORQUE
Engine Speed MAP Spark Timing BMEP 0-90% MFB Duration MFBR @ 50% MFB Actual Flame Speed ST @ 50% MFB Laminar Flame Speed SL @ 50% MFB Actual FSR @ 50% MFB Burned Gas Temperature @ 50% MFB NOX Concentrations 2,500 rev/min ~100 kPa 25 BTDC 1,050 kPa 46 CA (3.1 ms) 0.05/deg 10 m/s 1.0 m/s 10 4,809 ppm 8,000 rev/min ~100 kPa 30 BTDC 1,260 kPa 56 CA (1.2 ms) 0.05/deg 22 m/s 0.6 m/s 36 2,025 K 3,256 ppm 6,000 rev/min 270 kPa 11 BTDC 2,490 kPa 50 CA (1.4 ms) 0.03/deg 12 m/s 0.3 m/s 35 1,910 K 2,980 ppm

301

9.7.2 Further Suggested Knock Control Strategies

The literature presents numerous knock control strategies as previously outlined in Chapter 7.2. However, new and/or less well documented methods are proposed to enable the extension of the engines operating limits. These include: Spark plug location Piston and piston ring design Valve cooling Flame speed enhancement Repositioning the spark plug towards the inlet side may be an option to offset the flame speed differences associated with wall influences (i.e. the slower propagation across the cooler intake side surfaces). Consequently, the traveling flame front would consume the intake and exhaust side end-gas region simultaneously, thus reducing the likelihood of end-gas knock. Moreover, the simultaneous arrival of the flame in the end-gas around the piston increases efficiency and power, as the reduced burn time reduces the time loss from ideal Otto cycle performance. This may also allow further increases in MAP or CR, with additional benefits. However, the positioning of the spark plug is largely dictated by space limitations due to the compact nature of the pent roof combustion chamber. As a result of the larger intake valves, the spark plug is normally offset towards the exhaust side in conventional cylinder head design. Intake boosting would allow for the use of smaller intake valves and thus the repositioning of the spark plug closer to the bore centre, or even biased towards the intake side. This concept is shown in Figure 9.24, which illustrates the hypothetical flame propagation pattern with the new spark plug location, where the flame front has been distorted by the temperature gradient within the combustion chamber. ignition avoidance and engine reliability. Decreasing the intake valve size ensures adequate cooling around the spark plug, which is vital for pre-

302

INTAKE

EXHAUST

Proposed offset spark plug towards the intake side from its current position in pent roof combustion chambers

Geometry Layout INTAKE EXHAUST

Burning Profile INTAKE EXHAUST

Standard Design
End-gas

NA and Boosted
Larger intake valves. Valve diameter ratio (Ex/In) ~ 0.85. Offset spark plug towards exhaust ~ knock limited

Proposed Design Boosted

Smaller intake valves. Valve diameter ratio (Ex/In) ~ 1. Allows the spark plug to be relocated ~ reduced knock propensity

Figure 9.24: Proposed pent roof design to reduce the intake side knock propensity found from experiments. (Upper): Offset spark plug towards the intake side. (Lower): Effect of ignition point on combustion burning profiles and suggested knock location.

303

Piston cooling development indicated that significantly increasing the underside centre oil cooling had no effect on the propensity for knock (Table 9.1). Instead, experiments have shown that cooling the piston periphery through the piston rings significantly reduced knock (Chapter 5.4). Hence, the effect of increasing the piston ring bore contact area, through an increase in the number of rings, their thickness and location, reduces piston land temperatures by increasing heat transfer to the cooling system [55, 211, 220]. Hence, it is plausible for boosted engines to operate with more than two compression rings, which are normally found in modern engine designs. However, piston ring changes must be balanced with dynamic effects (which can occur at high piston speeds) and increased friction losses from the ring drag [98, 220]. The implementation of sodium filled exhaust valves which operate with lower temperatures [220] is known to prevent surface ignition in high BMEP applications. Although surface ignition is not prevalent in the modern pent roof combustion chamber, sodium cooled valves would promote a more even temperature distribution within the combustion chamber. This has advantages in reducing the flame speed differences between the intake and exhaust side, resulting in a less distorted burn profile and thus reduced knock propensity. Flame speed enhancement through chemical or turbulent effects can also increase burn rates thus reducing the knock likelihood and extending the engines operating limits. Enhancement from increased turbulence can be achieved through improved combustion chamber geometry and increases in engine speed (Figure 9.6). Alternatively, new technologies such as the Bishop rotary valve [49, 235, 236] can increase in-cylinder turbulence as the rotating induction mechanism enhances tumble and swirl patterns [103], resulting in increased flame speeds. Enhancement by chemical effects can also be achieved via supplementing gasoline with alternative fuels (hydrogen or ethanol [37, 79, 230]) or with new ignition technologies such as Hydrogen Assisted Jet Ignition [92]. These methods have been proven to chemically enhance combustion, enabling the operating limits to be extended.

304

9.8 Summary
In this chapter, test engine combustion analysis results for both NA and boosted modes are displayed and discussed. The effects of varying CR, engine speed, MAP and fuel delivery on combustion parameters are investigated. NA combustion results across both carbureted and PFI fuel delivery systems for this particular engine indicate negligible combustion variations between the two modes of fuel delivery at WOT, as verified by the similar performance and efficiency trends in Chapter 7. However, CR increases result in faster initial burn rates, which lead to higher IMEP. Additionally, increases in engine speed do not linearly correlate with combustion duration, with increasing turbulence levels promoting faster flame speeds and hence faster burning rates. Peak flame speeds approaching 40 m/s were recorded in both NA modes at engine speeds reaching 10,500 rev/min. MAP effects on combustion were investigated with TC in-cylinder pressure data. Analysis illustrates that increasing MAP results in faster initial burn rates. However, the faster initial burn is offset by the slow main energy release which increases the 10-90% burn duration. The poorer combustion is associated with the retarded spark timing, which causes the peak temperature and hence flame speed to reduce, although the peak pressure continues to rise due to the higher charge density. Although the retarded spark timing causes the CA burn location to be further delayed for rising MAP, the overall burn duration (0-90% MFB) does not change. The turbulent to laminar FSRs were shown to increase for rising engine speeds and MAP over all modes. This is despite decreasing turbulent flame speeds for rising MAP in the TC mode. However, the rate of laminar flame speed decline exceeded the turbulent flame speed resulting in increasing FSR. FSR values near 50 were recorded over all modes at engine speeds close to 10,000 rev/min. The boosted combustion effects illustrate that the TH is highly dependent on the combustion efficiency, which deteriorates rapidly if spark knock is encountered. Deterioration is associated with spark retard and/or fuel enrichment strategies which although suppress knock to extend the power limit, dramatically reduce combustion and hence TH.

305

In Chapter 7, spark knock was highlighted as the dominant factor in limiting the performance of small engines. In this chapter, further detail is given surrounding the phenomenon. This includes the effects on engine components, which allowed for the clear identification of the knock location in pent roof combustion chambers. The intake side was shown to be the region of highest knock intensities, with the cooler surfaces indicating that true spark knock and not surface ignition occurred in experiments. Knock severity was shown to be highest on the intake side, due to the knock onset locations dependence on the flame front. As the flame front travels over the cooler surfaces of the intake valves, the flame speed significantly reduces resulting in the last burning end-gas occurring on the intake side. Trialed strategies used to control knock are highlighted together with their effects. The effects of knock on combustion for varying intensities are also displayed. Further methods to control spark knock are also suggested targeting new ideas and new technologies. Combustion results for the small cylinder bore engine are also compared to a typical larger bore engine at maximum engine speed and peak torque. Comparisons highlight that the laminar and turbulent flame speeds vary substantially between both bore size engines, resulting in higher turbulent to laminar FSRs for the reduced bore engine. Although the flame speeds vary between both engines, comparisons show very similar 0-90% CA burn durations at these conditions, suggesting that good smaller bore combustion efficiency can be maintained at high engine speeds due to flame speed changes. However, this is only comparable to larger bore engines after small engine CR optimization.

306

CHAPTER 10
P O W E R E D B Y W A T T A R D

Conclusions

10.1 Introduction
This chapter concludes the work presented in this thesis, beginning by highlighting the project achievements on both a research and Formula SAE level. Although, the original intent of the project surrounded achieving success in Formula SAE, focus later evolved into PhD research on small engine performance limits. Following this, conclusions are presented to answer the six research questions outlined in Chapter 1.2. Lastly, recommendations for future work are suggested which include the possible implementation of the small downsized engine (operating in the turbocharged mode) as a power unit for the compact sized, regular passenger vehicle class.

307

10.2 Research Achievements

10.2.1 Mechanical Design and Development
A small downsized 0.43 liter inline twin cylinder engine was designed, constructed and developed at the University of Melbourne. overhead camshafts. The modern engine design featured a four valve pent roof combustion chamber and double The majority of components were manufactured inhouse, either specially cast, fabricated or machined from billets. Additional support was received from sponsors as acknowledged in Appendix A. The final developed turbocharged version is currently on display in the foyer of Mechanical Engineering, Building 170 and depicted in Figure 10.1.

Figure 10.1: The UniMelb WATTARD engine (TC - PFI mode) on display in the foyer of Building 170, Mechanical Engineering at the University of Melbourne.

A novel gasketless interface for an open deck, dry liner block configuration has been designed and developed. The gasketless interface has many benefits, including improving the structural integrity of the weak open deck design. Improving the block structural integrity was critical to the success of the turbocharged engine program due to the extreme operating conditions associated with nearly 300 kPa manifold absolute pressure operation. The

308

gasketless design also provides a cheap and reliable solution to frequent cylinder head removal, as no components require replacement and is therefore ideal for a prototype engine design. Moreover, the application of the gasketless interface provides performance, efficiency and emission benefits due to the elimination of any crevice volume near the interface, which can exist when conventional head gaskets are used. design. The possibility of gasket failure due to abnormal combustion is also eliminated with the gasketless The design has been proven to withstand the harshest operating environments, including high cooling system temperatures and sustained high knock intensities, with peak in-cylinder pressures exceeding 10 MPa. Extensive thermal analysis of the gasketless interface highlighted that removing the conventional head gasket improves the heat flow between the cylinder head and block assembly. This is due to the absence of a gasket, which behaves as an insulator. The improved heat flow reduces the likelihood of hot spots forming around the periphery of the chamber, which reduces the propensity for knock to occur in the end-gas region. gradient across the interface. A novel oil control solution has been developed, permitting the turbocharger to be placed downstream of the throttle. This allows the continued use of cheap and reliable piston ring type seals on the compressor side. The throttled compressor layout has additional benefits in reducing turbo-lag and increasing turbocharger rotational acceleration during engine transients, without the need for costly and sophisticated technologies. The importance of using CAE tools as an aid in reducing engine design and development time has been demonstrated. These tools have enabled improved performance, reliability and understanding at reduced costs and lead times. Excellent agreement between actual and predicted findings highlighted the success of the simulations. Additional benefits include hydrocarbon emission reductions due to the more uniform temperature

309

10.2.2 Experiments
A stationary engine test rig was commissioned and developed to suit the odd firing test engine. The rig included sensors and a data acquisition system Combustion analysis was achieved with inenabling real time sampling.

cylinder pressure measurement, permitting real time and post-test analysis. Experiments were successfully completed in the prototype test engine, which has been operated in a variety of normally aspirated and boosted modes. Normally aspirated modes included carburetion and port injection fuel delivery, while boosted running included supercharged and turbocharged intake systems. The engine control and design parameters that were varied included engine speed, manifold absolute pressure, compression ratio and air-fuel ratio. During experiments, engine speeds exceeding 10,000 rev/min, manifold absolute pressures reaching 270 kPa and compression ratios ranging from 9.6 to 13 were evaluated. Experimental results of performance, efficiency and emissions have been gathered and documented over the aforementioned domains, generating several thousand data points. Operating limits have been defined for a small engine operating in normally aspirated, supercharged and turbocharged modes, thus defining the practical and effective operating envelopes for future engines operating on the downsized increased power concept. In the highly turbocharged mode, the test engine achieved 25 bar brake mean effective pressure at 6,000 rev/min, believed to be the highest recorded value for small engines while operating on pump gasoline. This value is especially significant as it was achieved using no complex knock preventive strategies, relying on traditional tuning strategies and a small reduction in compression ratio. Hence, turbocharging has been successfully applied to an odd firing inline twin configuration. In the normally aspirated modes, experimental results showed minimum brake specific fuel consumption or maximum brake thermal efficiency values in the order of 220 g/kWh or 37% could be achieved. A maximum brake mean The effective pressure of 13 bar was also recorded at 8,000 rev/min.

310

performance and efficiency results demonstrate that smaller bore engines can match or exceed typical larger bore engines found in passenger vehicles. However, this was only possible after compression ratio optimization to compensate for the higher levels of dissociation, friction and heat losses associated with the smaller cylinder capacity. When altering the engine speed, manifold absolute pressure and compression ratio parameters, experimental results showed similar trends in performance, efficiency and emissions when compared to typical larger bore engines found in passenger vehicles. The relationship between engine speed, manifold absolute pressure and compression ratio parameters and their effect on the operating limits have been described for this small engine. exceeding 13. Experimental results showed the potential for normally aspirated engine operation at a compression ratio Furthermore, turbocharged results demonstrated engine operation at a compression ratio of 10 for manifold absolute pressures reaching 270 kPa. The increased performance of this particular small engine in comparison to larger bore engines is attributed to the physical size reduction, particularly the reduced bore size and fast burn combustion chamber. This results in engine speed increases together with end-gas volume reductions around the periphery of the chamber, allowing compression ratio and/or manifold absolute pressure to be increased before exceeding the engine damage limit. This has major benefits in extending the operating limits in downsized applications. The effects of varying engine speed, manifold absolute pressure, compression ratio and fuel delivery on combustion parameters across normally aspirated, supercharged and turbocharged modes were investigated. Peak flame speeds approaching 40 m/s and turbulent to laminar flame speed ratios of nearly 50 were found at engine speeds around 10,500 rev/min.

311

10.2.3 Formula SAE

The initial projects performance, fuel consumption and packaging goals for a Formula SAE rules-compliant new powertrain have been achieved. The test engine was installed into 2003 and 2004 Melbourne University Formula SAE vehicles and entered into competition. finishing all events. The 2004 entry became the first prototype engine to successfully compete in the competitions 25 year history, With engine performance that exceeded larger four cylinder engines and fuel consumption which rivaled the best single cylinders, the engine and vehicle package proved to be very competitive in competition. The first variant tested by the team clocked within one second of the fastest lap time at the 2003 Australasian competition. The turbocharged version later won the fuel economy event in 2004. Near constant power could be achieved over half the speed range in the highly turbocharged mode due to the Formula SAE regulated intake restriction, which limits airflow and thus performance when operating at the choked condition. The extension of the maximum power speed range enabled the vehicle to complete both 21 km endurance track events with the avoidance of gear shifting, except for initial vehicle takeoff. This improved on track power, making the vehicle easier to handle for novice student drivers.

10.3 Conclusions to the Research

The main goal of this extended research was to explore small engine (~0.5 liter) performance limits and define operating boundaries, especially as engine downsizing grows in popularity due to increasing world concerns. Defining these performance boundaries requires answers to questions about the effects of the engine component design, turbocharging, gas exchange and combustion to determine if they limit small engine performance. Throughout this thesis, performance, efficiency and emissions associated with the reduced bore size test engine are presented and found to be comparable to or surpass that of typical larger bore engines found in passenger vehicles. in Chapter 1.2 are now presented. To justify these claims qualitatively for future engine research, answers to the research questions posed

312

1. A four valve, inline two cylinder engine was designed and constructed to overcome the mechanical and thermal component limitations found in current production engines, which inherently restrict engine performance. The new engine was designed to reliably withstand the high mechanical and thermal loads associated with nearly two bar boost pressure and engine speeds exceeding 10,000 rev/min. Extensive engine development and experiments over hundreds of hours of engine operation later confirmed the design, concluding that the mechanical and thermal aspects of the component design can be overcome so as not to compromise engine performance limits. 2. The engine configuration, specifically the odd fire two cylinder design caused vibration and torque pulsation problems not witnessed in engine configurations with a greater number of cylinders. Engine reliability and longevity concerns associated with these effects were further magnified by the high mechanical and thermal loads associated with the combined high boost pressures and engine speeds. However, component development enabled these problems to be overcome. Hence, it is foreseeable that a two or other configuration with odd firing (e.g. V4) will not restrict small engine performance. 3. The intake and exhaust gas flow abnormalities associated with the reduction in the number of cylinders and unequal firing spacing were shown to be overcome for the inline two cylinder test engine. Experiments and simulation highlighted that the exhaust manifold volume was the most important parameter to optimize for highly turbocharged operation, with the intake and valve events having small effects in comparison. In addition to the volume, the exhaust geometry effects were investigated with both constant pressure and pulse turbocharging systems. Results from experiments and simulation showed significant brake power improvements when employing constant pressure type turbocharging to an odd firing two cylinder configuration. This demonstrates that with exhaust geometry optimization, high pressure ratio turbocharging can be successfully applied to any odd firing configuration ( two cylinders). Consequently, the gas flow abnormalities associated with the unequal firing spacing can be successfully overcome so as not to restrict engine performance.

313

4. Experimental and simulation results have clearly shown that the engine peak performance is not limited by the turbocharger. Experiments also demonstrated that the compressor is not the peak performance limiting factor, with other obstacles encountered prior to the compressor reaching its flow and hence performance limit. However, the low speed engine performance was found to be limited by the rate at which the turbocharger could supply the required airflow to increase power, as well documented in larger bore engines. Additionally, the turbocharger mechanical design proved adequate after development to suit the compressors throttled application. The novel compressor throttled solution reduced turbo lag and increased turbocharger rotational acceleration during engine transients, which aided in achieving adequate vehicle transient response under high boost pressures.
5. Combustion limitations at high engine speed, manifold absolute pressure and

compression ratio were investigated across normally aspirated, supercharged and turbocharged modes to establish if combustion restricted small engine performance. Experiments highlighted that abnormal combustion, specifically spark knock in the end-gas region, was the dominant factor in limiting the performance of this small engine. Highest knock intensities were observed when knock was deduced to occur on the intake side of the pent roof combustion chamber. Hence, the extent to which the larger engines can
be downsized while still maintaining equal performance is combustion limited.

6. In order to determine how emissions, especially CO2, from the test engine compare to modern larger passenger vehicle engines with similar power output, a case study was performed which analyzed the feasibility of replacing a larger 1.25 liter normally aspirated engine found in the 2007 Ford Fiesta. Results showed that the performance of the larger engine could be readily matched with the smaller turbocharged unit, with a 66% reduction in engine capacity while using no complex knock preventative strategies. This indicates the opportunity for the swept capacity of all normally aspirated engines fitted to automobiles to be halved with no loss in performance. Although peak power could be matched, the smaller downsized engine had increased fuel

314

consumption and hence CO2 emissions at high power outputs, which are seldom used in normal driving patterns. However, when the downsized test engine was simulated to operate in the Ford Fiesta chassis and analyzed over the New European Drive Cycle, results showed a 22% best reduction in fuel consumption and CO2 emissions, including the reduction to 62% at idle conditions. These are a consequence of operating the test engine closer to peak efficiency under light load conditions, even though both engines have similar peak efficiencies of ~30%. Additional CO2 benefits are due to engine and chassis mass reductions attributed to the smaller powertrain. g/km. The reduction in CO2 would shift the vehicle under the 2012 Euro target of 120

10.4 Recommendations for Future Work

There is potential to complete 3-dimensional modeling with a detailed chemical kinetic scheme to better understand the knock phenomenon, as spark knock has been shown to be the dominant performance limiting factor in small engines. However, it is imperative that experimental results are used to validate any simulation findings. Modeling may also allow the operating limits to be extended by reducing the propensity for intake side knock through combustion chamber and cylinder head geometry optimization (including the spark plug location together with the port and valve geometry). It is envisaged that modeling may be a useful tool in overcoming the flame speed discrepancies between the inlet and exhaust side as the flame travels over the cooler intake surfaces. The piston design, in particular the piston ring number, thickness and location requires further investigation under highly turbocharged operating conditions, as experiments have shown this area to be vital in controlling spark knock. Optimization may be completed through further experimentation in conjunction with CAE simulation to analyze the temperature profile and hence heat flux around the periphery of the combustion chamber. efficiency and emission benefits. Extending the operating limits through spark knock reductions would enable performance,

315

 High octane gasoline-toluene fuel blends were used briefly during engine development in order to reduce knock intensities and thus improve engine reliability at compression ratios beyond the knock limit with pump fuel. Results showed that the higher octane fuels successfully extended the engines operating limits enabling enhanced power, efficiency and emission benefits. Further high octane experiments could be completed to determine the operating limit boundaries and the associated effects on performance, efficiency and emissions across the test domain for varying fuel qualities. High octane fuels such as ethanol blends and higher grades of gasoline could be used in experiments, especially as these fuels become more widespread. A well-to-wheel analysis could be completed to determine the cost and CO2 effects of increasing fuel octane at the refinery, allowing the engine operating limits to be extended. This would give performance and efficiency benefits, which would then be offset against higher refinery costs and greenhouse impacts to determine the viability of refining higher grades of gasoline. This would allow the optimum fuel octane number of the highly turbocharged engine to be established. Extensive durability experiments to document the long term effects associated with high output turbocharged operation could be completed. This would help verify if the performance levels achieved in experiments could be sustained over OEM durability test cycles, while still maintaining engine robustness and reliability. In addition, results from durability tests would give insight into whether power level reduction or design and control strategy improvements are required. Direct injection, exhaust gas recirculation, intercooling, alternative fuels and ignition system technologies are documented to have positive effects on spark ignition engine performance, efficiency and emissions by extending the engines operating limits. Experiments could be completed to determine the viability and quantify the effects of these options in a highly boosted small engine. Turbocharged low speed performance in a small engine requires further development as experiments and comparison analysis highlighted weaknesses

316

prevalent in larger bore engines.

This could potentially involve both

turbocharger (variable turbine geometry) and engine (through scavenging with the aid of variable valve timing) further work. Further investigation could be completed to determine if the findings presented in this work for two cylinder configurations carry over to single cylinder designs. This has significant relevance to emerging markets such as China and India, with trends suggesting booming potential for single or twin cylinder configurations due to the cost benefits associated with producing engines with fewer cylinders. Furthermore, the constant power benefits of the flow restricted engine lend itself to a very simple low cost three speed transmission, with the test engine also not requiring an intercooler. The extent to which the bore size and cylinder capacity could be further reduced while still marinating performance and efficiency levels comparable to larger bore engines found in passenger vehicles also requires analysis.

10.4.1 Implementation into a Passenger Vehicle

Throughout the experimental phase of the research, many parameters were explored. However, the engine design, development and testing focused on supplying an engine that would meet Formula SAE regulations in an effort to achieve success in competition. Hence, the potential exists to develop and test the highly turbocharged powertrain to meet the conditions required for compact sized regular passenger vehicles. As an intake flow restriction would not be required if the engine was implemented into passenger vehicles, experiments could be completed to determine the engines operating limits without the flow restriction. Furthermore, CAE modeling validated by experimental testing is required to determine new engine parameters and variables (intake, exhaust, valve timing) to suit the unrestricted flow condition. This is particularly relevant for the turbocharged mode, which would most likely be employed in a passenger vehicle. In addition, the turbocharged peak power potential at high speeds without a flow restriction needs to be investigated to determine further swept

317

capacity reductions when compared to the test engine for larger engine replacement. Further experiments with an emission control system (three-way catalyst) could be completed to ensure the test engine can meet current emission standards over the drive cycle after recalibration to stoichiometric conditions. In addition, any effects due to the emission control system (higher exhaust backpressure) require investigation. Although vehicle transient response while the test engine was fitted to a 220 kg Formula SAE vehicle from driver feedback appeared uncompromised, further transient work is required with the engine fitted to a 1,000 kg passenger vehicle. Furthermore, the potential exists to further optimize the turbocharger compressor and turbine geometry to match the intended application. Reliability, cost and packaging analysis could be completed to determine the feasibility of implementing the test engine into the small passenger vehicle class. This includes material, manufacturing and maintenance analysis to determine the well-to-wheel impact of replacing larger engines with this type of powertrain. Although the performance, efficiency and emissions of this particular two cylinder configuration may give promise as a replacement for larger engines found in passenger vehicles, it is anticipated that the poor noise, vibration and harshness attributed to the odd firing interval may limit the application. Hence further work is required to reduce these effects for the odd fire twin configuration to levels found in engines with a greater number of cylinders. Possibilities include the addition of balance shafts in the inline twin, or flat configurations which would permit an equal firing spacing. Equal firing for a two cylinder configuration would give an identical firing frequency to that of a four cylinder configuration operating at half the engine speed, thus greatly reducing the noise, vibration and harshness. The implementation of an equal firing twin cylinder configuration would require further optimization, but results from the odd fire engine would provide a solid grounding for future development.

318

Awards and Publications

P O W E R E D B Y W A T T A R D

List of awards received: 2007 Faculty of Engineering MIRAMS Award for Outstanding Research 2007 Australian Young Engineer Nominee (SAE Australasia) 2006 Faculty of Engineering MIRAMS Award for Outstanding Research 2006 Max Bentele Award for Engine Technology Innovation (SAE International)

List of first authored publications:

2007-24-0083

The Feasibility of Downsizing a 1.25 Liter Normally Aspirated Engine to a 0.43 Liter Highly Turbocharged Engine (2007 ICE -

Naples, Italy)
2007-01-3623

Compression

Ratio

Effects

on

Performance,

Efficiency,

Emissions and Combustion in a Carburetted and PFI Small Engine (2007 APAC - Hollywood, CA, USA)
2007-01-1562

Highly Turbocharging a Flow Restricted Two Cylinder Small Engine - Turbocharger Development (2007 SAE World Congress

- Detroit, MI, USA)

2006-01-3637

Highly Turbocharging a Flow Restricted, Odd Fire, Two Cylinder Small Engine - Design, Lubrication, Tuning and Control (2006

SAE Motorsports - Dearborn, MI, USA)

2006-32-0036

Design and Development of a Gasketless Cylinder Head / Block Interface for an Open Deck, Multi Cylinder, Highly Turbocharged Small Engine (2006 SAE Small Engine - San

Antonio, TX, USA)

2006-32-0072

Comparing the Performance and Limitations of a Downsized Formula SAE Engine in Normally Aspirated, Supercharged and Turbocharged Modes (2006 Small Engine - San Antonio, TX,

USA)
2006-01-0745

Development of a 430 cm3 Constant Power Engine for Formula SAE Competition (2006 SAE World Congress - Detroit, MI, USA)

319

320

References
P O W E R E D B Y W A T T A R D

[1]

AAMA,

World

Motor

Vehicle

Data

1993.

American

Automobile

Manufacturers Association, Motor Vehicle Facts and Figures, 1996. [2] Abraham, J., Williams, F.A. and Bracco, F.V., A Discussion of Turbulent

Flame Structure in Premixed Charges. SAE Technical Paper Series

(850345), 1985. [3] Afzal, A. and Fatemi, A., A Comparative Study of Fatigue Behavior and Life

Predictions of Forged Steel and Pm Connecting Rods. SAE Technical Paper

Series (2004-01-1529), 2004. [4] Alger, T., Mangold, B., Mehta, D. and Roberts, C., The Effect of Sparkplug

Design on Initial Flame Kernel Development and Sparkplug Performance.

SAE Technical Paper Series (2006-01-0224), 2006. [5] Alkidas, A., Effects of Operational Parameters on Structural Temperatures

and Coolant Heat Rejection of a S.I Engine. SAE Technical Paper Series
(931124), 1993. [6] [7] All-Plastics, http://www.allplastics.com.au. Accessed: 20 May, 2004. Amann, C.A., Cylinder Pressure Measurement and its use in Engine

Research. SAE Technical Paper Series (85206), 1985.

[8] Arai, T., Yajima, J., Murata, K. and Hibino, M., Second Generation of High-

Response V6 Engine Series (3.0 and 3.5 Liters). SAE Technical Paper
Series (2000-01-0668), 2000. [9] Arnold, S.D., Balis, C., Jeckel, D., Larcher, S., Uhl, P. and Shahed, S.M.,

Advances in Turbocharging Technology and its Impact on Meeting Proposed California GHG Emission Regulations. SAE Technical Paper Series
(2005-01-1852), 2005.

321

[10]

Attard, W.P. and Khan, M., Exploration of the Effects of Nickel-Teflon

Coating on a Spark Ignition Internal Combustion Engine. Thesis (B.E.),

Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003. [11] Attard, W.P., Konidaris, S., Hamori, F., Toulson, E. and Watson, H.C.,

Compression Ratio Effects on Performance, Efficiency, Emissions and Combustion in a Carbureted and PFI Small Engine. SAE Technical Paper
Series (2007-01-3637), 2007. [12] Attard, W.P., Konidaris, S., Toulson, E. and Watson, H.C., The Feasibility of

Downsizing a 1.25 Liter Normally Aspirated Engine to a 0.43 Liter Highly Turbocharged Engine. SAE Technical Paper Series (2007-24-0083), 2007.
[13] Attard, W.P. and Watson, H.C., Development of a 430 cm3 Constant Power

Engine for FSAE Competition. SAE Technical Paper Series (2006-01-0745),

2006. [14] Attard, W.P., Watson, H.C. and Konidaris, S., Highly Turbocharging a

Restricted, Odd Fire, Two Cylinder Small Engine - Design, Lubrication, Tuning and Control. SAE Technical Paper Series (2006-01-3637), 2006.
[15] Attard, W.P., Watson, H.C. and Konidaris, S., Highly Turbocharging a Flow

Restricted Two Cylinder Small Engine - Turbocharger Development. SAE

Technical Paper Series (2007-01-1562), 2007. [16] Attard, W.P., Watson, H.C., Konidaris, S. and Khan, M.A., Comparing the

Performance and Limitations of a Downsized Formula SAE Engine in Normally Aspirated, Supercharged and Turbocharged Modes. SAE
Technical Paper Series (2006-32-0072), 2006. [17] Attard, W.P., Watson, H.C. and Stryker, P., Design and Development of a

Gasketless Cylinder Head/Block Interface for an Open Deck, Multi Cylinder, Highly Turbocharged Small Engine. SAE Technical Paper Series (2006-320036), 2006.

322

[18]

Australian-Standard, Measurement of Fluid Flow in Closed Conduits: AS

2360.1.1-1993. 1993.
[19] Australian-Standard, Internal Combustion Engines - Performance: AS

4594.1-1999. 1999.
[20] [21] Auto-Diagnostics, ADS 9000 Super Four Gas Analyzer. Operator's Manual. Automotive-Components-Limited, http://www.acl.com.au. Accessed: 10 January, 2003. [22] Automotive-Racing-Products, February, 2003. [23] Babu, M.K.G. and Deslandes, J.V., A Preliminary Analysis of Matching a

http://www.arp-bolts.com.

Accessed:

22

Turbocharger to the 1.3 litre Opel Passenger Car Engine including Cam Timing. Department of Mechanical and Manufacturing Engineering,
University of Melbourne, Report T/54, 1983. [24] Badami, M., Marzano, M.R. and Nuccio, P., Influence of Late Intake-Valve

Opening on the S.I. Engine-Performance in Idle Condition. SAE Technical

Paper Series (960586), 1996. [25] Baker, P.A., A Comparison of Multipoint Liquid and Gaseous Phase

Injection. Current Thesis (Ph.D.), Department of Mechanical and

Manufacturing Engineering, University of Melbourne, 2007. [26] Baniasad, S., Khalil, E. and Shen, F., Exhaust Valve Thermal Management

and Robust Design Using Combustion and 3D Conjugate Heat Transfer Simulation with 6-Sigma Methodology. SAE Technical Paper Series (200601-0889), 2006. [27] Baratta, M., Catania, A.E., Spessa, E. and Vassallo, A., Development and

Assessment of a Multizone Combustion Simulation Code for SI Engines Based on a Novel Fractal Model. SAE Technical Paper Series (2006-010048), 2006.

323

[28]

Barry, A., Errors in Practical Measurement in Science, Engineering and

Technology. John Wiley & Sons, ISBN 0471031569, 1978.

[29] Beer, T., Grant, T., Watson, H.C. and Olaru, D., Life-Cycle Emissions

Analysis of Fuels for Light Vehicles. Report HA93A-C837/1/F5.2E Australian

Greenhouse Office, 2004. [30] Boisvert, J. and Hill, P.G., Test of Thickness Flame Combustion Estimates

in a Single-Cylinder Engine. ASME, Vol.109, pp. 410-418, October 1987.

[31] Boretti, A.A., Jin, S.H., Zakis, G., Brear, M.J., Attard, W., Watson, H., Carlisle, H. and Bryce, W., Experimental and Theoretical Study of a Direct

Injection Spark ignition Engine. SAE Technical Paper Series (2007-011419), 2007. [32] Borman, G.L. and Ragland, K.W., Combustion Engineering. McGraw-Hill, New York, ISBN 0-07-1195978-9, 1998. [33] Bosch, Automotive Handbook Fifth Edition, Robert-Bosch-GmbH , ISBN 08376-0614-4, 2000. [34] Bradley, D., Hicks, R.A., Lawes, M., Sheppard, C.G.W. and Woolley, R.,

The Measurement of Laminar Burning Velocities and Markstein Numbers for Isooctane-Air and Isooctane-n-Heptane-Air Mixtures at elevated Temperatures and Pressures in an Explosion Bomb. Combustion and
Flame, Vol. 115, pp. 126-144, 1998. [35] Brewster, S., Initial Development of a Turbocharged Direct Injection E100

Combustion System. SAE Technical Paper Series (2007-01-3625) 2007.

[36] [37] British-Standard, Flow Engineering: BS-1042. 1984. Bromberg, L., Cohn, D.R. and Heywood, J.B., Calculations of Knock

Suppression in Highly Turbocharged Gasoline/Ethanol Engines using Direct Ethanol Injection.

Report, Massachusetts Institute of Technology Cambridge MA 02139, February 23, 2006.

324

[38]

Brown,

B.R.,

Combustion

Data

Acquisition

and

Analysis.

Thesis

(M.Eng.Sc.), Loughborough University, 2000. [39] Brunt, M., Pond, C. and Biundo, J., Gasoline Engine Knock Analysis Using

Cylinder Pressure Data. SAE Technical Paper Series (980896), 1998.

[40] Brunt, M.F.J. and Erntage, A.L., Evaluation of Burn Rate Routines and

Analysis Errors. SAE Technical Paper Series (970037), 1997.

[41] [42] Buzzi, L., Crankshaft Balancing. Booklet No. 19, Dott. Ing., 1982. Cairns, A., Blaxill, H. and Irlam, A., Exhaust Gas Recirculation for Improved

Part and Full Load Fuel Economy in a Turbocharged Gasoline Engine. SAE
Technical Paper Series (2006-01-0047), 2006. [43] CARB, Technology and Cost Assessment for Proposed Regulations to

Reduce Vehicle Climate Change Emissions. Pursuant to Assembly Bill 1493,

State of California, Air Resources Board, April 1, 2004. [44] Caton, J. and Heywood, J., An Experimental and Analytical Study of Heat

Transfer in an Engine Exhaust Port. Int. J. Heat Mass Transfer, Vol 24, No.
4, pp. 581-595, 1981. [45] Cauchi, J., Farrugia, M. and Balzan, N., Engine Simulation of a Restricted

Formula SAE Engine, Focusing on Restrictor Modelling. SAE Technical

Paper Series (2006-01-3651), 2006. [46] Challen, J., International Engine of the Year Awards 2004. Cover Story, Engine Technology International, September, 2004. [47] Chang, J., Guralp, O., Filipi, Z., Assanis, D., Kuo, T., Najt, P. and Rask, R.,

New Heat Transfer Correlation for an HCCI Engine Derived from Measurements of Instantaneous Surface Heat Flux. SAE Technical Paper
Series (2004-01-2996), 2004. [48] Chinouilh, G., Santacreu, P.O. and Herbelin, J.M., Thermal Fatigue Design

of Stainless Steel Exhaust Manifolds. SAE Technical Paper Series (2007-010564), 2007.

325

[49]

Chow, P.H.P., Combustion in a Rotary Valve Spark Ignition Engine. Thesis (Ph.D.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2005.

[50]

Chow, P.H.P., Watson, H.C. and Wallis, T., Combustion in a High Speed

Rotary Valve Spark Ignition Engine. Proc. IMechE, Part D: Journal of

Automobile Engineering, 2007. [51] Chun, K.M. and Kim, K.W., Measurements and Analysis of Knock on a SI

Engine using the Cylinder Pressure and Block Vibration Signals. SAE
Technical Paper Series (940146), 1994. [52] Clark, J.S. and Owen, B.D., Turbocharging the 1.3 litre BL Metro car. I.Mech.E, Turbocharging and Turbochargers Conference Publications 198283, pp. 49-82, 1982. [53] Clarke, D.P., Such, C.H., Overington, M.T. and Das, P.K., A Lean Burn

Turbocharged, Natural Gas Engine for the US Medium Duty Automotive Market. SAE Technical Paper Series (921552), 1992.
[54] Claywell, M. and Horkheimer, D., Improvement of Intake Restrictor

Performance for a Formula SAE Race Car through 1D and Coupled 1D/3D Analysis Methods. SAE Technical Paper Series (2006-01-3654), 2006.
[55] Clough, M.J., Precision Cooling of a Four Valve per Cylinder Engine. SAE Technical Paper Series (931123), 1993. [56] CONCAWE, Motor vehicle emission regulations and fuel specifications - Part

1. Summary and annual 1997/1998 update. Brussels, (Report no.9/98),

1998. [57] Conte, E. and Boulouchos, K., Experimental Investigation into the Effect of

Reformer Gas Addition on Flame Speed and Flame Front Propagation in Premixed, Homogeneous Charge Gasoline Engines. Combustion and Flame,
Vol. 146, pp. 329-347, 2006.

326

[58]

Daugherty, R.L. and Franzini, J.B., Fluid Mechanics with Engineering

Applications. Twelfth Edition, McGraw-Hill, ISBN 0-07-015427-9, 1982.

[59] [60] Davies-Craig, http://www.daviescraig.com.au. January, 2003. Davis, R.S. and Patterson, G.S., Cylinder Pressure Data Quality Checks and

Procedures to Maximize Data Accuracy. SAE Technical Paper Series (200601-1346), 2006. [61] DieselNet, http://www.dieselnet.com/standards/. Accessed: 10 August, 2007. [62] Dober, G.G., Geometrical Control of Hydrocarbon Emissions. Thesis (Ph.D.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2002. [63] Dogan, A., Gledhill, M. and Tsioutras, T., Honda CBR-600 in Formula SAE-

A.

Thesis

(B.E.),

Department

of

Mechanical

and

Manufacturing

Engineering, University of Melbourne, 2005. [64] Duchaussoy, Y., Lefebvre, A. and Bonetto, R., Dilution Interest on

Turbocharged SI Engine Combustion. SAE Technical Paper Series (200301-0629), 2003. [65] Ecker, H.J., Gill, G. and Schwaderlapp, M., Downsizing of Diesel Engines:

3-Cylinder/4-Cylinder. SAE Technical Paper Series (2000-01-0990), 2000.

[66] Edgar, J., 21st Century Performance. Clockwork Media, ISBN 0-947216-901, 2000. [67] Edison, M.H., The Influence of Compression Ratio and Dissociation on

Ideal Otto Cycle Engine Thermal Efficiency. SAE Prog. in Technology, Vol.
7, pp. 49-64, 1964. [68] Edison, M.H. and Taylor, C.F., The Limits of Engine Performance-

Comparison of Actual and Theoretical Cycles. SAE Prog. in Technology, Vol.

7, pp. 65-81, 1964.

327

[69]

Erickson, W.D. and Prabhu, R.K., Rapid Computation of Chemical

Equilibrium Composition: An application to Hydrocarbon Combustion.

AIChE Journal, Vol. 32, No. 7, July 1986. [70] Eriksson, L., Documentation for the Chemical Equilibrium program package

CHEPP. Department of Electrical Engineering, Linkping University, October

3, 2000. [71] [72] [73] [74] ExxonMobil, The Outlook for Energy: A View to 2030. Report, 2007. Ford-Australia, http://www.ford.com.au. Accessed: 27 February, 2007. Ford-UK, http://www.ford.co.uk/ie/fiesta/. Accessed: 15 January, 2007. Ford-US, http://media.ford.com/newsroom/release_display.cfm. Accessed: January 10, 2007. [75] [76] Formula-SAE, www.fsae.com. Accessed: 10 February, 2006. Fox, J.W., Cheng, W.K. and Heywood, J.B., A Model for Predicting Residual

Gas Fraction in Spark-Ignition Engines. SAE Technical Paper Series

(931025), 1993. [77] Freame, B.,

Australian

Precision

Engine

Parts

(APEP).

Personal

conversation, January 5, 2002. [78] Gauci, A. and Yousuff, F., Formula SAE - Restrictor (Experimental). Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2006. [79] Gerty, M. and Heywood, J.B., An Investigation of Gasoline Engine Knock

Limited Performance and the Effects of Hydrogen Enhancement. SAE

Technical Paper Series (2006-01-0228), 2006. [80] Gillespie, L., Lawes, M., Sheppard, C.G.W. and Woolley, R., Aspects of

Laminar and Turbulent Burning Velocity relevant to SI Engines. SAE

Technical Paper Series (2000-01-0192), 2000.

328

[81]

Goddard,

G.,

Green

Motorsports.

SAE

International

Motorsports

Engineering Conference & Exposition, Dearborn, Michigan, 2006. [82] Gorr, E., Kerr, C., Stevens, K. and Archer, J., Techniques for Manufacturing

and Coating Liners of Small Engines. SAE Technical Paper Series (2000-010904), 2000. [83] Grandin, B. and Angstrom, H.E., Replacing Fuel Enrichment in a

Turbocharged SI Engine: Lean Burn or Cooled EGR. SAE Technical Paper

Series (1999-01-3505), 1999. [84] Green, G.W., Design of Crankshafts By the Finite Element Method. SAE Technical Paper Series (870579), 1987. [85] Greene, D.L. and Hopson, J.L., Running Out of and Into Oil: Analyzing

Global Depletion and Transition Through 2050. ORNL/TM-2003/259, Oak

Ridge National Laboratory, Oak Ridge, Tennessee, October 2003. [86] Groff, E.G. and Matekunas, F.A., The Nature of Turbulent Flame

Propagation. SAE Technical Paper Series (800133), 1980.

[87] Guezennec, Y.G., and W. Hamama, Two-Zone Heat Release Analysis of

Combustion Data and Calibration of Heat Transfer Correlation in an I.C Engine. SAE Technical Paper Series (1999-01-0218), 1999.
[88] Guzzella, L., Wenger, U. and Martin, R., IC-Engine Downsizing and

Pressure-Wave Supercharging for Fuel Economy. SAE Technical Paper

Series (2000-01-1019), 2000. [89] Halpin, D., Repco Engine Developments. Personal conversation, January 5, 2002. [90] Hamilton, B.,

What

Parameters

Determine

Octane

Requirement?,

http://www.faqs.org/faqs/autos/gasoline-faq/part3/. Created: 15 January,

2004, Accessed: 8 December, 2004. [91] Hamori, F., Experimental Combustion Burn Rate Analysis (E-CoBRA). Software, 2005.

329

[92]

Hamori, F., Exploring the limits of Hydrogen Assisted Jet Ignition. Thesis (Ph.D.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2006.

[93]

Hamori, F. and Zakis, G., Formula SAE Engine and Gearbox Design. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2000.

[94]

Harralson, J., Design of Racing and High Performance Engines. SAE International, PT-53, ISBN 1-56091-601-X, 1995.

[95]

Heisler, H., Advanced Engine Technology. SAE International, ISBN 156091-734-2, 1995.

[96]

Heisler, H., Advanced Vehicle Technology. Butterworth-Heinemann, ISBN 0-7506-5131-8, 2002.

[97]

Heywood, J.B., Automotive Engines and Fuels: A Review of Future Options. Progress in Energy and Combustion Science, Vol. 7, Iss. 3., pp. 155-184, 1981.

[98]

Heywood, J.B., Internal Combustion Engine Fundamentals. McGraw-Hill, ISBN 0-07-100499-8, 1988.

[99]

Heywood, J.B., Combustion and its Modeling in Spark-Ignition Engines. International Symposium, COMODIA 94, 1994.

[100] Holden-Australia, http://www.holden.com.au. Accessed: 20 March, 2007. [101] Horiba, MEXA-700 Lambda Analyzer. Operator's Manual, 2004. [102] Horrocks, G., Intake Manifold Pressure Wave Tuning Thesis (B.E.), Department of Mechanical Engineering, University of Technology - Sydney, 1998. [103] Horrocks, G., A Numerical Study of a Rotary Valve Internal Combustion

Engine. Thesis (Ph.D.), Department of Mechanical Engineering, University

of Technology - Sydney, 2002.

330

[104] Houghton, J.T., Climate Change 1995: The Science of Climate Change. Published for the Intergovernmental Panel on Climate Change in collaboration with the World Meteorological Organization and the United Nations Environment Program, Cambridge University Press, Cambridge, U.K., 1996. [105] Hubbert, M.K., Techniques of Prediction as Applied to Production of Oil and

Gas. US Department of Commerce, NBS Special Publication 631, May,

1982. [106] IEA, http://www.iea.org/Textbase/subjectqueries/index.asp. International Energy Agency, Energy Information Centre, Created: 2000, Accessed: 10 June, 2007. [107] Incropera, F.P., Fundamentals of Heat Transfer. John Wiley and Sons, ISBN 0471612464, 1990. [108] Ishigaki, I., Kitagawa, J. and Tanaka, A., New Evaluation Method of Metal

Head Gasket. SAE Technical Paper Series (930122), 1993.

[109] Jackson, M.W., Analysis for Exhaust Gas Hydrocarbons - Nondispersive

Infrared Versus Flame Ionisation. Air Pollution Control Ass., Vol 16, pp.
697-702, 1966. [110] James, E.H., Laminar Burning Velocities of Isooctane-Air Mixtures - A

Literature Review. SAE Technical Paper Series (870170), 1987.

[111] Jang, S.H., Lee, Y.G., Oh, T.Y. and Park, K.S., An Experimental Study on

Knock Sensing for a Spark Ignition Engine. SAE Technical Paper Series
(931902), 1993. [112] Jeng, Y.R., Topographical Changes of Engine Bores During Engine Running

In. SAE Technical Paper Series (2000-01-1791), 2000.

[113] Karagounis, T. and Straus, A., Development of a 90 bhp Turbocharged

Engine for use in Formula SAE-A. Thesis (B.E.), Department of Mechanical

and Manufacturing Engineering, University of Melbourne, 2002.

331

[114] Kastner, L.J., Williams, T.J. and White, J.B., Poppet Valve Characteristics

and their Influence on the Induction Process. I.Mech.E. 178 pt.1, No.36,
1963. [115] Kato, S., Nakamura, N., Kato, K. and Ohnaka, H., High Efficiency

Supercharger Increases Engine Output, Reduces Fuel Consumption through Computer Control. SAE Technical Paper Series (861392), 1986.
[116] Kato, T., Nakashima, T., Akiyama, K. and Shimizu, R., Development of

Combustion Behaviour Analysis Techniques. SAE Technical Paper Series

(2007-01-0643), 2007. [117] Kawasaki, ER500 Operator Manual. 1993. [118] Keppeler, S., Schulte, H. and Ecker, H.J., The Technical Ramifications of

Downsizing HSDI Diesel Technology to the 300cc Displacement Class. SAE

Technical Paper Series (981916), 1998. [119] Keribar, R. and Morel, T., Thermal Shock Calculations in I.C Engines. SAE Technical Paper Series (870162), 1987. [120] Khan, M., LPG Throttle Body Injection for Enhanced Air-Fuel Mixing,

Reduced Emissions Exploring and Improved Efficiency. Current Thesis

(Ph.D.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2007. [121] Khan, M., Watson, H.C., Baker, P.A., Liew, G. and Johnston, D., SI Engine

Lean-Limit Extension Through LPG Throttle-Body Injection for Low CO2 and NOx. SAE Technical Paper Series (2006-01-0495), 2006.
[122] Kistler-Instrument-Corporation, Operating Instructions, Universal Pressure

Transducers, Types: 211B(X), 601B(X), 603B(X).

[123] Kleeberg, H., Tomazic, D., Lang, O. and Habermann, K., Future Potential

and Development Methods for High Output Turbocharged Direct Injected Gasoline Engines. SAE Technical Paper Series (2006-01-0046), 2006.

332

[124] Konidaris, S. and Stickney, S., Optimizing a Modified Kawasaki GPZ500S. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2004. [125] Konig, G. and Sheppard, C.G.W., End Gas Autoignition and Knock in a

Spark Ignition Engine. SAE Technical Paper Series (902135), 1990.

[126] Kreyszig, E., Advanced Engineering Mathematics. Eighth Edition, McGrawHill, ISBN 0-471-15496-2, 1999. [127] Kuck, H.A., Fleischer, V. and Schnorbus, W., New 1.3L High Performance

SI-Engine Supercharged by the VW G-Lader. SAE Technical Paper Series

(860102), 1986. [128] Kummer, J.T., Catalysts for Automobile Emission Control. Prog. Energy Combustion Sci., Vol. 6, pp. 177-199, 1981. [129] Kuratle, R., Combustion Engine Measurement Technology. Report, Kistler Instruments AG, 1990. [130] Lake, T., Stokes, J., Murphy, R., Osborne, R. and Schamel, A.,

Turbocharging Concepts for Downsized DI Gasoline Engines. SAE Technical

Paper Series (2004-01-0036), 2004. [131] Lambe, S.M., Hydrogen dual-fuel Combustion. Thesis (Ph.D.), Department of Mechanical Engineering, University of Melbourne, 1991. [132] Lancaster, D.R., Effects of Engine Variables on Turbulence in a Spark-

Ignition Engine. SAE Technical Paper Series (760159), 1976.

[133] Lancaster, D.R., Krieger, R.B. and Lienesch, J.H., Measurement and

Analysis of Engine Pressure Data. Engine Research Dept., General Motors

Research Laboratories, SAE Technical Paper Series (750026), 1975. [134] Lancaster, D.R., Krieger, R.B., Sorenson, S.C. and Hull, W.L., Effects of

Turbulence on Spark Ignition Engine Combustion. SAE Technical Paper

Series (760160), 1976.

333

[135] Lavoie, G.A., Correlations of Combustion Data for SI Engine Calculation -

Laminar Flame Speed, Quench Distance and Global Reaction Rates. SAE
Technical Paper Series (780229), 1978. [136] Lecointe, B. and Monnier, G., Downsizing a Gasoline Engine Using

Turbocharging with Direct Injection. SAE Technical Paper Series (2003-010542), 2003. [137] Liang, L., Modeling Knock in Spark-Ignition Engines Using a G-equation

Combustion Model Incorporating Detailed Chemical Kinetics. SAE Technical

Paper Series (2007-01-0165), 2007. [138] Lim, C., Hall, L. and Scarfe, M., Formula SAE: Engine. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2006. [139] Louey, B., Formula SAE-A Electrics Team. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2004. [140] Lu, J.H., Ezeboye, D., Iiyama, A., Greif, R. and Sawyer, R.F., Effect of

Knock on Time-Resolved Engine Heat Transfer. SAE Technical Paper Series

(890158), 1989. [141] Lumsden, G., Eddleston, D. and Sykes, R., Comparing Lean Burn and EGR. SAE Technical Paper Series (970505), 1997. [142] Lundstrom, R.R. and Gall, J.M., A Comparison of Transient Vehicle

Performance Using a Fixed Geometry, Wastegated Turbocharger and a Variable Geometry Turbocharger. SAE Technical Paper Series (860104),
1986. [143] MacInnes, H., Turbochargers. HP Books, ISBN 0-912656-49-2, 1984. [144] Magnecor, http://www.magnecor.com. Accessed: 10 June, 2004. [145] Mamiya, N., Masuda, T. and Noda, Y., Thermal Fatigue Life of Exhaust

Manifolds Predicted By Simulation. SAE Technical Paper Series (2002-010854), 2002.

334

[146] MatWeb, http://www.matweb.com. Material Property Data, Accessed: 10 October, 2007. [147] McKee, R.H., McCullough, G., Cunningham, G., Taylor, J.O., McDowell, N., Taylor, J.T. and McCullough, R., Experimental Optimisation of Manifold and

Camshaft Geometries for a Restricted 600cc Four-Cylinder Four-Stroke Engine. SAE Technical Paper Series (2006-32-0070 ), 2006.
[148] Mehrani, P., Knocking in a HAJI Engine. Current Thesis (Ph.D.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2007. [149] Metghalchi, M. and Keck, J.C., Burning Velocities of Air with Methanol,

Isooctane, and Indolene at High Pressure and Temperature. Combustion

and Flame, Vol.48, pp. 191-210, 1982. [150] Miliken, D., Durability Aspects of Turbocharged vs Naturally Aspirated

Racing Engines. SAE Technical Paper Series (2002-01-3362), 2002

[151] Miliken, W.F. and Miliken, D.L., Race Car Vehicle Dynamics. SAE International, ISBN 1-56091-526-9, 2005. [152] Mitsubishi, Lancer

EVO-7 Piston Technical Drawings. Supplied by

Mitsubishi-Australia, 2002. [153] Moen, R.D., Nolan, T.W. and Provost, L.P., Improving Quality Through

Planned Experimentation. First Edition, McGraw-Hill., ISBN 0070426732,

1991. [154] Monaghan, 2000. [155] Montazersadgh, F.H., Dynamic Load and Stress Analysis of a Crankshaft. SAE Technical Paper Series (2007-01-0258), 2007. M.L.,

Future Gasoline and Diesel Engines - Review.

International Journal of Automotive Technology, Vol. 1, No. 1, pp. 1-8,

335

[156] Morel, T. and Keribar, R., A Model for Predicting Spatially and Time

Resolved

Convective

Heat

Transfer

in

Bowl-in-Piston

Combustion

Chambers. SAE Technical Paper Series (850204), 1985.

[157] Moroso-Performance-Products, September, 2006. [158] Motec, Motec Interpreter. Software, 2003. [159] Mott, R.L., Machine Elements in Mechanical Design. Third Edition, PrenticeHall Inc., ISBN 0-13-841446-7, 1999. [160] Muller, U.C., Bollig, M. and Peters, N., Approximations for Burning

http://www.moroso.com. Accessed: 15

Velocities and Markstein Numbers for Lean Hydrocarbon and Methanol Flames. Combustion and Flame, Vol. 108, pp. 349-356, 1997.
[161] Muranaka, S., Takahi, Y. and Ishida, T., Factors Limiting the Improvement

in Thermal Efficiency of S.I. Engine at Higher Compression Ratio. SAE

Technical Paper Series (870548), 1987. [162] Nakamura, Y., Small High Speed, High Performance Gasoline Engine. SAE Technical Paper Series (640664), 1964. [163] Namazian, M. and Heywood, J.B., Flow in the Piston-Cylinder-Ring Crevices

of a Spark Ignition Engine: Effect on Hydrocarbon Emissions, Efficiency and Power. SAE Technical Paper Series (820088), 1982.
[164] National-Institute-of-Standards-and-Technology, Accessed: 20 June, 2007. [165] NCDC, 2007. [166] Negus, C.H., An Interactive Chemical Equilibrium Solver for the Personal

http://webbook.nist.gov.

ftp://ftp.ncdc.noaa.gov/pub/data/anomalies/annual.land.90S.90N.

Climatic Data Center, U.S. Department of Commerce, Accessed: 10 August,

Computer. Thesis (M.Eng.Sc.), The Faculty of the Virginia Polytechnic

Institute and State University, Blacksburg, Virginia, 1997.

336

[167] OConnor, F. and Smith, M., Variable Nozzle Turbochargers for Passenger

Car Applications. SAE Technical Paper Series (880121), 1988.

[168] Oak-Ridge-National-Laboratory, Accessed: 24 May, 2007. [169] Okamoto, K., Ichikawa, T., Saitoh, K., Oyama, K., Hiraya, K. and Urushihara, T., Study of Antiknock Performance under various Octane

http://cdiac.esd.ornl.gov/ftp/trends/co2.

Numbers and Compression Ratios in a DISI Engine. SAE Technical Paper

Series (2003-01-1804), 2003. [170] Olikara, C. and Borman, G.L., A Computer Program for Calculating

Properties of Equilibrium Combustion Products with some application to I.C. Engines. SAE Technical Paper Series (750468), 1975.
[171] Otobe, Y., Goto, O., Miyano, H., Kawamoto, M., Aoki, A. and Ogawa, T.,

Honda Formula 1 Turbocharged V6 1.5 liter Engine. SAE Technical Paper

Series (890877), 1989. [172] Overington, M.T. and Boer, C.D., Vehicle Fuel Economy - High

Compression Ratio and Supercharging Compared. SAE Technical Paper

Series (840242), 1984. [173] Paterson, D.J. and Henein, N.A., Emissions from Combustion Engines and

Their Control. Report, Ann Arbor Science Publishers, 1972.

[174] Petitjean, D., Bermardini, L., Middlemass, C. and Shahed, S.M., Advanced

Gasoline

Engine

Turbocharging

Technology

for

Fuel

Economy

Improvements. SAE Technical Paper Series (2004-01-0988), 2004.

[175] Pietrobello, G.P., Cocchi, M. and Bolleta, A., Piston and Piston Ring Design

and Development for the Formula 1 Engine. SAE Technical Paper Series
(942518), 1994.

337

[176] Poepperi,

M.,

Hedrich,

G.,

Heusler,

H.,

Koehler,

D.,

Marin,

B.,

Rothenberger, P. and Scharrer, O., Adaptation of a Dual Continuous

Variable Cam Phasing to a 4-Valve, 4-Cylinder Engine Thermodynamic Benefits and Engine Hardware Requirements. SAE Technical Paper Series
(2006-01-0408), 2006. [177] PolyFlex-Australia, http://www.polyflex.com.au/. Coupling Catalog, 15 June, 2004. [178] Poulos, S.G. and Heywood, J.B., The Effect of Chamber Geometry on

Spark-Ignition Engine Combustion. SAE Technical Paper Series (830334),

1983. [179] Prakash, V., Venkatesh, D.N. and Shrinivasa, U., The Effect of Bearing

Stiffness on Crankshaft Natural Frequencies. SAE Technical Paper Series

(940697), 1994. [180] Quader, A.A., Kirwan, J.E. and Grieve, M.J., Engine Performance and

Emissions Near the Dilute Limit With Hydrogen Enrichment Using An OnBoard Reforming Strategy. SAE Technical Paper Series (2003-01-1356)
2003. [181] Rai, H.S., Brunt, M.F.J. and Loader, C.P., Quantification and Reduction of

IMEP Error Resulting from Pressure Transducer Thermal Shock in an S.I. Engine. SAE Technical Paper Series (1999-01-1329), 1999.
[182] Randolph, A.L., Cylinder-Pressure-Based Combustion Analysis in Race

Engines. SAE Technical Paper Series (942487), 1994.

[183] Rassweiler, G.M. and Withrow, L., Motion Pictures of Engine Flames

Correlated with Pressure Cards. SAE Transaction, Vol. 42, 1938.

[184] Renold,

http://www.renold.com/Support/Catalogue_Download.

Chain

Catalogue, Accessed: 15 July, 2004. [185] Ricardo-Inc., WAVE. Software, Burr Ridhe, IL, USA, 1996. [186] Ricardo-Inc., WAVE Manual. Operator Manual, Burr Ridhe, IL, USA, 1996.

338

[187] Rice, A. and Whiten, M., Formula SAE-A Engine. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2001. [188] Richardson, S.H., Shaw, C.T., Lee, D.J. and Pierson, S., Modeling the Effect

of Plenum-Runner Interface Geometry on the Flow Through An Inlet System. SAE Technical Paper Series (2000-01-0569), 2000.
[189] Rodrigues, M. and Porto, S., Torque Plate Honing on Block Cylinder Bores. SAE Technical Paper Series (931679), 1993. [190] Rogers, T.E., Basic Digital Filter Design Routines. Textbook, Stearns & Hush, 1997. [191] Rosenkranz, H.G., Watson, H.C., Bryce, W. and Lewis, A., Drivability, Fuel

Consumption and Emissions of 1.3 litre Turbocharged Spark Ignition Engine developed as a Replacement for a 2 litre Normally Aspirated Engine. Proc. I.Mech.E., C 118/86: pp. 139-150, 1986.
[192] Rothe, M., Heidenreich, T., Spicher, U. and Schubert, A., Knock Behaviour

of SI Engines: Thermodynamic Analysis of Knock Onset Locations and Knock Intensities. SAE Technical Paper Series (2006-01-0225), 2006.
[193] Rototest-Research-Institute, http://www.rri.se. Accessed: March 8, 2007. [194] Rhl, C., Energy in Perspective. BP Statistical Review of World Energy, 2007. [195] Russ, S., A Review of the Effect of Engine Operating Conditions on

Borderline Knock. SAE Technical Paper Series (960497), 1996.

[196] Russ, S. and Scholl, D., Air-Fuel Ratio Dependence of Random and

Deterministic Cyclic Variability in a Spark-Ignited Engine. SAE Technical

Paper Series (1999-01-3513), 1999. [197] SAE-International,

http://students.sae.org/competitions/formulaseries/.

Formula SAE, 6 October, 2004.

339

[198] SAE-International, Traveling the Long Road to Gasoline Direct Injection. Automotive Engineering International, SAE International Magazine, June Edition, 2006. [199] Salazar, V. and Ghandhi, J.B., Liquid Fuel Effects on the Unburned

Hydrocarbon Emissions of a Small Engine. SAE Technical Paper Series

(2006-32-0033), 2006. [200] Schneider, E.W. and Blossfeld, D.H., Effect of Break-In and Operating

Conditions on Piston Ring and Cylinder Bore Wear in Spark-Ignition Engines. SAE Technical Paper Series (2004-01-2917), 2004.
[201] Schneider, E.W., Blossfeld, D.H., Lechman, D.C., Hill, R.F., Reising, R.F. and Brevick, J.E., Effect of Cylinder Bore Out-of-Roundness on Piston Ring

Rotation and Engine Oil Consumption. SAE Technical Paper Series

(930796), 1993. [202] Shahed, S.M., The Power of Turbocharging. SAE International Automotive Engineering, September, 2005. [203] Shell-Australia, www.shell.com.au. Accessed: 15 August, 2007. [204] Shigley, J.E. and Mischke, C.R., Mechanical Engineering Design. ISBN 0-07008303-7. 2001. [205] Smart-Australia, www.smartaustralia.com.au. 2007. [206] Smith, C., Tune to Win. Aero Publishers Inc., ISBN 0-9651600-3-3, 1978. [207] Song, J., Seo, Y. and Sunwoo, M., Effects of Ignition Energy and System Accessed: 20 February,

on Combustion Characteristics in a Constant Volume Combustion Chamber.

SAE Technical Paper Series (2000-05-0016), 2000. [208] Sonntag, R.E., Borgnakke, C. and Van-Wylen, G.J., Fundamentals of

Thermodynamics. Fifth Edition, John Wiley & Sons, ISBN 0-471-18361-X,

1998.

340

[209] Spindt, R.S., Air Fuel Ratios from Exhaust Gas Analysis. SAE Technical Paper Series (650507), SAE Transaction, Vol. 74, 1965. [210] Steinmuller, K. and Khan, M., Formula SAE-A Engine Team. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003. [211] Stone, R., Introduction to Internal Combustion Engines. Third Edition, Society of Automotive Engineers Inc., ISBN 0-7680-0495-0, 1999. [212] Subhi, E. and Cronk, R., Formula SAE-A Drivetrain. Thesis (B.E.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003. [213] SuperFlow, SF600 Operator Manual. 2001. [214] Suzuki, RF600 Operator Manual. 1996. [215] Suzuki, GSX-R600 Operator Manual. 1998. [216] Suzuki, S., Toyama, K. and Konda, N., Development of As-Sintered P/M

Connecting Rod for Automobiles. SAE Technical Paper Series (880168),

1988. [217] Swarts, A., Yates, A., Viljoen, C. and Coetzer, R., A Further Study of

Inconsistencies between Autoignition and Knock Intensity in the CFR Octane Rating Engine. SAE Technical Paper Series (2005-01-2081), 2005.
[218] Systat-Software-Inc., http://www.systat.com. SigmaPlot Ver. 9, Accessed: 10 January, 2004. [219] Tait, N., Piston and Ring Design for Performance Engines. Report, MahleACL, 2001. [220] Taylor, C.F., The Internal Combustion Engine in Theory and Practice, Vol. 1

and 2. The MIT Press, ISBN 0262700271, 1977.

341

[221] Taylor, J., McKee, R., McCullough, G., Cunningham, G. and McCartan, C.,

Computer Simulation and Optimisation of an Intake Camshaft for a Restricted 600cc Four-Stroke Engine. SAE Technical Paper Series (2006-320071), 2006. [222] Teraji, A., Tsuda, T., Noda, T., Kubo, M. and Itoh, T., Development of a

Novel Flame Propagation Model (UCFM: Universal Coherent Flamelet Model) for SI Engines and Its Application to Knocking Prediction. SAE
Technical Paper Series (2005-01-0199), 2005. [223] Thring, R.H. and Overton, M.T., Gasoline Engine Combustion The High

Ratio Compact Chamber. SAE Technical Paper Series (820166), 1982.

[224] Tinker, D., WaveView Manual. Eagle Technology [225] Tomanik, E., Sobrinho, R.M.S. and Zecchinelli, R., Influence of Top Ring

Gap Types at Blow-By of Internal Combustion Engines. SAE Technical

Paper Series (931669), 1993. [226] Topinka, J.A., Gerty, M.D., Heywood, J.B. and Keck, J.C., Knock Behavior

of a Lean-Burn, H2 and CO Enhanced, SI Gasoline Engine Concept. SAE

Technical Paper Series (2004-01-0975), 2004. [227] Toulson, E., Watson, H.C. and Attard, W.P., The Effects of Hot and Cool

EGR with Hydrogen Assisted Jet Ignition. SAE Technical Paper Series
(2007-01-3627), 2007. [228] Toulson, E., Watson, H.C. and Attard, W.P., The Lean Limit and Emissions

at Near-Idle for a Gasoline HAJI System with Alternative Pre-Chamber Fuels. SAE Technical Paper Series (2007-24-0120), 2007.
[229] Towers, J.M. and Hoekstra, R.L., Engine Knock, A Renewed Concern in

Motorsports - A Literature Review. SAE Technical Paper Series (983026),

1998.

342

[230] Tully, E.J. and Heywood, J., Lean-Burn Characteristics of a Gasoline Engine

Enriched with Hydrogen from a Plasmatron Fuel Reformer. SAE Technical

Paper Series (2003-01-0630), 2003. [231] US-Census-Bureau, Highway Statistics. US Department of Transportation, Federal Highway Administration, 2005. [232] US-EPA, Annual Fuel Economy Reports, http://www.epa.gov/. 5 June, 2007. [233] US-EPA, Light-Duty Automotive Technology and Fuel Economy Trends:

1975

through

2007.

(Report:

EPA420-R-07-008),

Compliance

and

Innovative Strategies Division and Transportation and Climate Division, Office of Transportation and Air Quality, September, 2007. [234] Wall_Colmonoy_Corporation, June 30, 2005. [235] Wallis, T., The Most Amazing F1 Engine you Never Saw. Motorsport News, February, 2007. [236] Wallis, T., The Bishop Rotary Valve. AutoTechnology Fisita Magazine, Motorsports Edition, November 2007. [237] Walsh, M.P., Croes, B., Ayala, A., Corey, R. and DeLucchi, M., Motor

http://www.wallcolmonoy.com.

Accessed:

Vehicles: Overview, Alternatives and Issues. Air Pollution as Climate Forcing: Alternative Scenarios Their Benefits and Costs, http://www.walshcarlines.com. Accessed: 15 July, 2007.
[238] Wang, G., Deljouravesh, R., Sellens, R.W., Olesen, M.J. and Bardon, M.F.,

An Optical Spray Pattern Analyzer. ILASS Americas '97, 1997.

[239] Ward's-Communications, Wards World Motor Vehicle Data. 2003. [240] Ward's-Communications, Ward's World Motor Vehicle Data. 2005.

343

[241] Watanabe, Y., Shiratani, K., Iwanaga, S. and Nishino, K., Thermal Fatigue

Life Prediction for Stainless Steel Exhaust Manifold. SAE Technical Paper
Series (980841), 1998. [242] Watson, H.C., A 3-fold Increase in the Power Output of an Engine for

Racing. I.Mech.E. Australian Southern Sub-Branch. I.E. Australian Meeting,

1970. [243] Watson, H.C., High Performance Vehicles. Lecture Notes, Department of Mechanical and Manufacturing Engineering, University of Melbourne, 1997. [244] Watson, H.C. and Andrews, G.E., Road Transport Engine Emissions Course. Lecture Notes, Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003. [245] Watson, H.C., Gauci, A., Yousuff, F. and Boretti, A., Optimizing the Design

of the Air Flow Orifice or Restrictor for Race Car Applications. SAE
Technical Paper Series (2007-01-3553), 2007. [246] Watson, H.C. and Milkins, E.E., Cylinder Head Design - Modern Engine

Developments. Lecture 7246/1, SAE-A, Melbourne, 1972.

[247] Watson, H.C. and Milkins, E.E., Power Generation Systems. Lecture Notes, Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003. [248] Watson, H.C., Milkins, E.E., Roberts, K. and Bryce, W., Turbocharging for

Fuel Efficiency. SAE Technical Paper Series (830014), 1983.

[249] Watson, H.C., Milkins, E.E., Roberts, K., Deslandes, J.V. and Bryce, W.,

Turbocharging for the Fuel Efficient Urban Car. SAE Japan, Proc. 2nd
International Pacific Conference on Automotive Engineering, pp. 368-381, SAE Technical Paper Series (830192), 1983. [250] Watson, N., Turbochargers for the 1980's - Current Trends and Future

Prospects. SAE Technical Paper Series (790063), SAE Transaction, Vol. 88,
1979.

344

[251] Watson, N. and Banisoleiman, K., A Variable-Geometry Turbocharger

Control System for High Output Diesel Engines. SAE Technical Paper Series
(880118), 1988. [252] Watson, N. and Janota, M.S., Turbocharging the Internal Combustion

Engine. Macmillan Press, ISBN 0-333-242904, 1982.

[253] Wentworth, J.T., Piston and Ring Variables Affect Exhaust Hydrocarbon

Emissions. SAE Technical Paper Series (680109), SAE Transaction, Vol 77,
1968. [254] Westin, F., Optimization of Turbocharged Engines Transient Response with

Application on a Formula SAE/Student Engine. SAE Technical Paper Series

(2005-01-2113), 2005. [255] Willeke, W., Mohrmann, R., Seifert, T. and Hartmann, D., Fatigue Life

Simulation for Optimized Exhaust Manifold Geometry. SAE Technical Paper

Series (2006-01-1249), 2006. [256] Woschni, G., A Universally Applicable Equation for the Instantaneous Heat

Transfer Coefficient in the Internal Combustion Engine. SAE Technical

Paper Series (670931), 1967. [257] WRI, Calculation based on compiled data from various editions of: International Energy Agency, Energy Statistics and Balances of Non-OECD Countries and Energy Statistics and Balances of OECD Countries (Organisation for Economic Co- Operation and Development, Paris, various editions). 2006. [258] WWU-Formula-SAE, http://dot.etec.wwu.edu/fsae/. Western Washington University, 8 January, 2006. [259] Yagi, S., Fujiwara, K., Kuroki, N. and Maeda, Y., Estimation of Total Engine

Loss and Engine Output in Four Stroke S.I. Engines. SAE Technical Paper
Series (910347), 1991.

345

[260] Yagi, S., Isgizuya, A. and FujiiI, I., Research and Development of High

Speed, High Performance, Small Displacement Honda Engines. SAE

Technical Paper Series (700122), 1970. [261] Yagi, S., Ishibasi, Y. and Sono, H., Experimental Analysis of Total Engine

Friction in Four Stroke SI Engines. SAE Technical Paper Series (900223),

1990. [262] Yamaguchi, K., Sato, A., Goto, E., Fujiki, R., Kawai, Y. and Nakata, K.,

Development of a New Metal Cylinder Head Gasket. SAE Technical Paper

Series (980844), 1998. [263] Yamamoto, H., Motorcycle engines:

Past, Present and Future. SAE

Technical Paper Series (1999-01-3347), 1999. [264] Yates, A.B.Y., Swarts, A. and Viljoen, C.L., Correlating Auto-Ignition Delays

and Knock-Limited Spark-Advance Data for Different Types of Fuel. SAE

Technical Paper Series (2005-01-2083), 2005. [265] Yoshimura-Exhausts, http://www.yoshimura-rd.com. January, 2003. [266] Zakis, G., Alternative Ignition Systems for CNG in Diesel Applications. Thesis (M.Eng.Sc.), Department of Mechanical and Manufacturing Engineering, University of Melbourne, 2003.

346

Appendix A
P O W E R E D B Y W A T T A R D

Sponsor Acknowledgments

A sincere thank you to all sponsors whose generosity made the UniMelb WATTARD engine possible. ACL Bearing Company APEP Pistons Argo Engineering Bishop Innovation Bohler Uddeholm CadCore Pty Ltd Concentric Asia Pacific Cosway Motorcycles Davies Craig Electromold Farley Laserlab Garrett Honeywell Kawasaki Australia Magnesium Technologies MAME Centre for Manufacturing M&W Ignitions Penrite Lubricants QMI Solutions Ricardo Robert Bosch Siemans VDO Southside Cylinder Heads University of Melbourne Wade Camshafts Whitehorse Industries Tasmania, Australia Victoria, Australia New South Wales, Australia New South Wales, Australia Victoria, Australia Victoria, Australia South Australia, Australia Victoria, Australia Victoria, Australia Victoria, Australia Victoria, Australia New South Wales, Australia New South Wales, Australia Victoria, Australia Victoria, Australia New South Wales, Australia Victoria, Australia Queensland, Australia Sussex, United Kingdom Victoria, Australia Victoria, Australia Victoria, Australia Victoria, Australia Victoria, Australia Victoria, Australia

347

348

Appendix B
P O W E R E D B Y W A T T A R D

Formula SAE

B.1 Introduction
In this appendix, background material relevant to FSAE is presented, including an explanation of the competition and its history. Furthermore, FSAE engine rules and regulations together with a review of various powertrain used in the competition is given. The advantages and disadvantages of fitting OEM high performance engines in FSAE trim are also discussed, along with engine performance and packaging concepts which ultimately affect vehicle performance in competition.

B.2 What is FSAE?

FSAE is a competition that was devised by the Society of Automotive Engineers (SAE - International) to raise tertiary students awareness of the automotive industry. For the purpose of the competition, students are to assume that a The intended sales manufacturing firm has engaged them as a production team to design and construct a prototype open wheel racecar for evaluation. market is the non-professional weekend autocross racer. Vehicles constructed are then scrutinized by an international panel of judges in various categories, with a breakdown of points given in Table B.1 [197]. Categories include static inspection and engineering design, solo performance trials and high performance track endurance.

349

Table B.1: FSAE event points breakdown [197]. Static Events Presentation Engineering Design Cost Analysis Acceleration Dynamic Events Skid-Pad Event Autocross Event Fuel Economy Event Endurance Track Event Total Points 75 150 100 75 50 150 50 350 1,000

Originally conceived in the US in 1981 [197], the formula has gained worldwide popularity in the engineering community with yearly sanctioned competitions held in all parts of the world, including Australasia, Brazil, Italy, UK and the US. All events are open to entrants from any country, with international competition encouraged. The inaugural Australasian FSAE competition attracted six local entrants in December 2000 [16] with the competition quickly establishing itself as the premier event for mechanical engineering students. Past Australasian events have also attracted international entrants from all parts of the world including Germany, India, Japan, New Zealand, Sweden and the UK [16]. Figure B.1 shows that the quality of the vehicles entered in past Australasian competitions are very professional in appearance. Local interest has grown rapidly, with 21 entrants from almost every university with a relevant department participating in the 2005 competition [13]. The success of the local competition is a great credit to its US originators and SAE support from the president, officers and volunteers of the Australasian branch.

Figure B.1: International and local entries at the 2004 Australasian FSAE competition.

350

B.3 Engine Rules and Regulations

Vehicles entering any FSAE event worldwide must conform to the rules and regulations of the competition [197]. An Addendum may be issued to reflect any changes or corrections made to the rules to suit the local competition. The FSAE rules pertinent to engine selection and design are summarized below [197]. The engine used to power the car must be of a four-stroke piston design. More than one engine may be used. Total engine displacement may not exceed 0.61 L per cycle. All engine airflow must pass through a circular restrictor, with dimension constraints not exceeding 20 mm in diameter. Throttle location must be upstream of the restrictor. Specifically, the restrictor must be placed in the intake system between the throttle and the engine. Throttling the engine downstream of the restrictor is prohibited. Turbochargers or superchargers are allowed if the team designs the application. However, the only sequence allowed is throttle, restrictor, compressor, engine. Only ambient air may be used to cool an intercooler. Any transmission and drivetrain combination may be used. Each car must be equipped with an on-board starter, and be able to start without any outside assistance at any time during the competition. Water-cooled engines must only use plain water, or plain water with cooling system rust and corrosion inhibitor at concentration levels not exceeding 1.5% by volume. Premium unleaded gasoline must be used. The temperature of the fuel introduced into the fuel system must not be altered by artificial means. All parts of the fuel storage and supply system including the complete air intake system must lie within the surface defined by the top of the main roll hoop and the outside edge of all four tires. The fuel tank must not exceed 7.57 liters in volume. The maximum permitted sound level shall not exceed 110 dB at the test engine speed. The test speed corresponds to a mean piston speed of 15.24 m/s (50 ft/s) rounded to the nearest 500 rev/min. Measurements will be made in the horizontal plane, 0.5 m from the end of the exhaust outlet at a 45 angle.

351

B.4 Improving Engine Performance

The primary objective for improving engine performance in FSAE is to improve the competitiveness of the vehicle. This can be done through engine specific performance increases coupled with packaging improvements.

B.4.1 Specific Output

Improving FSAE specific engine output is conceptually no different to any other SI engine. The brake specific output can be improved by increasing air consumption and hence fuel and energy, improving the combustion process or reducing the mechanical losses [98]. However, peak air consumption and hence peak power is primarily limited by the intake flow restriction [13, 15]. One method in improving engine performance is to extend the maximum power speed range, which increases on road performance during vehicle operation [13]. This is often cited as increasing the engine power integral [198].

B.4.2 Packaging
Engine and vehicle packaging are important parameters to consider during engine selection since they significantly affect vehicle dynamic characteristics, as the engine contributes a large portion of the vehicle mass. Hence, minimizing the engine mass, COG and volume improves the handling characteristics of the vehicle [151]. This also allows student designers more freedom in vehicle layout and design in order to optimize vehicle specifications and weight distributions. Engine layout and configuration in FSAE vehicles also affects auxiliary components such as manifold geometry, cooling and fueling systems. The design and placement of these components and systems can also be optimized with similar goals of improving on track vehicle performance. A ratio of engine power to vehicle mass can be used as a performance measure when comparing different engine types and vehicle configurations. Although larger capacity NA engines generally have brake output advantages when compared to smaller capacity engines, they suffer from increased fuel consumption coupled with increased complexity and mass. FSAE engine packages. This makes engine selection difficult for most teams. Table B.2 outlines the two extremes of typical

352

Table B.2: Engine power to vehicle mass comparisons for two extreme cases in FSAE engine selection. Engine Configuration and Capacity
Four cylinder (0.6 L) Single cylinder (0.45 L)

Power (kW)
~60 ~35

Vehicle Mass (kg)

~220 ~160

Power/Weight Ratio (kW/kg)

~0.27 ~0.22

B.5 Review of Past Formula SAE Powertrains

With several hundred FSAE teams located around the world, it is not surprising to have such a diverse field of powertrains used in this formula. With teams having varying philosophies on what makes a competitive FSAE package, engines are chosen to meet specific team design goals. Engine selection can be divided into two main categories, smaller capacity and larger capacity, with the option of either running NA or with forced induction. The advantages and disadvantages of these differing engine packages are summarized in Table B.3.
Table B.3: Advantages and disadvantages of differing FSAE engine packages. Engine Intake Background Advantage Disadvantage

NA Large Capacity Forced

Usually four cylinder 0.6 L OEM motorcycle engines

Simplicity Peak power

Weight/COG Packaging

Usually four cylinder 0.6 L OEM motorcycle engines that have been TC

Mid-range and peak power

Complexity Reliability Weight/COG Packaging

NA Small Capacity Forced

Usually single cylinder 0.25 - 0.45 L OEM motocross engines

Packaging Weight/COG Simplicity Fuel efficiency

Poor engine performance

Very rare, usually single cylinder 0.25 - 0.45 L motocross engines that have been SC or TC

Packaging Weight/COG

Complexity Reliability

353

Forced induction powertrains first entered the US competition in 1985, with both SC and TC entries [197]. The problems associated with forced induction are well documented in the literature, with increasing combustion pressures and temperatures reducing engine reliability [98, 220]. Applying boosted technology to FSAE based motorcycle and motocross engines results in BMEP increases when compared to unmodified OEM configurations. These OEM configurations already feature high CRs with highly stressed internal components. Generally speaking, internal modification is required to ensure engine reliability, except for very low boost applications. Many FSAE teams have previously experimented with forced induction during the engine development phase, but have abandoned further research due to engine reliability concerns, time constraints and a lack of expertise. However, with internal engine modification and an extensive engine development program, TC four cylinder 0.6 L motorcycle engines have been successful, winning the US competition on several occasions, the first in 1990 by American University UT Arlington. In the past decade, this package has been very competitive with Cornell University frequently winning US competitions. Locally, it has also had success with the University of Wollongong winning the 2004 Australasian event. Over the past decade, single cylinder engines have generally been uncommon due to a lack of performance and absence of a suitable, four stroke, lightweight engine design. However, in recent years, there has been an increasing trend in single cylinder entries, spurred by the recent introduction of four stroke technology to motocross bikes. Furthermore, vehicle masses have greatly reduced due to composites and improved design, making the underpowered single cylinder engine competitive against larger and heavier powertrains. Besides the reduced capacity, single cylinder performance reductions are associated with poor cylinder filling due to the intake flow restriction. Hence, the total induced air consumption occurs over a small CA interval. To smooth out the large flow fluctuations through the restrictor, large volumes downstream are needed, compromising transient response. However, some teams are willing to compromise brake performance for mass and packaging gains, as was successful for RMIT University who have won recent local and international competitions.

354

TC single cylinder FSAE engines are extremely rare as they require large volumes between the engine and turbocharger to minimize the velocity fluctuations experienced by the exhaust turbine [13, 15]. These large plenum volumes in the intake and exhaust system can lead to poor transient response due to increased mass transport times. Repositioning the throttle system in the intake system would minimize this effect, however upstream throttle location is mandatory by FSAE regulation. Turbo lag further deteriorates vehicle transient response. Past unique FSAE engines have been rare, with only one student entry from Western Washington University ever entered, an ambitious 19,000 rev/min 0.55 L transaxle V8 (Figure B.2). The team was the first to design and construct its own custom engine specific for the formula. Although unsuccessful in finishing any events, it was still a tremendous achievement for all 12 team members. The unit utilized many off-the-shelf components including cylinder heads, pistons and the transmission. In addition, they manufactured their own crankcase, crankshaft and auxiliaries. The unit only ran at one US FSAE competition in 2001 and suffered a huge penalty with the engine not having an onboard starting system. It was scrapped in subsequent years due to the added complexity, with reliability still being a major concern after several years of development. Performance was also not fantastic, with the engine struggling to match the 0.6 L four cylinder engines, due to the excessive friction losses associated with high engine speeds and many moving components. The unit was also heavy, with the bare engine weighing in at over 70 kg [258].

Figure B.2: Two extremes in FSAE engine selection. (Left): OEM Yamaha YZF 0.45 L single cylinder motocross engine. (Right): Western Washingtons prototype 0.55 L V8.

355

B.6 New Engine Targets for Formula SAE

Performance, efficiency and packaging targets were set for the new FSAE engine design when compared with a typical, four cylinder, 0.6 L reference engine. The reference engine was the Melbourne University FSAE teams previous powertrain, a Suzuki GSX-R600. This engine had proven to be a competitive and well developed four cylinder package with ample test data available for comparison. Specific areas targeted for improvement for the new engine design are outlined below. Maintain choked airflow through the intake restriction over half the engine speed Increased peak power by 5-10% through a combination of TH improvements and friction loss reductions associated with reducing the swept capacity (Chapter 3.3) 20% reduction in vehicle fuel consumption over the endurance track event 25% reduction in engine mass 50% reduction in physical size 25% reduction in the COG height Engine design packaged to suit a FSAE styled vehicle Maximum throttle response making the engine as tractable as possible with a minimum number of gear changes

356

Appendix C
P O W E R E D B Y W A T T A R D

Fuel Properties and Emission Regulations

C.1 Fuel Properties

A high quality pump gasoline labeled Shell Optimax was used as the base fuel in experiments, with manufacturer specifications listed in Table C.1 [203]. Table C.2 compares gasoline fuel properties to other alternative fuels used in SI engines [228].
Table C.1: Shell Optimax fuel specifications [203]. Property
Octane Number (ASTM-D2699) Density (ASTM-D4052) Molecular mass Specific Heat Volatility (ASTM D86) 10% 50% 90% Final evaporated evaporated evaporated Boiling Point K K K K % vol % mass 318 378 428 469 3.3 0.015

Unit
RON kg/m
3

Value
98.4 760 114.62 2.4

g/mol kJ/kg K

Benzene (ASTM D3606) Sulphur (ASTM D1266)

357

Table C.2: Properties of various fuels used in SI engines [228]. Property

Energy Density (MJ/kg) Energy Density (kJ/mol) Stoichiometric AFR (by mass) Stoichiometric AFR (by vol.) Fuel (% by mass) in Stoichiometric Mixture w/ Air Fuel (% by vol.) in Stoichiometric Mixture w/ Air Energy in 1 mol Stoichiometric Fuel-Air Mixture (kJ/mol) Density (kg/m3) 0C and 1atm Laminar Flame Speed (m/s) LFL (vol. %) UFL (vol. %) at LFL at UFL Auto Ignition Temperature (C) Minimum Ignition Energy (mJ)

H2
120 240 34.3 2.4 2.8 29.5 70.8 0.09 ~2.3 4 75 10.0 0.14 572 0.02

CNG
45 810 14.5 9.0 6.5 10.0 80.8 0.72 0.39 5.6 15 1.9 0.63 632 0.33

LPG
46.4 2,002 15.7 23.9 6.0 4.0 80.4 2.0 0.45 2.2 9.5 1.9 0.4 500 0.30

Gasoline
44 4,840 14.6 57.7 6.4 1.7 82.5 735 0.4 1.4 7.6 1.2 0.21 530 0.8

To ensure consistency between seasonal and various fuel blends used in experiments, fuel quality testing was completed using an ASTM CFR single cylinder research engine (Table C.3). Fuels included an alternative high grade gasoline (BP Ultimate) with comparable octane rating to Shell Optimax and an 80% gasoline/20% toluene blend. Minor testing was completed with BP Ultimate at several high BMEP conditions to verify experimental results found with the Optimax fuel. Moreover, blends of Toluene C7H8 (Methylbenzene) and Shell These Optimax were used during development to alleviate abnormal combustion concerns related to the fuel quality (Chapter 5.4.1 and Appendix J.7). blends significantly increased the fuel octane as shown in Table C.3.
Table C.3: Documented and measured RON numbers for fuels used during development and experiments. Fuel Blend
Shell Optimax BP Ultimate 80% Shell Optimax/20% Toluene

Published RON Measured RON Testing Method

98.4 98 108 97.8 97.5 112 CFR (ASTM-D2699) CFR (ASTM-D2699) CFR (ASTM-D2699)

358

C.2 Emission Regulations

Table C.4: Current and future US federal emissions regulations and Californian standards for light-duty vehicles [56, 61]. Category CO NMHC/NMOG3 g/mile g/mile
0.25 (0.31) --- (0.31) 0.125 (0.156) 0.075 (0.090) 0.04 (0.055) 0.0 0.075 (0.090) 0.040 (0.055) (0.01) 0.0 --- (4.2) 3.4 (4.2) 3.4 (4.2) 1.7 (2.1) 0.0 3.4(4.2) 1.7(2.1) (1.0) 0.0

NOX g/mile
0.4 (0.6) --- (1.25) 0.4 (0.6) 0.2 (0.3) 0.2 (0.3) 0.0 0.05 (0.07) 0.05 (0.07) (0.02) 0.0

Formaldehyde4 Particulates5 g/mile g/mile

--- (---) --- (---) 0.015 (0.018) 0.015 (0.018) 0.008 (0.011) 0.0 0.015 (0.018) 0.008 (0.011) (0.004) 0.0 --- (---) --- (0.08) --- (0.08) --- (0.08) --- (0.04) 0.0 (0.01) (0.01) (0.01) 0.0

Tier I Gasoline 3.4 (4.2)6 Tier I Diesel TLEV1 LEV1 ULEV1 ZEV1 LEV2 ULEV2 SULEV2 ZEV2 1) 2) 3) 4) 5) 6)

Emissions categories phasing out 2004 - 2007. Emission limits phasing in 2004 onwards. NMOG = reactivity corrected values for alternative fuelled vehicles. Methanol and flexible-fuel vehicles only. Diesels only. Limits for 50,000 miles with 100,000 miles in parentheses.

Table C.5: Future Californian fleet average GHG emission standards [61]. Year
2009 2010 2011 2012 2013 2014 2015 2016

GHG Standard g CO2/mi (g CO2/km) PC/LDT1

323 (201) 301 (188) 267 (166) 233 (145) 227 (142) 222 (138) 213 (133) 205 (128)

CAFE Equivalent mpg (l/100 km) PC/LDT1

27.6 (8.52) 29.6 (7.95) 33.3 (7.06) 38.2 (6.16) 39.2 (6.00) 40.1 (5.87) 41.8 (5.63) 43.4 (5.42)

LDT2
439 (274) 420 (262) 390 (243) 361 (225) 355 (221) 350 (218) 341 (213) 332 (207)

LDT2
20.3 (11.59) 21.2 (11.10) 22.8 (10.32) 24.7 (9.52) 25.1 (9.37) 25.4 (9.26) 26.1 (9.01) 26.8 (8.78)

PC/LDT1 = All passenger cars and light-duty trucks below 3,750 lbs gross vehicle weight LDT2 = All light-duty trucks between 3,750 - 10,000 lbs gross vehicle weight

359

Table C.6: EU emission standards for passenger cars (Category M1*), g/km [61]. Tier Date CO HC Diesel
Euro 1 Euro 2, IDI Euro 2, DI Euro 3 Euro 4 Euro 5 Euro 6 1992.07 1996.01 1996.01
a

HC+NOX

NOX

PM

2.72 (3.16) 1.0 1.0 0.64 0.50 0.50 0.50

0.97 (1.13) 0.7 0.9 0.56 0.30 0.23 0.17

0.50 0.25 0.18 0.08

0.14 (0.18) 0.08 0.10 0.05 0.025 0.005e 0.005e

2000.01 2005.01 2009.09b 2014.09

Gasoline
Euro 1 Euro 2 Euro 3 Euro 4 Euro 5 Euro 6 1992.07 1996.01 2000.01 2005.01 2009.09
b

2.72 (3.16) 2.2 2.30 1.0 1.0 1.0

0.20 0.10 0.10

c

0.97 (1.13) 0.5 -

0.15 0.08 0.06 0.06

0.005d,e 0.005d,e

2014.09

0.10c

* At the Euro 1..4 stages, passenger vehicles > 2,500kg were type approved as Category N1 vehicles Values in brackets are conformity of production (COP) limits a) Until 1999.09.30 (after that date DI engines must meet the IDI limits) b) 2011.01 for all models c) And NMHC = 0.068 g/km d) Applicable only to vehicles using DI engines e) Proposed to be changed to 0.003 g/km using the PMP measurement procedure

360

Appendix D
P O W E R E D B Y W A T T A R D

Restrictor Calculations

D.1 Maximum Theoretical Restrictor Mass Flow Rate

The maximum mass flow rate of an ideal gas through an isentropic nozzle is limited by the stagnation conditions and the minimum cross-sectional area. Hence, no matter how much the downstream pressure is reduced or the dimensions of the duct either side of the throat are changed, this maximum flow rate cannot be exceeded [14, 15, 245]. From this theory, the maximum mass flow rate is defined in Equation D.1 [98, 208]. In this equation, the Mach number (Ma) at the throat is assumed to be one, since the maximum mass flow that can be achieved through a convergent-divergent nozzle occurs when the flow is sonic at the throat region [245].
choked m
+1

P A* = o To

2 1 R +1

(D.1)

Assuming the following conditions, the theoretical maximum air mass flow rate achievable through the restrictor is calculated as in Equation D.2.

Assumed:
Atmospheric conditions, T = 20C, P = 101.3 kPa Adiabatic index for air, = 1.4 Orifice diameter, D = 20 mm

choked = m

(101.3 103 ) (10 10 3 )2 293

1.4 2 1.4 1 287 1.4 + 1

1.4 +1

(D.2)

= 0.075 kg/s

361

D.2 Flow Through a Venturi

Figure D.1 displays three cases of flow through an intake venturi. The first is subsonic flow where the pressure recovery is good (exit pressure is almost unity,

P is small). In the second case with sonic flow at the throat (Ma = 1), the P at
exit is a little larger because of the irreversibility associated with the shock at the throat. In the third case with supersonic flow downstream of the throat, the pressure loss is larger. The objective in restrictor design is to maximize the mass flow through the throat assuming that the engine has enough air consumption to deliver the required pressure drop at the exit.
P/PO

Subsonic Just sonic at throat Supersonic Shock

Critical pressure ratio = 0.528

Ma

A/A5

Figure D.1: Subsonic, near sonic and supersonic flow for air ( = 1.4) in the divergent section of an intake venturi. (Upper): Pressure ratio (P/PO). (Middle): Mach number (Ma). (Lower): Throat area ratio (A/A5) [245].

362

D.3 Compressible Flow Intake Model

A compressible flow analytical model was created for the intake system allowing the calculation of mass and volume flow rates for a variety of conditions together with static pressures and temperatures through the entire system. Models were created for all major induction components including the restrictor, turbocharger, inlet manifold and fuel system. This enabled the restrictor choking engine speeds to be evaluated for a variety of engine configurations and volumetric efficiencies, as outlined in Chapter 3.

D.4 Pre-Restrictor Fuel Injection

Calculations were performed to assess the viability of injecting fuel directly into the restrictor nozzle to reduce intake temperatures and hence increase density (mass flow) at the choked condition (Chapter 3.7).

D.4.1 Restricted Mass Flow

Assumed:
Atmospheric conditions, T = 20C, P = 101.3 kPa, = 1.204 kg/m3 Gasoline stoichiometric mixture, AFR = 14.6:1 Reduction in charge temperature through fuel vaporization caused by the

enthalpy of the phase change hfg, T = 20C [220]

Adiabatic index for a fully vaporized stoichiometric mixture, = 1.35 [98] Gas constant for a fully vaporized gasoline stoichiometric mixture, R = 243

J/kg K

Calculated:
The reduction in charge temperature through fuel vaporization increases the charge density by 26% as shown in equation D.3 but also reduces the speed of sound by 13%, affecting choking limits.

P 101.3 103 = = 1.525 kg/m3 RT 243 273

(D.3)

c = RT = 1.35 243 273 = 300 m/s

(D.4)

363

Thus, the increase in mass flow is dependent on the increase in density and decrease in speed of sound, which is confirmed by recalculating the restricted mass flow using equation D.1 with the altered parameters. This gives a mass flow increase of 11.4% at the choked condition when compared to no fuel being injected. However, the vaporized fuel occupies volume and thus displaces air. The exact amount assuming unit depth with the volumetric AFR is calculated as follows.

nf [FUEL]+ na [AIR]
Using Gasoline C8.26H15.5 :

a CO2 + b H2O + c N2

(D.5)

C8.26 H15.5 + na [O2 + (79/21) N2 ] Balancing: C: H: O: N: 8.26 = a 15.5 = 2b 2na = 2a + b (79/21)na = c

a CO2 + b H2O + c N2

(D.6) (D.7)

b = 7.75 na = 12.13 c = 45.65

(D.8) (D.9) (D.10)

Rewriting: C8.26 H15.5 + 12.13 [ O2 + (79/21) N2 ] 8.26 CO2 + 7.75 H2O + 45.65 N2 (D.11)

AFR volume =

nO2 + nN 2 nf

12.13 + 45.65 = 57.78 :1 1

(D.12)

With an increase of 11.4% in choked mass flow, when accounting for the quantity of fuel injected, the air mass flow is increased by 9.5% as shown in Equation D.13.

57.78 air = Increase of m 111.4 57.78 + 1 100 = 9.5%

(D.13)

364

D.4.2 Manifold Air Temperature

Injecting fuel directly into the restrictor and thus into the turbocharger compressor has benefits in minimizing the temperature rise due to inlet boosting as shown below. This occurs because the turbocharger compresses the mixture at a much lower temperature, minimizing the temperature rise during compression.

Assumed:
Atmospheric conditions, T = 20C, P = 101.3 kPa, = 1.204 kg/m3 Reduction in charge temperature through fuel vaporization caused by the

enthalpy of the phase change hfg, T = 20C [220]

Adiabatic index for a fully vaporized stoichiometric mixture, = 1.35 [98] Pressure ratio due to inlet boosting, Pratio = 2.5 (1.5 bar boost) Turbocharger compressor efficiency, COMP = 0.75 No intercooler No temperature loss through manifold walls

Calculated:
Injecting fuel after the compressor ( = 1.4)
1 T P T ratio + T T =

T manifold

COMP

(D.14)

= 390 K

117C

Pre-injecting fuel into the restrictor before the compressor ( = 1.35)

1 (T T ) P (T T ratio =

)
+

T manifold

COMP

(T

(D.15)

= 371 K

98C

365

D.5 Maximum Power

With a known theoretical maximum air mass flow rate, maximum power can be estimated using the consumed fuel energy. The mass flow rate of fuel can be calculated, assuming a stoichiometric AFR mixture for complete combustion. Although peak power does not occur at stoichiometric conditions (Figure M.2), this offers a good estimate. Hence, the total energy released from the fuel mass can be calculated with the maximum power deduced from an assumed TH value. A maximum brake TH of 30% peak power conditions [124]. is estimated for high performance motorcycle engines [263], with previous FSAE experimental studies finding lower values at

Known:
Theoretical maximum air mass flow rate, choked = 0.075 kg/s (Section D.1) Gasoline LHV = 44,300 kJ/kg [98] Gasoline stoichiometric mixture, AFR = 14.6:1

Assumed:
Discharge coefficient, CD = 0.9 (vena contracta and frictional losses) [58].

Later experimentally found to be 0.96 (Appendix H.5.4)

Assumed maximum air mass flow rate, corrected choked = 0.068 kg/s Peak brake thermal efficiency, TH = 30% [124, 263]

Calculated:
Mass flow rate of fuel

fuel = m

choked m AFR

0.068 = 0.0046 kg/s 14.6

(D.16)

Total energy released

fuel = 44,300 0.0046 = 204 kW LHV m

Maximum power

(D.17)

fuel TH = 204 0.3 = 61.2 kW LHV m

(D.18)

366

Appendix E
P O W E R E D B Y W A T T A R D

Turbocharger Selection

E.1 Introduction
This Appendix provides further detail surrounding turbocharger selection. In

particular, in-depth analysis and reasoning is supplied to support the decision table criteria and weighting information displayed in Chapter 6.2 (Table 6.1), which summarizes the turbocharger selection process for the Garrett GT-12 and GT-15 units. The selection process is only described for these two units as maps from competitive turbochargers could not be obtained. Furthermore, the requirements for the turbocharger were set from the theoretical analysis performed in Chapter 3.3, where maximum air consumption rates due to the flow restriction are outlined together with engine design specifications and operating parameters. These values correspond to the turbocharger delivering pressure ratios at the required airflow rate and corresponding engine speeds.

367

E.2 Aspect
E.2.1 Availability and Cost
Background searches revealed many turbocharger units which satisfied the desired airflow requirements at higher engine speeds. However, few units could meet the operating requirements at low engine speeds, largely associated with the small engine capacity and FSAE airflow objectives (maintaining a choked inlet throughout a wide operating engine speed range). Two units which fulfilled these requirements were the GT-12 and GT-15. These units were manufactured by Garrett Honeywell and were readily available for use with diesel and gasoline engines. Although these units were manufactured offshore, they could be sourced locally through the Australasian branch with relatively short lead times.

E.2.2 Matching
In all applications, turbocharger matching to the engine is a compromise between low and high speed performance. this is delivered. Hence, sizing and matching must take into account two major considerations, the desired level of boost and the rate at which In this particular application, the maximum boost was determined by the airflow required to choke the intake flow restriction at the minimum desired engine speed. With a known desired choked operating speed range set from the theoretical analysis (Chapter 3.3), a preliminary engine turbocharger matching study was performed comparing the GT-12 and GT-15. Operating maps with engine operating points overlaid are displayed in Chapter 6.6 and this Appendix. Initially, the study compared the compressor side matching due to the complexity of the variable nozzle turbine (VNT), with the turbine geometry incorporated shortly afterwards. Calculations are shown in the following pages, with the results highlighting that the GT-12 is more suited to the desired operating conditions. The GT-12 is able to choke the flow restriction at the minimum required engine speed with higher compressor efficiencies, while still operating below the surge line. Although this has reduced compressor work advantages, a further benefit is the reduction in intake temperatures, an important factor in minimizing knock as no intercooler was incorporated into the design (Chapter 3.7)

368

Calculations

Engine

air

consumption

requirements

are

outlined

to

compare

engine

turbocharger matching for both the GT-12 and GT-15. Figure E.1 outlines the test engines maximum air consumption for varying engine speeds, dictated by the flow restrictions choking limits.

Known:
Theoretical maximum air mass flow rate, choked = 0.075 kg/s (Appendix

D.1)
Engine swept volume = 0.43 L Desired choked operating speed range = 6,000 - 10,000 rev/min

Assumed:
Discharge coefficient, CD = 0.9 (vena contracta and frictional losses) [58].

Later experimentally found to be 0.96 (Appendix H.5.4)

Assumed maximum air mass flow rate, corrected choked = 0.068 kg/s

Calculated:
From Figure E.1, a VOL of 150-240% is required to maintain choked flow

within the operating speed range.

400
Theoretical Maximum Airflow Maximum Airflow at C Cd D = 0.9

300 VOL VOL (%)

200

100
Desired choked operating speed range

0 2000

4000

6000

8000

10000

12000

Engine Speed (rev/min)

Figure E.1: Theoretical and assumed test engine air consumption needed to cause choked flow through the intake restriction for varying engine speeds.

369

With calculated values of engine air consumption required to cause choked flow, turbocharger delivery rates are calculated from the flow model (Appendix D.3).

Known:
150 - 240% VOL required to maintain choked flow Maximum pressure ratios and air delivery rates for modern turbochargers

Assumed:
Compressor efficiency, COMP = 0.75

Calculated:
From Figure E.2, the pressure ratio required to achieve choked flow over

the operating speed range (6,000 - 10,000 rev/min) varies from 2.8 - 1.7. With known values of pressure ratio and required mass flow to achieve choked flow, engine operating points can be overlaid on OEM supplied maps, as shown in Figures 6.20 (GT-12) and E.3 (GT-15). As described in Chapter 6.4.3, the pressure drop across the restrictor was limited with wastegate control, as there was no need to reduce the suction pressure when the nozzle was already choked. Hence, the MAP is similar to the pressure ratio across the compressor. Avoiding the compressor surge line is vital in ensuring unit reliability, with weighted values during selection (Table 6.1) based on efficiency values for the required operating conditions. From the maps supplied, the GT-12 has a matching advantage when compared with the GT-15 as it has operating points at higher efficiencies.
Maximum pressure ratio achievable by Garrett GT-12 and GT15

3.0

Compressor Pressure Ratio

2.5

Maximum pressure ratio achievable by IHI RH3 turbocharger

2.0

1.5

Desired choked operating speed range 10,000 rev/min 6,000 rev/min

1.0 100 125 150 175 200 225 250

VOL (%)

Figure E.2: Turbocharger limits and corresponding compressor pressure ratio and engine air consumption required to maintain choked flow over the desired operating speed range.

370

E N I L E G R U S
W O T

Pressure ratio and corresponding mass flow to achieve choked flow at 6,000 rev/min

Pressure ratio and corresponding mass flow to achieve choked flow at ~10,000 rev/min

Figure E.3: Garrett GT-15 compressor map with WOT engine operating points over the choked operating speed range overlaid.

371

Figure E.4: Garrett GT-15 turbine maps (VNT).

372

E.2.3 Implementation
Turbocharger implementation problems were faced for both units due to FSAE rules, which regulated upstream throttle and restrictor locations in order to limit airflow as described in Chapter 6.3 (Figure 6.1). Due to identical bearing and seal designs, both turbocharger units faced equal oil consumption problems under compressor throttling conditions as further outlined in Chapter 6.3. The GT-15 also faced further implementation problems due to complexity and reliability of the VNT feature as described in Section E.2.4.

E.2.4 Control
Intake boost regulation varied substantially between both units. The GT-12

controlled boost pressure by regulating the exhaust mass flow to the turbine, bypassing the unwanted mass through a conventional internal wastegate flap valve (Figure 6.6), with development described in Chapter 6.4.3. However, the GT-15 featured VNT technology, which directs exhaust mass flow onto the turbine wheel through a series of moving vanes as depicted in Figure E.5. As these vanes move, the area between the tips change leading to a variable aspect ratio (A/R ratio). Both turbocharger units controlled moving components for boost regulation via simple diaphragm actuators which in turn limited positional movement, with reference signals taken from the intake manifold.
Nozzle vanes near wide open position

Actuating Ring

Nozzle vanes near closed position

Turbine inlet

Figure E.5: Turbocharger exhaust turbine fitted with VNT technology highlighting the vane position movement.

373

The benefits of VNT are well documented in the literature [143, 167, 251], involving transient response improvements and allowing gains in both high and low speed performance. The technology is especially suited to diesel applications where exhaust gas temperatures are lower than gasoline engines, making the delicate movable vanes more robust and reliable. Hence, the reliability of the GT15 (VNT) was questioned if installed on the test engine as it was initially designed for the diesel application. Furthermore, to obtain the maximum benefit, the control system would need to be carefully designed as a simple mechanically operated boost control system is only suitable with low BMEP engines [251]. High BMEP applications, such as in this one, required a more sophisticated system due to the high exhaust pressure that can be generated with a small turbine area. Micro-electronic control systems have previously been demonstrated but add significant complexity to the system [251]. In passenger vehicle and commercial applications where larger units are used, VNT has been shown to improve transient response with the ability to accelerate the shaft from an initial low speed, in response to the sudden increase in engine torque demand during vehicle acceleration [167]. This is achieved by overcoming the rotor assembly inertia at low engine speeds. However, it is anticipated that the transient benefits from VNT may only be minor in this specific application due to the turbochargers small size, as the rotational inertia is a power function of rotor diameter. Furthermore, the increased flow complexity of the VNT can The reduce turbine efficiencies, which leads to higher exhaust backpressures. cylinder scavenging and increased pumping losses or reduced pumping gain. It was also anticipated that VNT would not deliver torque advantages seen in other applications at both high and low engine speeds due to the intake flow restriction which limited maximum airflow. Hence, high engine speed boost pressure was not required in this application due to the reduction in MAP needed to choke the flow restriction as the engine speed increased. Therefore, the simple wastegate system of the GT-12 was superior for this application as it could match or exceed the VNT at low engine speeds. Further detail surrounding the GT-12 is outlined in Chapter 6.4.3.

increased pressure differential results in high fuel consumption due to poor

374

E.2.5 Mass and Packaging

The GT-12 and GT-15 differ with regards to unit mass and packaging. The GT-15 is heavier and more difficult to package due to its larger size, associated with the VNT. The compressor and centre bearing housing for both units are very similar in mass and size, however the VNT of the GT-15 increases the mass and size of the turbine housing. Hence, the GT-15 has a ~25% increase in volume and a 1.5 kg weight disadvantage, weighting in at 5.2 kg.

E.2.6 Cooling and Lubrication

Both turbocharger units required oil lubrication for the centre housing plain bearing assembly which supported the rotating shaft. acceleration [66], but were unavailable. pump to produce a hydrodynamic film. Ball bearing assemblies were preferable due to their documented benefits in improving shaft rotational For reliable plain bearing operation, The oil quality and feed pressure are filtered engine oil needed to be pressure fed into the assembly by the engine oil known to be vital in ensuring bearing longevity as shaft speeds were expected to exceed 250,000 rev/min. Furthermore, the GT-12 also incorporated water cooling in the centre bearing housing to avoid problems associated with turbocharger oil heat soak after quick engine shutdown (Chapter 6.4.2). Hence the water cooling feature of the GT-12 was initially seen as advantageous in improving engine and turbocharger reliability.

E.2.7 Final Selection

A decision table was used for the final selection process between both units, with results shown in Table 6.1. Hence, the GT-12 was selected with a summary of salient differences in the selection process outlined below.
The GT-12 featured improved matching and hence was more suitable to

FSAE applications where mass airflow is restricted

The GT-12 featured a simple boost control method, which was easily

adjustable through conventional mechanical or electronic techniques

The GT-12 has mass and packaging advantages

375

 The GT12 features water cooling in the centre bearing housing which could

be advantageous in improving reliability

The GT-12 can withstand higher exhaust temperatures due to limitations

associated with the VNT of the GT-15

376

Appendix F
P O W E R E D B Y W A T T A R D

Engine CFD Simulation

F.1 Introduction
Simulation tools in engine design minimize costs and lead times by predicting characteristics of engines that have not been constructed. In order to use these simulation programs effectively and ascertain the validity of the data that is being produced, the operator must have extensive knowledge of the variables and their effects. Hence, a CAE code produced by Ricardo Software Incorporated named WAVE [185] was used to conduct CFD engine simulations. Simulations concentrated on improving engine performance, with selected formulae behind the software outlined in this appendix, together with reference material used to establish formula constants. WAVE is not a parametric software package which converges to an optimal solution for a particular set of variables. The software only simulates results from pre set user defined variables. This makes the engine design and evolution In essence, the heavily reliant on the operator to use his knowledge and skills to use the full capabilities of the software to improve engine performance. experimental data to determine engine parametric constraints. software is just another design tool, together with hand calculations, empirical and

377

F.2 Ricardo WAVE Models

WAVE is a CAE code developed by Ricardo to analyze the dynamics of pressure waves, mass flows and energy losses in ducts, plenums and the intake and exhaust manifolds of various systems, including engines. The software provides a fully integrated treatment of time dependent fluid dynamics and thermodynamics by means of a 1-dimensional finite difference formulation, incorporating a general thermodynamic treatment of working fluids (air, air-hydrocarbon mixtures, liquid fuels and products of combustion [186]). The software models a general network of pipes, volumes and junctions in terms of a set of building blocks. These include:
Constant area or conical pipes or ducts Passages with abrupt changes of area Junctions of multiple ducts Elbows, orifices and plenums Terminators such as infinite plenums (ambient conditions)

Furthermore, the software includes a library of components including engine cylinders and turbochargers. These components can be attached to the pipe network to serve as the sources or absorbers of pulsating flows [186]. Some of these models are now described.

F.2.1 Cylinder Model

Conservation equations for mass and energy (Appendix N) are solved to describe in-cylinder processes for the cylinder model. The conservation of mass equation accounts for the mass changes within the cylinder due to varying inlet and exhaust valve flows. These flow rates have been computed using isentropic compressible flow through a restriction for a given effective area. The discharge coefficients for the computation of these effective areas have been calculated from experimental data gathered on a steady state flow bench as described in Appendix H. The combustion model for premixed SI engines is based on the Sshaped mass fraction burned profile represented by the Wiebe function relationship (Equation F.1), widely used in thermodynamic calculations [98]. This relationship allows the independent input of function shape parameters and burn durations, with the function documented to accurately represent experimentally observed trends of combustion heat release.

378

X b = 1 exp a

m +1

F.1)

where:

Xb = cumulative MFB, = crank angle, o = start of combustion,

= combustion burn duration, constants m = 2 and a = 5 [98]

The fuel rate delivered to the cylinder is computed during simulation by multiplying the calculated air consumption by the user defined AFR. The heat transfer within the cylinder is calculated using the Woschni correlation, as outlined in Appendix N.5.

F.2.2 Pipe Model

The WAVE pipe model incorporates heat transfer, wall friction and additional losses due to flow obstructions within the duct. For smooth walls and turbulent flow, the code calculates the friction coefficient Cf using Equation F.2. coefficient is multiplied by a user defined skin friction factor. This For intake and

exhaust engine ports, the wall friction is set to zero since friction is already included in the valve flow coefficients from physical flow bench experimentation. Otherwise, the skin friction factor is a function of wall roughness.

C f = 0.096
where:

U D

(F.2)

Cf = friction coefficient, = dynamic viscosity, = instantaneous

local density, U = instantaneous flow velocity and D = pipe diameter

The convective heat transfer coefficient hc is computed using the Colburn analogy as outlined in Equation F.3 [107]. The heat transfer coefficient can also multiplied by a user defined factor. According to the Colburn analogy, this multiplier has always been set equal to the skin friction multiplier with exceptions in the modeling of intake and exhaust ports where the multiplier is set to unity as the valve flow coefficients have been found experimentally.

379

hc =
where:

Cf
2

U C p Pr 0.67

(F.3)

hc = convective heat transfer coefficient, Cp = gas heat capacity, Pr

= Prandtl number

Pressure losses P due to pipe geometry (bends) or flow obstructions within the duct have been accounted for via multiplying a coefficient Cpl by the specific kinetic energy. Throttle butterfly and primary exhaust pipe bends are some parameters that are modeled using this technique with Cpl values of 0.1 and 0.5 respectively. P = Cpl (0.5 r U 2) where: P = pressure loss, r = bend radius (F.4)

F.2.3 Complex Pipe Junction Model

Limitations in the code due to dimensional constraints require caution when modeling junctions in the pipe network, which cannot be described in terms of 1dimensional flow. These include plenum chambers and exhaust junctions. These elements can be modeled in a lumped fashion as a set of ducts, orifices and constant pressure volumes with tunable parameters such as volume, surface area, friction and heat transfer coefficients. discharge coefficients and pipe spacing. Parameter data can be derived from experimental pressure traces under varying engine operating conditions, however some operator calibration is often required to match experimental data. Handbooks can also be used to locate empirical correlations for loss coefficients but are usually for steady flow conditions since the values in pulsating flow are specific to each application. The intake plenum has been modeled as one junction connecting the intake runners, with the volume set to that of the entire plenum chamber. The model of the exhaust manifold feeding the turbine is also made up of one junction connecting the two headers, with a duct modeling the flow path from the junction to the turbine housing involute entry. For each pipe entering or leaving the volume, further parameters can be specified such as pipe orientation angles,

380

F.2.4 Turbocharger Model

The turbocharger model in WAVE is configured as two separate components with a compressor and turbine, which are coupled by matching rotational shaft velocities. Flow calculations through these components are carried out at each time step of the simulation with instantaneous input variables of pressure, temperature and common shaft speeds aiding in the interpolation of instantaneous mass flow rates from user defined maps. Each component is represented by a map, which relates mass flow rate, pressure ratio, rotational speed and efficiency. The turbocharger manufacture typically supplies a flow map which has been obtained via steady state flow experimentation. However, these maps are not well suited for use in WAVE simulations due to the sparse data generation, with the manufacturer concentrating on the regions of high efficiency. Off design steady state and transient operating points often lie outside these regions specified by the manufacturer, with problems also occurring due to the varying turbine pressure ratios depending on the CA position. To overcome this difficulty, the code includes a preprocessor, which enhances the data provided by the turbocharger manufacturer and builds up detailed maps suited for numerical simulations. These rearranged compressor and turbine maps have provision for calibration by the operator before the use in simulation. Figure F.1 represents the compressor and turbine maps that have been generated using the WAVE preprocessor TC-map. Copies of the original GT-12 maps supplied by Garrett are displayed in Chapter 6.6.

F.2.5 Turbocharger Wastegate Control Model

Initially, simulation runs were completed without a wastegate model, relying on the restrictor to limit airflow to the turbine. However, a wastegate valve was later implemented and modeled as one cylindrical duct bypassing the turbine. The effective area of this duct is varied at each operating condition until the numerical value matches the experimental boost pressure and turbocharger shaft speed data in the supplied maps.

381

Figure F.1: Garrett GT-12 maps generated using Ricardos TC-map. (Upper): Compressor. (Lower): Turbine.

382

F.2.6 Muffler Model

A coarse grain muffler model was implemented, represented by a duct with an increased diameter and an exit nozzle. The duct geometries were then tuned to predict exhaust back pressure values which were compared to previous pressure loss experimental data [265].

F.2.7 Friction Model

In order for WAVE to compute brake output quantities such as BMEP and brake power from indicated data, engine friction was modelled by a modified form of the Chen-Flynn correlation [186] as outlined in Equation F.5. The correlation has four constants needing user-defined values. The first term represents friction due to auxiliary components such as water pumps and alternators, the second varies with peak cylinder pressure, a third term linearly dependent on mean piston velocity for friction in the hydrodynamic regime and a fourth term with reference to mean piston velocity for windage losses. The values for constants A, B, C, D were taken from the literature describing engine friction in four stroke motorcycle engines [259, 261]. FMEP = A + B Pmax where:

s s + C N + D N 2 2

(F.5)

FMEP = friction mean effective pressure, constants A B C D [259, 261], N = engine rotational speed (rev/min), s = stroke length

383

384

Appendix G
P O W E R E D B Y W A T T A R D

Transmission Ratio Optimization

G.1 Gear Ratio Optimization

The role of the transmission is to provide torque and speed multiplication in order to supply engine power to the driving wheels at any road speed [96, 243]. Generally speaking, typical engine power is non-uniform throughout the operating speed range. By using multiple gear sets, the drivetrains conversion ratios adapt the available torque to the required tractive force to provide forward movement. Optimizing the spacing between and increasing the number of gear ratios optimizes the vehicles fuel consumption or acceleration by changing the engine operating region. An infinite number of gear sets will produce a vehicle which transmits maximum power to the driving wheels over all vehicle speeds. This is the theory behind a continuously varying transmission (CVT), which can vary the transmission ratio to maintain maximum power for any given vehicle speed. Some similarities exist between the CVT concept and the proposed FSAE powertrain described in Chapter 3 (near constant power over a wide speed range coupled with reduced gear sets), as they both can be optimized to deliver near maximum power to the driving wheels over all vehicle speeds subject to traction limitations associated with initial vehicle acceleration.

385

G.2 Final Matched Transmission

Gear ratios to suit the simulated constant power characteristics of the new engine design were tuned to suit FSAE vehicle operation. The required number of gear sets and corresponding ratios were determined using an analytical model, which estimated drive wheel power from the simulated engine performance shown in Chapter 3.3.2. Using this data, combined with typical FSAE wheel diameters and vehicle properties, it was possible to optimize the gear and final drive ratios to maximize the transfer of available engine power for any given vehicle speed. It is also noted that the gear ratios within the crankcase were chosen to minimize the final drive sprocket ratio in order to reduce the drivetrain sprocket diameter. Reducing the differential sprocket diameter significantly improved vehicle packaging, allowing the rear drivetrain to be positioned closer to the engine. Consequently, the selected gear ratios within the crankcase allowed the final drive sprocket diameter to be halved when compared to typical four cylinder engines used in FSAE [212]. Previous FSAE vehicle logged data was also utilized to gather typical vehicle speeds used in competition. Logged data over several years of competition revealed an average vehicle speed of 55 km/h over the endurance event, which typically involved a speed range varying from 30 to 90 km/h. Maximum vehicle speed was also limited to just over 100 km/h by FSAE regulation. The spacing between gear sets was constrained with aims of continuously supplying near maximum power to the drive wheels after initial vehicle acceleration. A three speed transmission was selected, which minimized the mass and complexity of the transmission but also allowed maximum power to be utilized over the majority of the endurance event without the need for a gear change. Alternatively, a gear up-change or down-change enabled near maximum power at higher or lower vehicle speeds (Figure G.1). Table G.1 displays the final selected gear ratios for the test engine, with Figure G.1 also making comparisons to the teams previous Suzuki GSX-R600 powertrain.

386

Table G.1: FSAE trim gear reduction ratios for the Suzuki GSX-R600 and the test engine. Suzuki GSR-R600 (FSAE Trim)
Engine Primary Reduction 1 Gear 2 3 4 5
nd rd th th st

Test Engine
2.652 2.571 1.777 1.125 2.5

1.756 2.866 2.058 1.65 1.428 1.285 1.181 4

Gear Gear Gear Gear

6th Gear Final Drive Reduction

800 700 Drive Torque (Nm) 600 500 400 300 200 100 0 0 20 40 60 80 100 120
1st 2nd 3rd 4th 5th CVT

1st Gear 2nd Gear 3rd Gear 4th Gear 5th Gear 6th Gear

53 kW 6th

Suzuki GSX-R600 - FSAE Trim

140 160

Vehicle Speed (km/hr)

800 700 Drive Torque (Nm) 600 500 400 300 200 100 0 0

CVT

1st Gear 2nd Gear 3rd Gear

1st

3rd

57 kW

UniMelb WATTARD Test Engine

20 40 60 80 100 120 140 160

Vehicle Speed (km/hr)

Figure G.1: Engine vehicle matching through transmission design using the engine performance data from Chapter 3.3.2. (Upper): 6-speed transmission for the Suzuki GSXR600 adapted to FSAE. (Lower): Optimized 3-speed transmission for the test engine.

387

388

Appendix H
P O W E R E D B Y W A T T A R D

Flow Bench Airflow Testing

H.1 Introduction
This appendix describes how flow bench testing was used to conduct steady state intake and exhaust port airflow measurements. Parameters varied during the investigation included bore and cylinder head geometry combinations, with results aiding in setting the engine specifications described in Chapter 3. Furthermore, the experimental flow results were used as data inputs in CFD simulations (Appendix F), which proved a valuable tool during engine optimization as described in Chapter 6. Moreover, the flow bench rig and airflow measurement technique is described, with results from experiments also presented for a variety of test configurations.

389

H.2 Steady versus Pulsating Flow

Engine airflow test rigs are commonly found in the OEM and aftermarket automotive industries as they can provide insight into engine breathing characteristics. However, the intake gas dynamics vary substantially between an engine and a test rig due to the inherently different flow regime. A flow bench is generally configured to simulate steady gas flow, while an engine in firing condition produces pulsating flow in the inlet and exhaust systems due to pipe resonance. However, at particular engine speeds with optimized inlet and exhaust lengths, the WOT performance of the engine tends to be limited by fluid friction, which restricts the flow of gas into and out of the engine [246]. Although there are small differences in the frictional effects of steady and pulsating flow, there is sufficient similarity to use steady flow testing as a guide to performance levels or performance improvement [114, 220]. Evidence of this was provided by Watson et al. who had previously shown that an increase in air flow on the flow bench could be used to accurately predict increases in power output, provided the inlet Mach index (Equation H.1) was below a critical value [242].

Inlet Mach Index =

Average gas velocity in inlet port Sonic velocity

(H.1)

H.3 Super-Flow SF600 Flow Bench

A Super-Flow SF600 flow bench fitted with a Flow-Com controller was used to conduct flow measurements. The SF600 flow bench is able to measure the airflow through a variety of engine components including cylinder heads, intake manifolds, carburetors, throttle systems, velocity stacks and restrictor plates. For intake flow measurements, ambient air is drawn into the intake system by a positive displacement air pump, passing through the orifice plate before exiting through the rear of the flow bench. For exhaust flow measurement, the path is reversed as displayed in the schematic of Figure H.1. The test pressure at the base of the test cylinder is adjustable by altering the flow rate controlled by the positive displacement pump.

390

Velocity Probe

P1+ P2P1 - Flow Pressure P2 -Test Pressure P3 - Velocity Pressure T1 - Lower Temperature T2 - Upper Temperature P2+ P1Flow Meter Flowscale Temp. Probe Dial Indicator and Valve Movement Test Head Cylinder Adaptor Flow Control Valve Test Pressure Monometer Orifice Plate Air Pump

T1

Figure H.1: SF600 flow bench schematic [213].

The air mass flow rate through the system is calculated by measuring the pressure differential across the orifice plate using a U-tube monometer. Flow straighteners are built into the SF600 to avoid disturbance at the flow measurement point, resulting from sharp angular geometry changes within the system. The placement of the orifice plate is also vital to ensure fully developed flow conditions and thus valid results. The SF600 adheres to Australian and British Standards (AS 2360.1.1-1993, BS 1042) regarding orifice plate setup and pressure tapping for measurement [18, 36]. The spacing and geometry requirements of this standard are shown in Figure H.2.
L1 L2 D and D/2 Pressure Tappings

Direction of Flow

DX

L1 = (0.9 - 1.1) D L2 = (0.48 0.52) D

L1 L2

Flange Tappings

Figure H.2: Orifice plate measurement detail for AS 2360.1.1-1993 and BS 1042 standards.

391

H.4 Flow Test Rig and Experiments

Prior to any flow tests being completed, various cylinder adaptor attachments for different bore sizes were manufactured, with the flow rig setup depicted in Figure H.3. Besides attaching the cylinder head to the flow bench, these attachments ensured no airflow leakage between components, with o-ring seals ensuring accurate and repeatable experimental data. The flow rig involved attaching To velocity stacks without manifolds to the intake of various cylinder heads.

minimize flow losses and hence improve discharge coefficients, the velocity stack was radiused with plasticine to minimize turbulent flow entry. Exhaust port tests were completed without radiused exits as this had no effect on measured values. Experiments were completed with a pressure drop of 6 and 10 kPa across the valves, measured at the base of the cylinder as depicted in Figure H.1. Intake and exhaust valve lifts were varied in 0.5 mm increments until maximum flow was achieved. The valve position was measured with a dial indicator placed on the valve tip to ensure valve lift accuracy as shown in Figure H.3.
Valve lift measurement and control

Cylinder head

Adaptor attachment

Flow bench

Radiused inlet port entry

Figure H.3: Experimental flow rig setup.

392

Table H.1: Test conditions and configurations trialed in airflow experiments. Test Pressure
6 and 10 kPa pressure drop across valve

Valve Lift
0.5 mm increments to maximum flow

Configuration
A B C D

Cylinder Head
Suzuki RF600 Suzuki RF600 Kawasaki ER500 Suzuki RF900

Bore Diameter
65 mm 69 mm 74 mm 74 mm

Table H.1 outlines a selection of flow tests for which results are shown (Section H.5) for various cylinder head and bore configurations, with cylinder head selection further discussed in Chapter 3.4.2. Bore sizes were varied (from that dictated by initial OEM cylinder head configurations) to document the valve shrouding and airflow effects for the test engines alternative bore size.

H.5 Flow Test Results

Figures H.4 and H.5 display inlet and exhaust airflow data gathered from experimental flow bench tests, for various configurations outlined in Table H.1. To enable fair comparisons between configurations, data is displayed versus the non-dimensional valve lift and valve diameter ratio (L/D). Discharge coefficients are calculated as in Equation H.2 [246], with no flow losses corresponding to a unity discharge coefficient.
Measured Mass Flow Maximum Choked Mass Flow

Discharge Coefficien t (C D ) =

(H.2)

Figure H.6 displays engine discharge coefficients, calculated by combining the valve airflow data with the inlet and exhaust camshaft profiles throughout the cycle. Hence, average Mach indexes are also calculated and displayed in Figure H.6. Intake restrictor flow bench results are also displayed in Figure H.7 [78, 245], with the geometry of both devices tested given in Table H.2. Restrictor device B was designed by the author due to packaging constraints and hence used in all experiments.

393

H.5.1 Cylinder Head Mass Flow

0.10

0.08 Inlet Mass Flow (kg/s)

0.06 (A) RF600 - 65 mm bore (6 kPa) 0.04 (A) RF600 - 65 mm bore (10 kPa) (B) RF600 - 69 mm bore ( 6 kPa) 0.02 (B) RF600 - 69 mm bore (10 kPa) (C) ER500 - 74 mm bore (6 kPa) (C) ER500 - 74 mm bore (10 kPa) 0.00 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Lift / Diameter (L/D)

0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00 0.00

Exhaust Mass Flow (kg/s)

(A) RF600 - 65 mm bore (6 kPa) (A) RF600 - 65 mm bore (10 kPa) (B) RF600 - 69 mm bore (6 kPa (B) RF600 - 69 mm bore (10 kPa) (C) ER500 - 74 mm bore (6 kPa) (C) ER500 - 74 mm bore (10 kPa) 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Lift / Diameter (L/D)

Figure H.4: Port experimental mass flow values at various test pressures. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Inlet. (Lower): Exhaust.

394

H.5.2 Cylinder Head Discharge Coefficients

0.6

0.5 Inlet Discharge Coefficient ( (Cv) CD)

0.4

0.3 (A) RF600 - 65 mm bore (6 kPa) (A) RF600 - 65 mm bore (10 kPa) 0.2 (B) RF600 - 69 mm bore (6 kPa) (B) RF600 - 69 mm bore (10 kPa) 0.1 (C) ER500 - 74 mm bore (6 kPa) (C) ER500 - 74 mm bore (10 kPa)

0.0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Lift / Diameter (L/D)

0.6

0.5 Exhaust Discharge Coefficient (Cv) (CD)

0.4

0.3 (A) RF600 - 65 mm bore (6 kPa) 0.2 (A) RF600 - 65 mm bore (10 kPa) (B) RF600 - 69 mm bore (6 kPa) (B) RF600 - 69 mm bore (10 kPa) 0.1 (C) ER500 - 74 mm bore (6 kPa) (C) ER500 - 74 mm bore (10 kPa) 0.0 0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Lift / Diameter (L/D)

Figure H.5: Port discharge coefficients at various test pressures. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Inlet. (Lower): Exhaust.

395

H.5.3 Engine Discharge Coefficients

0.8 0.7 Discharge Coefficient (CD) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -360

(A) RF600 - 65 mm bore (B) RF600 - 69 mm bore (C) ER500 - 74 mm bore

CAVER = 0.40 CAVER = 0.39 CAVER = 0.36 CAVER = 0.33 CAVER = 0.38 CAVER = 0.35

EXHAUST

INTAKE

-240

-120

0 Crank Angle (deg)

120

240

360

0.8 0.7 0.6

(A) RF600 - 65 mm bore (B) RF600 - 69 mm bore (C) ER500 - 74 mm bore

CAVER = 0.39 CAVER = 0.37 CAVER = 0.37 CAVER = 0.35 CAVER = 0.35 CAVER = 0.32

Discharge Coefficient(CD)

0.5 0.4 0.3 0.2 0.1 0.0 -360

EXHAUST

INTAKE

-240

-120

0 Crank Angle (deg)

120

240

360

Figure H.6: Engine discharge coefficients versus CA. Configurations A, B, and C correspond to various cylinder head and bore combinations shown in Table H.1. (Upper): Test pressure = 6 kPa. (Lower): Test pressure = 10 kPa.

396

H.5.4 Intake Restrictor Nozzle Mass Flow

0.08

0.07 A (50-20-12-40) Mass Flow Rate (kg/s) 0.06 B (35-36-14-42) A (No throttle body) 0.05 B (No throttle body)

0.04

0.03

0.02 0 4 8 12 16 20 24 28 32 Pressure Drop (kPa)

1.0

0.9 Discharge Coefficient (CD) A (50-20-12-40) 0.8 B (35-36-14-42)

0.7

0.6

0.5

0.4 0 4 8 12 16 20 24 28 32 Pressure Drop (kPa)

Figure H.7: Intake restrictor experimental flow bench airflow data [78, 245]. Device A and B correspond to various Dall restrictor nozzle and diffuser geometry outlined in Table H.2. (Upper): Mass flow rate. (Lower): Discharge coefficient.

397

Table H.2: Dall restrictor geometry outlined in Figure H.7. Geometry

Nozzle Device A B Diameter (mm) 50 35 Included Angle (deg) 20 36 Diffuser Included Angle (deg) 12 14 Diameter (mm) 40 42 Total Length (mm) 182 126

398

Appendix I
P O W E R E D B Y W A T T A R D

Engine Subsystem and Transmission Development

I.1 Introduction
This appendix provides further description surrounding aspects of engine development that were required to ensure reliability while the test engine was installed on the dynamometer and into successive FSAE vehicles. The additional detail supplied in this appendix follows previous developments described in Chapter 5. Information concerning transmission and engine subsystems, including the lubrication and water cooling systems, is supplied in this appendix.

399

I.2 Clutch Assembly

Engine torque transfer to the transmission was completed with a multi-plate wet clutch, with components outlined in the drawings of Appendix P. A wet clutch system differs from conventional dry systems as the assembly is partially immersed in lubricating fluid (engine oil), which limits the production of heat and keeps the mating surfaces clean. However, since the lubricating oil can often reduce friction, several clutch plates need to be stacked into a sandwich consisting of multiple layers of friction plates, driving discs and a single pressure plate to avoid slippage [96]. The clutch system used in the test engine was adapted from a similar capacity OEM motorcycle [117], which worked faultlessly throughout wet sump NA dynamometer and vehicle testing. However, when the dry sump was implemented due to oil surge reasons (Section I.4), the reduced oil in the crankcase caused clutch plate overheating. The overheating resulted in glazing of the friction plate material, which led to clutch slippage during dynamometer development. For NA testing, this was rectified by adding a central oil feed which flowed oil through the gearbox input shaft into the clutch housing. In addition, drillings into the centre hub allowed the oil flow to reach the clutch plates, increasing cooling (Figure I.3). These drillings were strategically located in areas where oil pooling would occur due to the centrifugal force of the rotating clutch assembly. Following oil cooling, there were no further problems encountered during dry sump NA testing. Once boosted testing commenced, the twofold increase in torque levels necessitated an improved clutch design. Due to geometric constraints and the integral clutch design within the crankcase, the implementation of a larger system was not feasible. Hence, development was centered on improving the torque capacity within the limited space. Initially, development focused on increasing the pressure plate spring force on the clutch pack in order to prevent slippage, through the use of stiffer springs (1.5 and 2 times stiffer than the OEM springs). However, a downside of the stiffer springs was the increased clutch cable force required for engagement during gear selection.

400

MPa 0 15

STRESS
m 30 45 60 0

DISPLACEMENT
30 60 90 120

OEM Pressure Plate

OEM spring load = k1

k1 = 25 N/mm

OEM Pressure Plate

Spring load = 2 k1

k2 = 50 N/mm

Designed Pressure Plate

(Never manufactured) Spring load = 2 k1

k2 = 50 N/mm

Figure I.1: FEM model analysis simulating varying spring loads applied to various pressure plate designs after clutch slippage due to turbocharging. Both pressure plate designs have similar mass and rotational inertia. (Left): Von Mises stress distribution. (Right): Resultant displacement.

401

Pressure plate spring development started with the 1.5 fold stiffer springs, with experimental results showing similar magnitude torque improvements prior to slippage. However, when the twofold stiffer springs were implemented, the clutch slipped at lower torque levels when compared to the OEM arrangement. FEM analysis (Figure I.1) revealed that the decreased torque capacity was a consequence of excessive spring tension on the pressure plate, which caused higher stresses and hence greater pressure plate deflection. The increased deflection reduced the pressure plate contact area resulting in lower torque levels before slippage occurred. FEM analysis also highlighted that the higher stresses in the centre of the pressure plate during clutch engagement, due to the twofold stiffer springs, could possibly cause failure. To resolve this, a new pressure plate was designed with equivalent mass, which gave similar stresses and deflections (Figure I.1 - Lower) as the OEM clutch, but was capable of withstanding the twofold stiffer springs. However, this design was never manufactured as the following changes were made to the wet clutch through a series of development steps, which enabled the system to withstand the boosted conditions. The developments taken were as follows:
The clutch plate material was changed to Kevlar. The OEM clutch pressure plate spring stiffness was increased 1.5 fold. The clutch basket was reinforced with a circumferential stiffening ring

(Figure I.3) to minimize radial deflection and resist the twofold increase in torque experienced due to turbocharging. FEM analysis in Figure I.2 displays the effects, with basket deflection reduced 6 fold.
The number of friction plates was increased from 7 to 8. This was made

possible as the clutch basket was extended by the circumferential ring.

The centre oil flow to the clutch was increased through a relief made in the

operating rod. Care needed to be taken to balance the oil pressures as the engine oil pump capacity was limited.
The turbocharger oil drain was relocated over the top of the clutch to aid

in cooling.

402

MPa 0 30

STRESS
m 60 90 120 0

DISPLACEMENT

30

60

90

120

Original Clutch Basket

(At peak TC torque)

Developed Clutch Basket

(At peak TC torque)

Figure I.2: FEM model analysis simulating the clutch basket at peak torque due to turbocharging. (Left): Von Mises stress distribution. (Right): Resultant displacement. (Top): Original basket. (Bottom): Developed version with circumferential stiffening ring.

The final version of the developed clutch is displayed in Figure I.3. This version enables a twofold increase in the torque capacity of a modern clutch with only a marginal increase in size, which did not affect operation within the crankcase. The newly designed clutch has packaging and engine mass reduction advantages applicable to other powertrains. Development has shown clutch limitations are due to the stiffness and deflection of the clutch components, with stiffer components allowing more load and hence torque to be applied.

403

Circumferential stiffening ring, also allowing a greater number of clutch plates to be stacked in the sandwich

Centre oil cooling

Oil drillings to cool clutch plates

Figure I.3: Developed clutch housing enabling a twofold increase in torque transmission in a confined space which was needed for boosted operation.

I.3 Transmission
Transmission development was required to ensure reliability for the new wide ratio three speed transmission. The optimized gear ratios are outlined in Appendix G.2. Transmission ratio optimization ensured compatibility with engine power characteristics, allowing vehicle gearshifts to be significantly reduced as near constant power could be maintained over a wide speed range [48, 241]. There were no problems recorded with the initial transmission design, which featured straight cut spur gears with dog interlocking drives. These were actuated by a rotating drum assembly as shown in the exploded views of Appendix P. The dog drives were initially designed with geometry perpendicular to the gear face as shown in Figure I.4. This design was successful for over 18 months of testing, however following dynamometer seizure and transmission destruction (Appendix J.3), problems associated with torque transmission occurred. Under heavy loading (boosted operation), the newly manufactured perpendicular dog drive would be thrown out as the two gears would separate. Close inspection of the initial failed components indicated that as the drive dogs bedded in under lighter NA load conditions, they developed a slight inverse taper angle. Hence, when larger loads

404

were applied with boost, the inverse taper would axially force the dog drive and gear together, locking the two components. further problems recorded. These findings indicate that perpendicular drive dogs, widely found in motorcycles, are adequate for torque transmission in NA powerplants which feature an integral engine and constant mesh transmission in a common crankcase. However, this type of powertrain rarely operates under boosted conditions. At the higher loads, experiments found that inverse taper was required on the drive dogs to ensure drive engagement under the high torque transmission. Although the inverse taper improves the drive engagement, the higher axial force drives the two components together causing more difficult gear shifts, thus requiring an occasional lift off throttle for gear disengagement. Consequently, the newly manufactured drive dogs were ground with a 1 inverse taper (Figure I.4) with no

90 Drive dog x 3

89

Isometric View

Original Design
(Perpendicular)

Final Design
(Inverse taper)

Figure I.4: Transmission drive dog development needed for the high torque output associated with intake boosting.

I.4 Lubrication System

The original engine design featured a wet sump lubrication system in order to minimize the parasitic losses associated with driving auxiliary systems such as a dry sump scavenge system. Minimizing all parasitic losses was important as this improved brake power [14]. Baffling with scraper plates was assumed to be sufficient in preventing oil surge for the wet sump system when the engine was

405

installed into a FSAE vehicle, as the vehicle would encounter lateral acceleration forces up to 1.5 g (acceleration due to gravity). Further detail behind the wet sump system is given in the exploded views of Appendix P. On the vehicles first outing, initial baffling attempts proved to be unsuccessful with the engine seizing after several minutes of operation. Oil surge had resulted in starving both big-end plain bearings of oil, causing cylinder two bearing failure and subsequent irreparable connecting rod and crankshaft damage, as depicted in Figure I.5. Besides the heat affected zone shown in Figure I.5, crankshaft damage included bending which resulted in severe values of rotational axis runout, with the crankpin also cracking along the inner radius. The connecting rod big-end diameter also elongated, with the high bending loads causing the beam to plastically deform at the region previously highlighted as having the highest stresses (Figure 3.8). The failed components highlight the validity of the calculations and FEM analysis performed during the design stage (Chapter 3.5), with seizure occurring before any internal components were expelled from the crankcase.

Cylinder 2 bearing failure and subsequent crankpin damage

Predicated connecting rod beam failure from the FEM analysis (Figure 3.8)

Figure I.5: Irreparable crankshaft and connecting rod damage as a result of oil surge under high lateral vehicle acceleration, leading to big-end bearing failure.

406

The oil surge problems in the wet sump system were largely associated with the sump pan design. This was flat to allow the engine to be positioned as low as possible in order to reduce the vehicles COG height and thus improve dynamic performance [14, 151]. Development continued around the wet sump system as there was insufficient time to incorporate a dry sump configuration before FSAE competition. Engine development centered on improving the baffling to reduce the oil flowing into the engines clutch and generator covers under lateral loading. Trial and error methodology was used to improve the sump baffling in order to reduce oil surge to acceptable limits. Vehicle tests at sustained lateral loading indicated that insufficient baffling caused sudden oil surge, with the oil pressure recovering as the driver reacted to the warning light. As the baffling effectiveness improved pickup oil retention, increased periods of sustained lateral acceleration could be maintained before surge. However with the improved baffling, oil pressure would not instantly recover as the baffling also restricted oil flow back into the sump. The oil surge problems were eventually reduced to acceptable limits with a combination of improved baffling and an increase in oil capacity, which was achieved by overfilling the sump. The final wet sump design iteration showed that oil pressure could be maintained for 10 seconds at a sustained lateral acceleration of 1.5 g. However, oil pressure would take up to three seconds to recover once surge occurred. This was adequate but not ideal for competition due to the higher parasitic losses and extra mass associated with the increased oil capacity. However, the wet sump solution proved successful, with no oil surge problems at FSAE competition. The following year saw the further developed engine design that featured a dry sump system as depicted in the exploded views of Appendix P. The engines prototype nature made it difficult to implement an aftermarket system into the design and thus a new system including scavenge pumps was designed and manufactured. The new system incorporated the oil and scavenge pumps on a common shaft, which was driven internally. Thus the inlet and outlet oil lines were routed through the crankcase to the external oil tank. The tank featured a swirl inducing inlet port [14, 157, 206] to enhance de-airation as no centrifuge was incorporated in the design, with baffling also implemented inside the tank to limit the oil moving up the tank walls during high g vehicle operation.

407

A scavenge ratio sizing of 1.5 (1.5 times the capacity of the pressure pump) was chosen for the scavenge pump [14]. This capacity was dictated by space However, provision had been limitations and from calculations thought to be marginal as the scavenge pump was required to evacuate both oil and blow-by. made in the design to increase the systems shaft speed if the design proved inadequate. The increased shaft speed would also cause oil flow increases and thus an external oil pressure relief valve would need to be implemented. The pump bodies were manufactured in-house with OEM sourced G-rotor internals. An internally contoured CNC machined sump pan encouraged oil flow to the scavenge pickup. This also allowed the engine to be positioned lower in the vehicle. Commissioning the new dry sump oil system on the dynamometer indicated that the oil tank level could not be maintained at high speeds or cold conditions. This was due to the internal oil pressure relief valve, which relieved the excess oil into the crankcase where there was insufficient engine scavenge to evacuate the oil back to the tank. This problem was overcome by designing and incorporating an external oil pressure relief valve which returned excess oil directly to the oil tank under its own pressure. An external oil relief valve also enabled the oil pressure to be easily adjusted. A schematic of the oil flow path is given in Figure I.6, with more drawing detail also given in Appendix P.

OIL GALLERIES PRESSURE RELIEF VALVE OIL PUMP SCAVENGE PUMP OIL VENTED OIL TANK VENTED CRANKCASE

OIL FILTER

OIL

Figure I.6: Dry sump oil path schematic, featuring internal oil and scavenge pumps with an external oil pressure relief valve.

408

It was advantageous to seal the crankcase and allow the scavenge pump to pull a partial vacuum in order to reduce windage and pumping losses and therefore increase delivered power [14, 94]. However, the power gains due to the reduced losses were offset by the increased parasitic power needed to create the vacuum. Furthermore, problems existed in maintaining crankcase vacuum under all operating conditions with the revised dry sump system. At some operating conditions, particularly at low load, the crankcase would pressurize causing gasket failures and hence oil to be expelled. This was caused by the high blow-by rates associated with insufficient gas loading on the piston ring at low MAP. These problems were magnified by the first iteration piston design, which featured a single compression ring (Chapter 3.5.3). The problems associated with crankcase pressurization were temporarily solved by venting the crankcase to the atmosphere in order to continue development. Figure I.7 displays the BMEP in NA operation for both the wet and dry sump systems. Engine specifications, condition, fuel and ignition settings were held constant over both runs. The effect of driving the dry sump scavenge pump can be seen with reduced BMEP over all speeds. Power losses are calculated via the differing performance, with the dry sump system absorbing more power as the engine speed increased.

1300 1200

4.0

3.0 BMEP (kPa) 1100 1000 900 800 700 4000 BMEP - W et Sump BMEP - Dry Sump Power Loss - Dry Sump
R = 0.93

2.0

1.0

5000

6000

7000

8000

9000

0.0 10000

Engine Speed (rev/min)

Figure I.7: Brake performance effects for both wet and dry sump lubrication systems.

409

Brake Power Loss (kW)

This performance lost was originally thought to be partially recovered by evacuating the crankcase as previously described. However, with insufficient scavenge due to the inadequate dry sump system, an externally driven scavenge pump was used to draw a minimum of 50 kPa crankcase absolute pressure (CAP) over all speeds at WOT. Back to back constant speed dynamometer testing showed no change in brake performance with varying CAP, with results lying within the bounds of the 1% measurement error. Because no clear performance gains were observed with the crankcase under vacuum, it remained vented for dry sump operation, performing successfully at the 2004 FSAE competition. However, dynamometer transient testing may have yielded differing results, as engine acceleration rates are expected to be improved under vacuum conditions. This is due to the reduced energy required to accelerate reciprocating and rotating components in a partial vacuum [14].

I.5 Water Cooling System

Initial NA dynamometer developments revealed no engine water cooling problems with the existing dynamometer setup which included an open water to air cooling tower as described Chapter 4.2.4. As the engine design did not feature an integral mechanical water pump due to the added complexity, a lightweight electric version manufactured by Davies Craig (Electric Booster Pump - EBP) [59], was incorporated into the system [210]. As the GT-12 turbocharger featured centre bearing water cooling, the engine water cooling circuit was re-routed to accommodate the turbocharger and improve its reliability (Appendix 6.4.2). However, significant TC engine reliability issues arose from the onset of initial engine run-in and calibration at light load operating conditions where the engine had previously operated successfully in NA modes. This involved high blow-by rates and knock/cooling concerns which led to piston reliability problems. At the time of initial turbocharger development, the precise cause of the piston reliability problems could not be identified. Consequently, turbocharger development ceased momentarily with the engine reverting back to the NA - PFI mode for fitment into the FSAE vehicle for competition. However, further NA in-

410

vehicle developments at competition led to similar piston reliability problems. At the time of competition, the precise cause of the poor engine reliability could not be determined, but was related to the water cooling circuit as temperatures exceeded 130C. These cooling problems led to the re-inspection of the TC components after competition. Re-inspection revealed significant upper cylinder bore wear around the periphery of both liners after only several hours of light load operation. The high wear rates were caused by the high thermal loading and hence bore distortion due to inadequate cooling, which resulted in bore wear caused by the piston ring motion. These high wear rates had previously not been present in NA dynamometer operation. Further engine cooling bench testing revealed that the engine reliability problems experienced in the TC mode were a result of compromising the water flow path through the engine. This was caused by inducing the turbocharger cooling water prior to entering the engine and expelling after the engine output, with the water cooling circuitry displayed in Figure I.8. Consequently, the majority of water flow bypassed the engine and flowed through the turbocharger. Bench testing revealed that this problem was further magnified by the use of a marginal capacity electric water pump. Further vehicle bench tests also revealed that the vehicle related cooling problems were also related to inadequate water flow, with the cooling radiator placing higher flow demand on the marginal electric water pump, which was not evident in dynamometer cooling tower testing. This was due to the increased back pressure in the cooling system as a result of the ~12 kPa pressure drop across the radiator [124], which resulted in halving the cooling system water flow. Consequently, the dynamometer cooling system was altered to be more representative of the vehicle (Chapter 4.2.4) and a larger flow capacity electric water pump was incorporated into the system.

Turbocharger Heat dissipation (radiator, cooling tower)

Electric water pump

Engine

Figure I.8: Initial inadequate turbocharger water cooling circuit, resulting in the majority of flow bypassing the engine.

411

412

Appendix J
P O W E R E D B Y W A T T A R D

Test Rig Development and Calibration

J.1 Introduction
In this appendix, further description regarding the experimental test rig, described in Chapter 4 is given. In particular, the initial setup and development of the subsystems of the rig is discussed. The development described focuses on the testing requirements due to the engine design which features high speed operation, an odd firing interval and an integral clutch and transmission. These features result in coupling, mounting and torque absorption difficulties, which required special remedies. and instrumentation. Furthermore, the variety of fuel and lubricating systems used between the test modes increased the complexity in measurement

413

J.2 Brake Torque Measurement and Calibration

Torque measurement was via a strain gauge type load cell attached to the test bed at a known distance from the dynamometer pivot axis. Calibration of the Interface-SSM500 load cell is shown in Figure J.1 together with a free body diagram outlining the methodology behind the beam calibration. To relieve any bearing stiction in the dynamometer pivot axis, the test bed plate was tapped during the calibration. Hence, static torques were computed from the beam calibration using Equation J.1 and compared to the dynamometer displayed value. Comparisons allowed torque correction, as displayed in Figure J.2.
Dynamometer (Diameter D) Removable arm (Mass B, Length L) Dynamometer pivot axis

Removable weights (Mass A)

Force X

Force Y

1 L 1 Applied Torque = X L + D + Y + D 2 2 2 1 1 1 = A 9.8 L + D + B 9.8 L + D 2 2 2

(J.1)

Removal beam

Dynamometer

Strain gauge load cell Incremental masses Torque damper

Figure J.1: Dynamometer torque calibration. (Upper): Methodology and calculation. (Lower): Actual load cell calibration.

414

300 250 Applied Torque (Nm) 200 150 100 50 0 0 50 100 150 200 250 Dynamometer Displayed Torque (Nm) y = 1.083x R2 = 1.000

Figure J.2: Brake torque calibration, beam calibration versus dynamometer displayed.

Safety stops, which limited dynamometer rotation on its trunnion bearings was implemented to ensure load cell reliably and longevity in the case of sudden dynamometer or engine seizure. Consequently, the safety stops limited load cell displacement, which ensured strain limits were not exceeded due to shock loading (eg. engine seizure - Section J.3). As result of the odd firing interval, the unequal torque pulsations experienced by the load cell were dampened by the high rotational inertia of the dynamometer field rotor. Mounting the load cell to an elastomer dampening cushion (Figure J.1) also aided in smoothing the torque pulsations.

J.3 Dynamometer-Engine Coupling Development and Calibration

An existing dynamometer-engine coupling [187], featuring a direct coupled shaft was retro-fitted to the transmission output shaft of the test engine. The shaft was linked to the dynamometer via a PolyFlex torsional damper [177], and connected to the engine transmission via a 30 pressure angle splined joint. Further detail together with a sectional view of the direct coupled shaft is given in Figures J.3 and J.5. However, following several hundred hours of engine testing, a dynamometer failure was experienced. It was found that the dynamometer input and engine output shafts exceeded the 0.12 mm allowable axis misalignment limits. It was also discovered that the shaft misalignment had increased during

415

engine operation due to the rubber mounting of the engine to the cradle (Chapter 5.7). The rubber mounting allowed small engine movements, however this placed high loads on shaft bearings, eventually resulting in simultaneous dynamometer bearing and engine failure caused by the rapid dynamometer deceleration. The bearing failure initiated further development to find a reliable technique to couple the transmission shaft of the flexible mounted engine to the fixed dynamometer. A new coupling was designed consisting of a chain drive, accommodating engine misalignment and vibrational movement. Moreover, the chain drive provided a weak failing link to prevent engine failure if dynamometer seizure reoccurred. The first iteration chain drive (Figure J.3 Upper Right), attached sprockets to the engine transmission and the existing shaft to avoid the manufacture of new components. To accommodate the chain drive, the engine cradle was repositioned to achieve the required shaft offset needed for operation. Chain and sprocket selection was based on availability and load requirements from calculations [159]. A 530-5/8 pitch Renold Synergy roller chain with a minimum tensile strength of 22 kN was selected [184]. The Renold chain featured a high safety factor, however load ratings were not excessive to ensure joining link failure upon reoccurring dynamometer seizure. In the first design iteration, the torsional damper was removed to accommodate the bending moment created by the chain load. It was assumed that the chain would act as a sufficient damper, however this proved inadequate with the load cell experiencing high torque pulsations. Consequently, a second iteration chain drive was implemented with a torsional damper (Figure J.3 Lower Left). This was achieved by designing and manufacturing a new shaft, which incorporated a self-aligning pillow block bearing on the end of the shaft to cancel the bending moment and hence transfer the chain load to the engine cradle. A pre-loaded chain tensioner was later incorporated into the design (Figure J.3 Lower Right) to accommodate chain wear and stretch over time, with no further problems encountered. Brake torque measurement correction, correlating chain drive to direct drive, is given in Figure J.4.

416

Figure J.3: Dynamometer coupling development. (Upper Left): Direct drive with torsional damping. (Upper Right): 1st iteration - Counter levered shaft with chain drive. (Lower Left): 2nd iteration - Chain drive with torsional dampening. (Lower Right): 2nd iteration with chain tensioner.
80 70 Direct Drive Torque (Nm) 60 50 40 30 20 10 0 0 10 20 30 40 50 60 70 Chain Drive Torque (Nm) y = 1.055x R2 = 0.995

Figure J.4: Brake torque measurement correction to compensate for the increased losses associated with the chain drive when compared to the direct drive system.

417

(Original) Direct coupling with torsional dampening

(1st Iteration) Chain drive

(2nd Iteration) Chain drive with torsional dampening

1 2 3 4 5 6 7 8

Coupling Component List Dynamometer shaft 9 - Dynamometer flange Retaining nut 10 - Flat washer Torsional damper 11 - Sprocket Spring washer 12 - Shoulder bolt Bolt 13 - Washer Coupling shaft 14 - Splined drive Spigot 15 - Bearing housing Bolt

Figure J.5: Dynamometer coupling development (Upper): Direct coupling with torsional damping. (Middle): Chain drive. (Lower): Chain drive with torsional dampening.

418

J.4 Fuel Measurement and Calibration

The varying induction modes and consequent carburetion and PFI fuelling requirements necessitated different fuel delivery systems to quantify fuel consumption rates. The carbureted system required a simple gravity feed from the fuel supply to the float bowl chamber with no fuel return. The fuel lines were kept as short as possible to minimize the likelihood of fuel expansion and vapor production. Hence, consumption was measured via direct mass measurement of the fuel tank reduction. Problems did exist with direct mass measurement of the fuel tank due to its large size which was needed to store the fuel consumed at high speed and load conditions. Thus, fine mass measurement resolution was unachievable using simple methods. The PFI system, however, increased the complexity of the measurement due to the recirculating fuel system with a gasoline return to the fuel tank. This layout would necessitate two measurements (supply and return) if in-line flow meters were used, thus reducing the accuracy of the measurement as the errors from both instruments could compound. Having a common fuel consumption measurement system for all modes is desirable for direct back to back comparison. This was achieved by gravimetric measurement, using a beam balance mechanism to cancel out the high tank mass via a counterbalance weight (Chapter 4.3.1). Injector calibration also allowed fuel flow rates to be cross checked for accuracy. Calibration involved performing the Motec ECU Injector Test, which fired the injector at a user defined pulse width and corresponding engine speed, allowing the mass to be measured for a given time. It is important to note that fuel mass flow rate per cycle may be determined from injector pulse width and corresponding engine speed or duty cycle alone. Figure J.6 displays the calibration curve for the Bosch injector. Injector calibration also ensured individual fuel injectors could be matched and paired, thus minimizing the fuel flow rate discrepancies between cylinders. This was important due to ECU limitations, allowing only a fixed individual cylinder trim for all MAP values at user defined speeds. Fuel distributions between injectors were quantified by repeatedly pulsing the injector over several cycles and visually observing the fuel spray pattern on sheets of paper, similar in principle to a mechanical patternator [238].

419

60 Injector Flow Rate (mg/pulse) 50 40 30 20 10 0 0 2 4 6 8 10 12 14 Injector Pulse Width (ms)

y = 4.363x - 3.846 R2 = 0.999

60 Injector Flow Rate (mg/pulse) 50 40 30 20 10 0 0 10 20 30 Injector Duty Cycle (%) 40 50 60

y = 1.035x - 3.143 0.998 R2 = 0.999

Figure J.6: Injector calibration using gasoline, Bosch 0-280-156-124. (Upper): Injector pulse width. (Lower): Duty cycle.

J.5 Blow-by Measurement and Calibration

Blow-by measurement was conducted volumetrically using a Fisher and Porter Pty Ltd (FP-1-20-G-8/78) rotameter. The unit required modification from its intended application, thus requiring recalibration at atmospheric conditions as shown in Figure J.7. The wet and dry sump lubrication systems required different piping arrangements in order to route the blow-by gas through the rotameter, before expelling the gas to atmosphere. This was trivial in wet sump operation, as all external crankcase vents were sealed, with the blow-by gas vented via the engine breather through

420

the rotameter. In dry sump operation, quantifying blow-by rates was inherently more complicated due to difficulties associated with maintaining scavenge pressure over all test conditions (Appendix I.4) [14]. At some test conditions, the crankcase would pressurize expelling blow-by gas through the rotameter. However, at all other conditions, the scavenge pump would draw atmospheric air into the crankcase, which it later vented with blow-by through the oil tank. This was overcome by implementing a positive crankcase ventilation (PCV) valve on the outlet of the crankcase breather, which acted as a one-way check valve. This ensured no atmospheric air could to be drawn into the crankcase by the scavenge pump, but permitted any blow-by gas to be expelled if the crankcase pressurized. This allowed both the crankcase breather and the oil tank to be vented to the rotameter.
50 40 Blow-by (l/min) 30 20 10 0 0 10 20 30 40 Rotameter Flow (%) 50 60 70 80 y = 0.582x R2 = 0.997

Figure J.7: Volumetric calibration of the modified Fisher and Porter FP-1-20-G-8/78 rotameter used to determine blow-by rates (T = 20C, P = 101.3 kPa).

J.6 Cooling System Calibration

Cooling system volume flow rates were determined using a Fisher and Porter Pty Ltd (FP-1-35-G-10/83) rotameter. The unit required modification for its intended application, thus requiring recalibration at atmospheric conditions as shown in Figure J.8.

421

40 Cooling System Flow Rate (l/min)

30 y = 0.407x R2 = 0.999 20

10

0 0 10 20 30 40 50 60 70 80 90 Rotameter Flow (%)

Figure J.8: Volumetric calibration of the modified Fisher and Porter FP-1-35-G-10/83 rotameter used to determine cooling system flow rates (T = 20C, P = 101.3 kPa).

J.7 Cylinder Pressure Measurement and Calibration

The Kistler 601-B1 piezoelectric pressure transducer was calibrated statically at regular intervals using a dead weight tester, with no static drift changes observed over the testing period. Figure J.9 [60, 131, 133]. Packaging limitations in mounting the transducer in the cylinder head dictated the use of a compact air cooled unit. The possibility of using a cooling housing with external water supply to regulate transducer body temperatures was also eliminated due to space limitations. Hence, to ensure reliability and longevity, the transducer was positioned in close proximity to the cylinder head water-cooling jacket, which allowed the engine cooling system to regulate transducer body temperatures. Due to access constraints, the pressure transducer was initially non-flush mounted, with an orifice connecting the diaphragm to the combustion chamber. This mounting configuration resulted in signal oscillations as shown in Figures J.10 and J.11, attributed to pipe resonance in the connecting bore [129]. These oscillations were originally thought to be caused by knock due to the associated problems and failures already experienced. However, significantly Two different charge amplifier gains and sensitivities were used for both high and low pressure signals, with the calibration shown in

422

increasing the fuel quality with toluene blends (Appendix C.1) together with installing a new pressure transducer did not reduce the frequency or magnitude of these pressure oscillations (Figure J.11). The oscillation problem was eventually eliminated by flush mounting the transducer.
60 50 Pressure (bar) 40 30 20 10 0 0 1 2 Volts 3 4 Range: 200 units/V, Sensitivity: 1.8 pc/unit Range: 200 units/V, Sensitivity: 1.4 pc/unit Range: 100 units/V, Sensitivity: 1.8 pc/unit y = 10.812x R2 = 1.000

y = 21.832x R2 = 0.999

y = 17.461x R2 = 1.000

Figure J.9: Pressure transducer calibration (Kistler 601-B1, SN: C65704).

Non-flush mounting with connecting bore

Flush mounting in combustion chamber

Figure J.10: Mounting comparisons for cylinder pressure transducers in combustion chambers and resulting pressure data quality effects. (Upper): Non-flush mounting with connecting bore. (Lower): Flush mounting [129].

423

7 6
Raw Pressure (MPa)

98 RON gasoline [Sensor A] 112 RON gasoline-tolulene [Sensor A] 112 RON gasoline-tolulene [Sensor B]

5 4 3 2 1 0 -180

-150

-120

-90

-60

-30

30

60

90

120

150

180

Crank Angle (deg)

Figure J.11: Experiment in-cylinder pressure oscillation effects due to the non-flush mounted transducer. Single cycle oscillations shown to be independent of fuel quality and transducer type. SC - PFI, 6000 rev/min, 150 kPa MAP, CR = 11, 15 BTDC spark timing.

Flush mounting was achieved with the aid of a National Instruments PCB-401A series connector. This series connector extended the length of the pressure transducer assembly, allowing direct access into the combustion chamber as shown in Figure J.12. However, as the National Instruments series connector was not designed to operate in conjunction with Kistler transducers, some modification of the pressure transducer assembly was necessary to ensure sealing and tightening integrity. Thermal shock was encountered during testing due to the excessively high transducer temperatures following flame arrival. This resulted in reduced charge output, characterized by pressure cycle inconsistencies in the pumping loop of a log-Pressure versus log-Volume graph [38, 133]. A high temperature RTV silicone compound was applied to the transducer diaphragm to minimize this phenomenon. This RTV coating has been proven to minimize thermal shock, with uncoated sensors recording errors of up to 5% in gross IMEP and 70% in PMEP [181]. Due to the excessively high BMEP and resultant heat loads, the RTV thermal barrier was constantly monitored to minimize thermal drift and shock effects to ensure the in-cylinder pressure data quality. During the entire testing program, the RTV layer was only reapplied on few occasions, with curing the RTV at elevated temperatures prior to engine installation vital to its longevity.

424

Flush mounted transducer

Radiused edges

Combustion sealing ring

Spacer

PCB-401A series connector

Intake Ports

Spark plug hole

Exhaust Ports

Kistler 601-B1 pressure transducer

Tightening nut

Figure J.12: Pressure transducer setup and mounting. (Left): Combustion chamber and pressure transducer location. (Right): Pressure transducer and associated components allowing direct flush mounting into the chamber.

J.8 TDC Alignment

In order to accurately phase the cylinder pressure and spark timing to the crankshaft position, TDC location was found using two methods. The first method involved measuring the piston movement using a dial indicator positioned on the piston crown. pin. This was achieved by rotating the crankshaft clockwise and counter-clockwise to minimize the piston slap effects in the non-offset gudgeon The midpoint between the two readings located TDC, allowing alignment with a TDC marker on the flywheel [133]. The estimated accuracy of the nondynamic measurement is expected to be within 0.1 CA, with results affected by piston to bore clearance and gudgeon pin offset values. Dynamic engine effects associated with reciprocating component deflections and rotating twist angles also results in distorting the TDC position relative to the crankshaft, as analyzed in Chapter 9.3. Due to these effects, the non-dynamic TDC location was compared to a dynamic case, achieved by examining a motoring trace. As the dynamometer was the non-motoring type, a motoring trace was recorded as outlined in Chapter 4.2.1. Since the peak pressure location in motoring traces generally occurs between 0.5-1.5 BTDC due to the heat transfer to the cylinder walls during compression [7, 133], further information was required to establish TDC.

425

6.5 6.3 6.1 5.9 log (Pa) 5.7 5.5 5.3 5.1 4.9 4.7 4.5 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 y = -1.348x + 6.314 R2 = 0.999 y = -1.316x + 6.340 R2 = 1.000

100 kPa

log (Clearance Volume / ActualVolume)

Figure J.13: Motoring trace, gathered with a first cylinder WOT ignition cut. NA - PFI, 6000 rev/min, 100 kPa MAP, CR = 10, drift compensated, single cycle.

Consequently, a log-Pressure versus log-Volume diagram was generated for the motoring condition as shown in Figure J.13. From the P-V diagram and observing rules summarized in the literature [7, 92, 133], TDC location was calculated, with the phase accuracy recorded to lie within 0.2 CA. The relevance of TDC location and the effect on IMEP for SI engines is highlighted with a minimum phase accuracy of 0.3 CA needed to reduce IMEP errors below 1% [133].

J.9 Compression Ratio Calibration

As the engine design featured a fixed cylinder head relative to the crankcase, the CR could not easily be varied. In order to maintain the minimum squish height (0.8 mm) across all CRs to maximize in-cylinder squish and turbulence effects, CR variations were achieved through piston crown modification (Chapter 5.4.3). As the engine design coupled the cylinder head to the block via a gasketless interface (Chapter 3.6), the clearance volumes within the cylinder and cylinder head were measured using a burette filled with kerosene. With the piston crown geometry and resultant volume previously modeled, measured values could be compared and verified. Equation J.2 defines the CR calibration for this particular engine. It is noted that the static CR is defined, thus eliminating the possible volume changes associated with component flexure.

426

CR STATIC =

V S + VC VC
216.9 + (12 + 3.2 + Z) (12 + 3.2 + Z)

(J.2)

where:

CRSTATIC = static compression ratio, VS = swept volume (216.9 mL),

VC = clearance volume (cylinder head = 12 mL, squish height = 3.2

mL, piston crown = Z ).

427

428

Appendix K
P O W E R E D B Y W A T T A R D

Piston Repair

K.1 Piston Repair Techniques

Due to the numerous piston failures, limited supply and long lead times, pistons repairs were attempted. Piston repairs are not documented in the literature and unheard of in the piston manufacturing field [77]. For this reason, the technique used to repair pistons during this project is documented. As the pistons were forged from a high grade of weldable 2618A aluminum (Chapter 3.5.3), the feasibility of repairing pistons via welding was explored. A potential drawback from welding the heat treated pistons was the loss of mechanical properties from the welding and thus heating process. The welded pistons could not be re-heat treated as the distortion would make them unuseable due to the tight cylinder and gudgeon pin clearances. Fortunately, testing revealed that softer pistons actually improved engine reliability under heavy knocking conditions as the piston land was more pliable. However, it was found from the experiments that the softer lands allowed the top piston ring groove width to increase, resulting in severe piston ring flutter at high engine speeds. The ring flutter caused high blow-by rates which in turn expelled oil out of the crankcase breather. This could not be rectified and hence engine development continued at lower speeds (< 8,000 rev/min) where the combustion force exceeded the inertia of the piston ring. ring. Once the top ring groove clearance opened up excessively, the groove was re-machined to suit a new wider piston Although not the most desirable solution, piston repairs allowed

429

turbocharger engine development to continue for a further six months while awaiting the arrival of new pistons components. No failures were recorded for the repaired pistons, even under worse case heavy knocking conditions (Chapter 9.6). The repair technique is now described: 1) The failed top land was machined away with the oil ring groove retained. Distortion could be tolerated in the oil ring groove as the oil rails were pliable and forced to suit the piston groove by the ring expander. 2) The piston was preheated to 150C to minimize the distortion effect from welding. 3) The machined top land and valve pockets were tungsten inert gas (TIG) welded using a high silicon content filler material rod (10% silica). This gave the top land some structural integrity. unchanged. 4) Before the piston cooled down, a heavily lubricated piston pin was easily inserted into the piston pin boss, aided by the differing effects associated with thermal expansion. This ensured the pin bore did not heavily distort, with the oil ensuring galling did not occur on pin removal. 5) The piston was slowly air cooled under an insulated box. 6) Once cooled, the welded piston was machined to dimension, however slightly larger piston to bore clearances resulted (increased by 20-30 m). The increase clearance had no noticeable effect on oil consumption when compared to new pistons. A crossways welding technique was used to ensure the piston skirt geometry remained relatively

Figure K.1: Piston repairs needed due to failures and long lead times. (Left): Welded piston. (Right): Final repaired product with no failures recorded.

430

Appendix L
P O W E R E D B Y W A T T A R D

Engine Run-In

L.1 Engine Run-In Procedure

After engine disassembly and subsequent rebuild, an engine run-in was completed to ensure adequate bed-in of all components to minimize wear rates and improve reliability [112]. The adopted run-in procedure was based on previous studies completed by Repco Engine Development [89], with the run-in modified to suit the test engine. The run-in process involved a series of dynamometer tests up to 2/3 maximum engine speed and load conditions in the NA - Carbureted mode as shown in Figure L.1.

8000 7000 Engine Speed (rev/min) 6000 5000 4000 3000 2000 1000 0 0 20 40 60 80 Time (min) 100 120 140
1st point 2
nd

1
point

Engine Speed Load

0.8 0.6 0.4 0.2 0 160 Normalised Load

Figure L.1: Engine run-in procedure. NA - Carburetion.

431

The carbureted test mode was selected for the run-in process as this minimized the complexity of the fuel calibration, enabling focus to be placed on engine monitoring. The run-in sequence involved five minute interval tests starting at a high idle speed with a pre-heated water cooling system. The engine load and speed were then varied in 1/3 load and 1,000 rev/min increments, returning back to the high idle speed after each test point was completed. The run-in procedure took approximately 160 minutes to complete, with temperatures and pressures monitored during this time. Blow-by rates were also monitored during the run-in process to gauge component bed-in, with the effects upon returning to idle conditions shown in Figure L.2. Significant blow-by reductions were recorded during the run-in process when new components (pistons and piston rings) were installed due to the improved seal generated between the piston ring and liner surface [112]. However, negligible blow-by effects were seen if the engine was reassembled using identical components. After the engine run-in was completed, a series of WOT tests were completed to maximum speed, with individual cylinder compression later checked at warm cranking conditions. Prior to experiments continuing, blow-by rates, The lubricating oil and oil filter were also cylinder compression and brake performance were compared to previous baseline results to ensure engine reliability. inspected for metallic particles and consequently replaced.
10 Total Engine Blow-by (L/min) 2 Blow-by (% of induced air)

End of 1st point

9
End of 2nd point

1.8

1.6

1.4

6 0 10 20 30 Time (min) 40 50

1.2

Figure L.2: Blow-by effects upon returning to idle conditions during the engine run-in procedure shown in Figure L.1. New pistons and rings fitted. NA - Carburetion, fast idle speed (2300 rev/min), 50 kPa MAP, CR = 10.

432

Appendix M
P O W E R E D B Y W A T T A R D

Exhaust Gas Analysis

M.1 ADS-9000 Exhaust Gas Analyzer

Two Auto-Diagnostics ADS-9000 Super Five Gas Analyzers were used to measure dry emission concentrations, specifically HC, CO, CO2, O2 and NOX from each cylinders exhaust products. The NDIR method, where the infra-red absorption in a sample cell containing exhaust gas is compared to absorption in a reference cell, was used to measure the HC, CO and CO2 concentrations [173]. calibrated using known compositions of sample span gases. An electrochemical cell is used to measure O2 and NOX amounts, with detectors The ADS-9000 Oxygen calculates the AFR based on an algorithm described by Spindt [209].

concentrations are used to calculate the AFR in lean and stoichiometric regions and a combination of CO and CO2 concentrations are used in rich regions were O2 concentrations are minimal. The range and accuracy of the ADS-9000 is shown in Table M.1 [20]. To ensure ADS-9000 durability due to the NDIR system limitations, regular maintenance involved cleaning the particulate filters together with draining the wet products which were captured before entering the gas analyzer. Lengths of copper sampling tube were also used to reduce sample exhaust gas temperatures to within specified ranges, thus aiding in vapor condensation for the dry analysis.

433

Table M.1: ADS-9000 measurement ranges and accuracy [20]. Measurement Ranges

Channel
HC CO CO2 O2 NOX Temperature Engine Speed AFR Lambda () CO Correction

Range
0 - 10,000 ppm 0 - 10% 0 - 20% 0 - 23% 0 - 4,000 ppm 0 - 150C 0 - 10,000 rev/min 7 - 75 0.5 - 5.0 0 - 10%

Resolution
1 ppm 0.01% 0.01% 0.1% 1 ppm 1C 10 rev/min 0.01 0.001 0.01%

Accuracy

Channel
HC CO CO2 O2 NOX

Range
0 - 240 ppm 0.00 - 1.00% 1.00 - 2.00% 0.00 - 16.00% 0.00 - 2.00% 2.01 - 23.00% 0 - 1,000 ppm 1,001 - 2,000 ppm 2,001 - 4,000 ppm

Resolution
11 ppm 0.05% Absolute 0.06% Absolute 0.40% Absolute 0.10% Absolute 5.00% Relative 32 ppm 60 ppm 120 ppm

M.2 Exhaust Gas Sampling Position

The sampling position of the exhaust gas is quite critical in obtaining the actual exhaust product concentrations, needed for valid ADS-9000 and heated exhaust gas oxygen (HEGO) sensor analysis. The author had previously experienced exhaust gas concentration discrepancies when sampling too close to the exhaust ports in a high speed racing engine. This was associated with inhomogeneous sampled exhaust gas as the intake mixture entered the exhaust system during valve overlap. This effect is significantly magnified in high speed racing engines, were large degrees of valve overlap are used to improve air consumption and reduce the residual gas content. Published data from Horiba Pty Ltd recommends an exhaust sampling length ranging from 1-5 m to ensure homogenous mixture sampling when using HEGO sensors [101].

434

In order to facilitate individual cylinder exhaust analysis, short sampling lengths were unavoidable due to the exhaust plenum manifold which featured short primary headers. Thus two ADS-9000 analyzers sampled exhaust products from individual cylinders, with sampling points located in close proximity to the exhaust ports. To quantify the level of inhomogeneity in the exhaust gas, a study at varying load and speed conditions with varying sampling positions was completed. Results found negligible differences when comparing exhaust gas compositions for this particular engine. This is explained by the minimal through scavenging which occurred during valve overlap due to the camshaft design (Chapter 6.5.3). This is associated with narrow valve overlap angles and low poppet valve discharge coefficients in this region (Appendix H.5). Thus the validity of the exhaust gas analysis completed by the ADS-9000 could be ensured. Packaging constraints limited the HEGO sensor positioning, making individual cylinder HEGO analysis unfeasible. Thus an overall engine-out HEGO analysis was completed by sampling further down the system as shown in Figure M.1. minimize wet product exposure, ensuring equipment longevity and reliability. All sampling positions for both HEGO and ADS-9000 systems were elevated to

Upstream individual cylinder emissions sampling positions

Downstream HEGO sampling position

Figure M.1: ADS-9000 and HEGO sensor sampling position. (Left): Dynamometer test rig setup. (Right): Vehicle setup.

435

M.3 ADS-9000 Emission Correction

Obtaining correct emission concentrations is vital in comparing any experimental results with published data. This was achieved with a number of pre and post sampling steps which were developed based on previous experiential studies [62, 92, 266]. The following procedures ensured correct emission concentration levels.

M.3.1 Eliminating Air Leakage

Air leakage into the exhaust sample gas dilutes the exhaust products resulting in false emission concentrations and AFR. To avoid this, an airtight system was developed between the sampling points positioned in the exhaust header pipes and the emissions analyzer. This involved welding bosses on the exhaust manifold and fitting Swagelok olive connections between lengths of annealed copper tubing. Tube diameters were selected on the basis of minimizing the flow Maintaining a positive gauge pressure at the sampling points also restrictions into the gas analyzer, with 8 mm diameter tubing used to match the feed input. ensured an adequate supply of exhaust gas. This was achieved via positioning the sampling points relatively close to the exhaust ports. Turbocharger testing also ensured positive gauge pressures due to the flow restrictions associated with driving the turbine.

M.3.2 Correcting for Hydrocarbon Type

Irrespective of the fuel used, the ADS-9000 displays HC concentrations based on the infra-red properties of hexane and thus displays C6 equivalent. It is important to note the molecular composition and the number of carbon atoms per molecule when discussing and comparing raw HC emissions. Thus, all HC values presented in this thesis are quoted in C6 equivalent. However, correction can be made to any form of HC from the ADS-9000. Consider the sampling of gasoline (C8.26H15.5) with the ADS-9000 displaying HC emissions as 100 ppm. This corresponds to 100 ppm HC in C6, however 73 ppm HC in gasoline (C8.26H15.5). Thus the type of HC concentration should be specified in the calculation of specific emissions (Section M.5).

436

M.3.3 Hydrocarbon Sensitivity

As previously mentioned, the ADS-9000 NDIR measurement detects HC concentrations for varying HC types with a different response, depending upon the emission spectra compared with hexane. For gasoline fueled engines, HC emissions determined from flame ionization detector (FID) analyzers are approximately twice the levels found by NDIR analyzers [109]. Dober and Zakis had previously determined calibration factors for different fuel types using known concentrations of sample span gases [62, 266]. measurement. These factors provide a correlation between actual HC concentrations versus that displayed by the NDIR Calibration was achieved by performing a direct concentration study between the ADS-9000 and a fast FID, which has a consistent response irrespective of the HC type. The results of this study are presented in Table M.2 with Static referring to the unburned mixture and Dynamic to the firing engineout condition.
Table M.2: ADS-9000 sensitivity to various fuels [62, 266]. Hydrocarbon Type Natural Gas Butane Propane Gasoline Static Calibration Factor 2.97 0.89 1 2.29 Dynamic Calibration Factor 3.03 0.89 -

M.3.4 Wet/Dry Analysis and Compensation

The ADS-9000 performs a dry analysis, thus displaying higher emission concentrations when compared to actual emissions which include the water vapor content in the breakdown. Hence, compensation for all dry analysis products is needed to ensure accurate results. The amount of compensation needed varies, since certain fuel compositions yield higher concentrations of water vapor than others. Thus wet concentrations can be obtained by multiplying the dry emission product with the vapor compensation ratio (VCR ) defined in Equation M.1. If the measurements are accurate and the calculations have been made correctly, the total composition of products should equal 100%.

VCR =

moles of products without water moles of products without water + moles of water

(M.1)

437

M.4 AFR Calculation

Determining the AFR by mass was undertaken using three methods. This ensured measurements could be cross-checked for accuracy and an average value computed.

Determination methods included: Calculation from air and fuel measurement Calculation from the exhaust products Exhaust gas oxygen sensor measurement

M.4.1 Calculation from Air and Fuel Measurement

This method of AFR determination is trivial if air and fuel consumption rates are known. Compensation needs to be made for temperature effects if volume flow rates are used. This method was largely not used as direct air consumption was not measured, but rather calculated from the AFR (Chapter 4.3.1). However, AFR checks were made when airflow rates were known, associated with choked flow operating conditions (Chapter 4.3.1). When directly measuring air consumption, careful attention should be paid due to the possibility of unsteady flow across the measurement device. Unsteady flow is associated with intake resonance amongst other factors, thus making measurement location critical for hot wire, vane or orifice plate devices.

M.4.2 Oxygen Measurement

AFR was measured directly by sampling the exhaust products using several oxygen based sensors, which produce a voltage signal dependent on the oxygen concentration in the exhaust gas stream. Advantages of this system include quick response times which enable near real time monitoring and logging. determination was used during dynamometer and vehicle calibration. This Overall eliminates the need for timely calculations and hence this form of AFR engine AFR was directly measured using a Bosch LSM-11 HEGO sensor (Table 4.1). The wideband sensor incorporated an electric heater, which allowed quick warm-up times for fast oxygen measurement. This was especially well suited for measurements in the lean and rich regions [33]. Individual cylinder AFR was directly measured using oxygen cells in the two ADS-9000 exhaust gas analyzers.

438

M.4.3 Calculation from the Exhaust Products

Performing a chemical balance from the known individual cylinder exhaust products gathered from each cylinders exhaust gas analyzer enabled individual cylinder AFR to be calculated. Although the ADS-9000 conducts a dry analysis, no compensation is needed in the AFR calculation as the relative concentration of products does not change. The general chemical equation expressed by M.2 has underlying species found from the ADS-9000, with subscripts y, k and x chosen to be representative of the molecular species for which the concentrations from the exhaust gas analyzer are known.
nf [FUEL]+ na [AIR] aCO2 + bH2O + cN2 + dO2 + eHyCk + f CO + gNOx (M.2)

Balancing:

C: H: O:

nf [ Cs] = a + ek + f nf [ Hs] = 2b + ey nf [ Os ] + 2na = 2a + b + 2d + f + gx

(M.3) (M.4) (M.5) (M.6)

Rearranging:

nf = a + ek +

f C ' s
2

b=

nf H ' s ey 2

(M.7) nf O ' s 2

na = a +

b
2

+d +

f
2

gx
2

(M.8)

Giving:

AFR mass =

na 1 +

nf M fuel

79 M air 21

(M.9)

where: Mfuel

= molecular weight of Optimax gasoline = 114.62 g/mol = molecular weight of air = 28.96 g/mol

Mair
Sample Calculation

A sample AFR calculation for a lean condition using gasoline is provided, to demonstrate actual individual cylinder exhaust emission concentrations and AFR. Used in the calculation is sample raw data from the ADS-9000 analyzer, shown in Table M.3.

439

Table M.3: Raw engine-out emission data from the first cylinders ADS-9000 emissions analyzer. TC - PFI, 8000 rev/min, 70 kPa MAP, CR = 10. CO2 14.22% O2 1.67% CO 0.26% NOX 1541 ppm C6H14 35 ppm

Applying the emission correction outlined in Section M.3.3, involving multiplying the HC concentrations by 2.29 (Table M.2), Table M.4 displays actual emission concentrations.
Table M.4: Actual emission concentrations after Table M.3 correction. CO2 14.22% O2 1.67% CO 0.26% NOX 1541 ppm C6H14 80 ppm

Rewriting Equation M.2, with x = 1 due to the major portion of NOX being NO k = 6 due to the ADS-9000 specifying HCs as C6 y = 11.2 (using a H:C ratio of 1.87 for gasoline)

nf ( C 8.26H15.5 ) + na O2 +

79 N2 21

(M.10)

80 1541 14.2CO2 + b H2 O + c N2 + 1.7O2 + H11.2 C6 + 0.26CO + NO 10000 10000

Balancing and rearranging: C:

80 6 8.26nf = 14.22 + + 0.26 10000 nf = 1.76 80 11.22 15.5nf = 2b + 10000 b = 13.59 1541 2na = 2 (14.22 ) + b + 2 (1.67 ) + ( 0.26 ) + 10000 na = 22.89 79 22.89 1 + 28.96 21 = = 15.65 1.76 114.62 (M.11)

H:

(M.12)

O: Giving:

(M.13)

AFR mass

(M.14)

The AFR value of 15.65 calculated from the exhaust products compares to the oxygen measured value of 15.54.

440

M.5 Brake Specific Emissions Calculation

Due to the multi-cylinder arrangement, emissions analysis was conducted on an individual cylinder basis to compensate for cylinder variations, (Section M.2). Therefore, mass emission values were also computed on an individual cylinder basis and combined to achieve total engine-out mass emissions. These were then incorporated with total engine-out brake power to calculate brake specific emissions. This method also allowed emission correction to constant values (Section M.6.2).
Cyl 1 Emission + Cyl 2 Emission Brake Specific Emission= Brake Power n s , cyl Cyl 1 Emission = n p , cyl n s , cyl Cyl 2 Emission = n p , cyl
1

(M.15)

Ms 1 M p , cyl Ms M p , cyl

m fuel , cyl 1

air , cyl +m

) )

2 2

m fuel , cyl

air , cyl +m

where:

np = moles of total products, ns = moles of sample product, Mp =

molecular weight of products, Ms = molecular weight of sample product (e.g. NO, CO, HC etc.), = mass flow

M.6 AFR Variation

M.6.1 Efficiency Correction to Alternate
Efficiency correction to constant ( = 0.9,1) allows clear comparisons and trends to be analyzed, which were previously unrepresentative due to mixture variations associated with the tuning strategy amongst other factors. sensitivity analysis on their effects. To facilitate the correction, several assumptions were made, which are outlined together with a The efficiency correction assumes no alteration to the engines operating limits as outlined in Chapter 7, even though fuel enrichment was used as a method of knock control. However, it is argued that knock may be controlled with the implementation of various other strategies (Chapter 7.2). This would allow near stoichiometric operation in regions which were heavy enriched to avoid the DL.

441

1000

975 BMEP (kPa)

950

925

900 0.85

0.875

0.9

0.925
Lambda Lambda( )

0.95

0.975

Figure M.2: BMEP effects for varying . NA - PFI, 6000 rev/min, WOT, CR = 10.

It is noted for simplicity that the correction assumes no brake power loss due to the fuel correction, with corrected efficiency values calculated from altered fuel consumption rates. Power changes are assumed to be negligible for the qualitative analysis. Figure M.2 displays peak BMEP occurring in the ~0.9 region for this particular test mode and operating conditions. Peak BMEP occurs slightly rich of stoichiometric due to dissociation at the high temperatures following combustion. At stoichiometric conditions, molecular oxygen is present in the burned gases enabling some additional fuel to be added and partially burned to increase temperature and pressure, thus increasing power and BMEP [98]. Figure M.2 shows less than 2% BMEP variation for ranging between 0.86-0.94. Furthermore, only a 4% reduction is recorded at stoichiometric fuel conditions. These results are comparable to other test engines, with 5% IMEP variation previously recorded for a variation from 0.75-1 [98]. Hence, it is assumed that the minor change in performance due to variation in experiments will have negligible effect on corrected efficiency trends, with corrected results lying within the bounds of experimental error.

M.6.2 Emissions Correction to Alternate

Due to the varying , exhaust emission calibrations were completed to observe the engine speed, MAP and CR effects for two constant values ( = 0.9,1). The calibrated results are intended for qualitative analysis and enable trends across all

442

modes to be compared on the same basis for this particular engine. Calibrations are made on an individual cylinder basis, with engine-out brake specific emissions then determined as outlined in Section M.5. Previous assumptions concerning the effect on fuel flow rate, brake power and operating limits outlined in the efficiency correction (Section M.6.1) are also incorporated into the emissions correction. Emissions correction to constant is more difficult to achieve as the emissions formation and engine-out products significantly vary with variation, as shown in Figure 7.12.
CO, CO2 and O2 Correction

Emissions correction analysis was initially performed for each data set (engine mode and test CR) using dimensionless dry concentrations, with emissions plotted for contours of constant power, achieved with engine speed and load variations. However, CO, CO2 and O2 results showed merging trends for varying power outputs in each data set. This allowed a general calibration within each data set, as exhaust concentrations were also found to vary substantially between data sets due to varying engine designs [98]. Figure 7.12 displays CO, CO2 and O2 dry concentrations on a molar basis for a specific data set for this particular engine. Results show that CO, CO2 and O2 are heavily dependent on fuel mixtures. The formation of these emission products is also reliant on the interacting species, with CO, CO2 and O2 concentrations affected by each other. CO, CO2 and O2 emission correction to alternate values was not performed using calibration constants, due to the possibility of error compound effects associated with high fold changes in emissions (e.g. 17 fold CO reduction from = 0.9 1). Exhaust product chemical equilibrium (Equation M.2) could also not be ensured. Hence a carbon balance initiating from the mole compositions of the dry exhaust products was used to correct individual data points to alternate values. Methodology behind the CO and CO2 correction is given in Figure M.3. It is noted that the HC portion is neglected in the carbon balance due to its insignificant contribution. The carbon balance also assumes equal air quantities at a given data point, with the carbon content and fuel consumption altering to accommodate for alternative values. Correction is made as in Equation M.16.

443

Experiment (A) CO, CO2 and HC emissions, A (mole basis) Fuel and air mass flow, A (g/h) Fuel correction from A to B

Alternate (B) CO determined from Figure 7.12, B (mole basis)

Neglect HC portion as insignificant

CO and CO2 mass flow rate, A (g/h)

CO mass flow rate, B (g/h)

CO2 mass flow rate, B (g/h)

Carbon out mass flow rate, A (g/h)

Carbon correction from A to B

Carbon out mass flow rate, B (g/h)

Brake Power (kW)

BSCO2 corrected to B (g/kWh)

Figure M.3: Block diagram displaying the carbon balance methodology used to correct BSCO2 emissions to constant values. Methodology also applied to BSCO emissions.

Carbon , A = ( A ) m Carbon , B m Carbon , A = m

M Carbon M CO , A + Carbon m M CO M CO 2 nCO , A n products , A nCO 2, A n products , A M CO 2, A + Carbon m M HC

(M.16) HC , A m 0

CO , A = m

M CO M products , A M CO 2 M products , A

air , A +m m fuel , A

) )

CO 2, A = m Carbon , B = m

air , A +m m fuel , A

M M Carbon CO , B + Carbon m M CO 2 M CO nCO , B n products , B

M CO 2, B + Carbon m M HC

HC , B m

CO , B = m CO 2, B m

M CO M products , B

fuel , A + m air , A ( A ) m

CO 2, B m Corrected BSCO2, B = Brake Power

444

where:

= mass flow, A = actual , B = corrected , n = number of

moles or mole fraction, M = molecular weight, nCO, Figure 7.12 (nCO, 1 = 0.0075, nCO,
0.9 B

is found from

= 0.043).

HC Correction

Due to the effects of speed and load variations on HC emissions [98], corrections to alternate ( = 0.9,1) were achieved using mole concentrations and brake power contours. Figure 7.12 shows that unburned HC exhaust concentrations vary substantially with engine design and operating conditions as previously documented [98, 220]. Hence correction was completed for each data set due to the changing surface to volume ratios and induction systems associated with CR and test mode changes.

550 HC concentrations (ppm C6)6) 450 350 250 150 50 0.7 0.8 0.9 Lambda ( ) 1.0

0-20 kW (Cyl 1) 20-40 kW (Cyl 1) 40-60 kW (Cyl 1) 0-20 kW (Cyl 2) 20-40 kW (Cyl 2) 40-60 kW (Cyl 2)

1.1

1.2

Figure M.4: Individual cylinder HC emission concentrations (ppm C6) for gasoline with and power variation (engine speed and load) within a given data set. TC - PFI, CR = 10.

Figure M.4 displays individual cylinder HC emissions versus for three power output ranges, low (0 -20 kW), medium (20-40 kW) and high (40-60 kW). Trends from both cylinders show clear HC trends relative to power in the data scatter, allowing HC emission correction to alternate for each data set. Trends show decreasing levels of HC emissions as power is increased for a given . Reductions in HC emissions are also shown as the mixture is leaned from the rich state, while still maintaining adequate combustion stability and avoiding the misfire region.

445

When comparing HC concentrations from both cylinders, values from the second cylinder are considerably lower when compared to the first. This is related to the crevice volume variations [163, 253] between both cylinders associated with the cylinder pressure transducer access port. Although the extra crevice volume is small (~ 0.3 mL), comparatively it contributes almost a 50% increase in the first cylinders total crevice volume due to the modern engine design. Hence, engineout HC emissions presented in this thesis are achieved using the second cylinders HC results to negate the pressure transducer crevice volume effects which would not be present in a production version.

446

Appendix N
P O W E R E D B Y W A T T A R D

Combustion Modeling

N.1 Introduction
This appendix describes the framework behind E-CoBRA [91], the two-zone quasidimensional model used to complete the combustion analysis in this thesis. Models and equations behind E-CoBRA are presented to justify the combustion results. In this appendix, only a brief summary of salient points is included and the reader is referred to Hamoris work [92] for further detail and validation behind the in-house developed software.

N.2 Model Outline

The E-CoBRA program is divided into three main loops: the compression of unburned charge, the two-zone combustion including compression and expansion, and the expansion of the completely burned products. The condition of the charge (pressure, temperature, AFR, type of fuel) and the engine specifications (engine speed, CR, spark timing) at the start of the compression are required as inputs. A basic solution flowchart is shown in Figure N.1, followed by a detailed derivation of all of the equations used in the software.

447

Start Input Experimental Text File and Initialize Engine Configuration

Convert Data to Crank Angle & Pressure

Peg Pressure Relative to TDC

Knock Analysis

Smooth Pressure Trace Y Setup Flame Table & Calculate Residual Gas Fraction

Compression (single-zone)

Combustion (two-zone)

Calculate Adiabatic Flame Temp., Burned/Unburned Volume N Is Flame Kernel Initialized? Y Calculate up to 100% burn duration, Burned/Unburned Flame Temperature, Equilibrium Concentrations

Expansion (single-zone)

Detailed Combustion Analysis

Print Results to Screen

Y Calculate Mean Temperature, Flame Surface Area, Flame Speed, FSRa, u, Turbulent Reynolds and Damkohler Numbers etc

Analyse more cycles?

End

Figure N.1: Solution procedure flow chat for E-CoBRA.

448

N.3 Engine Geometric Relationships

Variables which define the engine geometry are outlined and shown in Figure N.2.

TDC

b b = bore diameter
s

s l

= stroke length = connecting rod length

a = crank radius = s/2

BDC

R = l/a
CR = compression ratio = (Vs/Vc)/Vc

= distance between crankshaft axis and piston pin axis

Vs = swept volume a Vc = clearance volume

Figure N.2: Basic geometry of the reciprocating IC engine.

Piston position:

x = l 2 + a 2 sin 2 + a cos
Combustion chamber area:

(N.1)

A=

s 2 b 2 + b R + 1 cos ( R 2 sin 2 )1 / 2 + 2 2 CR 1

(N.2)

Combustion chamber volume and its derivative:

V =

Vs V + s R + 1 cos ( R 2 sin 2 )1 / 2 CR 1 2

(N.3)

dV Vs = sin 1 + cos ( R 2 sin 2 ) 1 / 2 d 2

(N.4)

449

N.4 Combustion Modeling

N.4.1 Residual Gas Mass Fraction
The residual gas fraction is estimated using the model developed by Fox et al. [76]. The regression equation to predict residual fraction is the following:

OF pi X r = 1.266 N pe
where:

0.87

pe pi + 0.632

( pi / pe )0.74
CR

(N.5)

N = engine speed, pi = intake pressure (kPa), pe = exhaust

pressure (kPa), CR = compression ratio, = fuel/air equivalence ratio, OF = overlap factor (mm2 /L)

The overlap factor is calculated using the following formula:

OF =
where:

1.45 (107 + 7.8 + 2 ) LV ,max2 DV b b

(N.6)

b = bore diameter, LV = maximum valve lift, DV = valve inner seat

diameter (all dimensions are in mm), = valve overlap in crank angle degrees at 0.15 mm valve lift.

N.4.2 Compression and Expansion Process

The compression and expansion processes equate when applying the first law to the charge as a single-zone, with derivation as follows. The single-zone model (Figure N.3) operates throughout the compression and expansion part of the cycle. The heat transfer to the working fluid is positive and the work transfer from the working fluid is positive. The differential first law for this model for a small crank angle change is:

dQ dW dU = d d d dQ dV dT P = mcv d d d

(N.7) (N.8)

450

Open system boundary Qht

W
Figure N.3: Open system boundary for single-zone combustion chamber.

From the ideal gas equation PV = mRT

dP dT 1 dV = P +V d R d d

(N.9) (N.10)

dU cv dV dP = P +V d R d d
The first law now becomes:

dQ dV cv dV dP P = P +V d d R d d
Rearranging:

(N.11)

dQ cv dV cv dP 1 + P = V d R d R d dP 1 R dQ R dV = 1+ P d V cv d cv d

(N.12) (N.13)

Note for N.13: R = Cp-Cv and R/Cv=(Cp/Cv)-1 Now from (N.9)

dT 1 dV 1 dP =T + d V d P d

(N.14)

Equation N.13 and N.14 are the governing equations for the compression and expansion part of the cycle.

451

N.4.3 Combustion Process

During the combustion phase, the charge is assumed to be divided into two-zones (Figure N.4) with the equations derived from the equation of state, the first law, and the conservation of mass and volume equations as follows. The two-zone model operates during the combustion part of the cycle. The heat transfer to the working fluid is positive, and the work transfer from the working fluid is positive. The energy equation for the entire system changed at any time instant (corresponding to CA) is described by the first law:

dU dQ dW = d d d mcv dT dQ dV = P d d d

(N.15) (N.16) Open system boundary

Qb

Qu

Tb mb

Tu m u

Wb

Wu

Figure N.4: Open system boundary for a two-zone combustion chamber.

Internal energy of the system:

U = mu uu + mbub
Differentiating N.17 yields

(N.17)

dU du dmu du dmb = mu u + uu + mb b + ub d d d d d

(N.18)

452

Mass conservation:

m = mu + mb

dmu dm = b d d

(N.19)

Applying N.19 to N.18 and combining with N.16 (Note: uu = cvu dT)

(ub uu )
Volume conservation:

dmb dT dT dV dQ + mu cv u u + mb cv b b + P =0 d d d d d dV dVu dVb = + d d d

(N.20)

V = Vu + Vb
From PV = mRT using the product rule

(N.21)

dV dP dT dm +V = mR + RT d d d d

(N.22)

1 dV 1 dT 1 dm 1 dP = + V d T d m d P d
Differential equation of state:

(N.23)

1 dmu 1 dTu 1 dP dVu = Vu + d m u d Tu d P d 1 dmb 1 dTb 1 dP dVb = Vb + d mb d Tb d P d

(N.24)

(N.25)

In N.24 and N.25, replace V/T with mR/P (since PV = mRT) and substitute N.24 and N.25 into N.21 while applying N.19.

dV Vb Vu dmb Ru mu dTu Rb mb dTb V dP + + = d P d P d P d mb mu d

Applying the first law to the unburned zone

(N.26)

1 dTu 1 dP 1 dQu duu = Pvu d mu d T d P d u

(N.27)

453

Expanding the bracket of N.27 and applying cp = R + cv

mu c pu

dTu dP dQu = Vu + d d d

(N.28)

From N.20, N.26, and N.28, the following equations govern the combustion phase.

dTu Vu dP 1 dQu = + d mu c pu d mu c pu d

(N.29)

dTb P = d mb Rb

dV (RbTb RuTu ) dmb RuVu dP Ru dQu + V dP (N.30) P d P d Pc p u d Pc pu d d

cvb dV dQ dP 1 = 1 + P d cvu cvb Ru cvb Rb d d Vu + V Vu c R c Rb p b p u u Ru dmb + (ub uu ) cvb T Tu b Rb d cv cv R dQ + u b u u c pu Rb c pu d

where: Rgas =

(N.31)

n Runiversal m

cv = c p R

u = h RT

N.5 Heat Transfer

The bulk heat transfer was calculated by a method extensively used in the literature and outlined below. The following equations are used to account for convection and radiation (5% of total heat transfer), assuming a 1-dimensional heat flow through the cylinder wall, piston and cylinder head [87, 98].

dQ dQu dQb = + d d d dQu 30 Au = hcu (Tg u Tw ) + (Tg4u Tw4 ) d N dQb 30 Ab = hcb (Tg b Tw ) + (Tg4b Tw4 ) d N

(N.32)

(N.33) (N.34)

454

where:

A = surface area, N = engine speed (rev/min), hc = convective

heat transfer coefficients, = 0.6, = Stefan-Boltzmann constant = 5.67*10-8 (W/m2*K4),

To be more precise, the total heat transfer in each zone is the sum of the heat transfer between the gas and cylinder, the gas and cylinder head, the gas and piston, and lastly the heat transfer between the two zones. Therefore, the heat transfer for the burned and unburned zones becomes:

dQu dQcylinder dQcylinder _ head dQ piston dQzone = + + + d d d d d dQb dQcylinder dQcylinder _ head dQ piston dQzone = + + + d d d d d

(N.35)

(N.36)

To estimate the heat transfer coefficient (hc) the Woschni correlation is used. This correlation has been extensively used in previous studies due to the simple adjustment of constant C in Equation N.37. Hence, the heat transfer coefficient of any engine can be correlated. The heat transfer coefficient must apply to both burned and unburned zones.

hc = C D 0.2 P 0.8T 0.55 0.8

where:

(N.37)

D = diameter of piston, P = cylinder pressure in (kPa), T = zone

temperature (K)

Woschni hypothesized that the average gas velocity is proportional to the mean piston speed. To account for the change in density due to combustion, he introduced a pressure rise due to combustion term (P-Pm). The average cylinder gas velocity, identical for both burned and unburned zones is therefore expressed as follows:

= C1S P + C2

Vd Tr1 (P Pm ) PrVr

(N.38)

455

where:

Vd = displaced volume in (m3), Vr (m3), T1 (K), P1 (Pa) represents

the known state of the working gas related to inlet closure or ignition, Sp = average linear engine speed, Pm = corresponding motoring pressure in the absence of combustion, C1 = 6.18 for gas exchange, C1 = 2.28 for compression and expansion, C2 = 0 for gas exchange and compression, C2 = 3.24*10-3(m/sK) for combustion and expansion

N.6 Chemical Equilibrium

A chemical equilibrium solver was implemented into the two-zone model as the detail and accuracy of a computationally demanding kinetic combustion model was not required for the purpose of this research. The rapid computation of chemical equilibrium compositions (RCCEC) method developed by Erickson and Prabhu was used for calculating the concentration of chemical species [69]. This method was chosen because it is up to 80 times faster than the often-used free-energy minimization method. The chemical system is composed of four elements (C, H, O, N) and ten reacting species: H2O, CO2, CO, O2, H2, N2, H, O, OH, and NO. Starting with an appropriate set of equilibrium and elemental balance equations, the ten equations are reduced to two equations and then to a single equation containing one unknown variable. This single equation is then solved with a Newton iteration scheme. For a reacting gas mixture containing ten species composed of four chemical elements, there are six independent chemical reactions that can be written. The chemical reactions considered in the scheme are: CO + H2O 2CO2 H2+O2 H2 O2 O2+N2 CO2 + H2 2CO + O2 2OH 2H 2O 2NO [1] [2] [3] [4] [5] [6]

456

The following describes the solution procedure of the product species for a given input: 1.) Input parameters:

Temperature (T, Kelvin) Density (, kg/m3) Equivalence Ratio () H/C ratio mol CO/ mol CO2 ratio ( = initial estimation) Nitrogen/Oxygen ratio (air = 79/21)

2.) Compute constants: Chemical equilibrium constants Kp were curve fitted using the famous vapor pressure model exp(a+b/x+c*log(x))

CO + H2O CO2 + H2

(N.39)

T < 800 K1 = Exp(-11.5723745 + 5255.72032 / T + 0.96398354 * Log(T)) T <= 2600 K1 = Exp(-11.927269 + 5193.65214 / T + 1.02850772 * Log(T)) T <= 6000 K1 = Exp(-9.90914812 + 4728.77893 / T + 0.79405057 * Log(T))
2CO2 2CO + O2
(N.40)

T < 800 K2= Exp(15.4600070 - 67772.0986 / T + 0.77535057 * Log(T)) T <= 6000 K2= Exp(30.2986377 - 69188.1029 / T - 1.18303008 * Log(T)) K2= K2* Po / (Density * Runi * T)
H2+O2 2OH
(N.41)

T < 800 K3 = Exp(4.81617915 - 9554.57691 / T - 0.14234455 * Log(T)) T <= 2700 K3= Exp(6.42554150 - 9664.76907 / T - 0.36267753 * Log(T)) T <= 6000 K3 = Exp(8.40910545 - 10346.7118 / T - 0.58193587 * Log(T))
H2 2H
(N.42)

T < 800 K4 = Exp(1.96193395 - 51996.3266 / T + 1.48000464 * Log(T)) T <= 2500 K4= Exp(4.62582493 - 52329.0289 / T + 1.14258246 * Log(T)) T <= 6000 K4 = Exp(11.8839457 - 54344.0182 / T + 0.31650772 * Log(T)) K4 = K4 * Po / (Density * Runi * T)

457

O2 2O

(N.43)

T < 800 K5= Exp(4.80924042 - 59545.6966 / T + 1.39032710 * Log(T)) T <= 3000 K5= Exp(10.9084500 - 60205.2613 / T + 0.59937809 * Log(T)) T <= 6000 K5 = Exp(15.2261961 - 61463.4378 / T + 0.11118509 * Log(T)) K5= K5 * Po / (Density * Runi * T)
O2+N2 2NO
(N.44)

T < 800 K6 = Exp(2.73981886 - 21709.6133 / T + 0.03721003 * Log(T)) T <= 3000 K6 = Exp(3.13453339 - 21774.4906 / T - 0.00959753 * Log(T)) T <= 6000 K6 = Exp(6.89000288 - 22968.6081 / T - 0.42951512 * Log(T)) Ka = (K4 / (K2 * K3)) ^ 0.5 Kc = K1 * K2 * K3
(N.45), (N.47),

Kb = (K2 * K5) ^ 0.5 Kd = K2 * K6

(N.46) (N.48)

Table N.1 Constant coefficients amn.

N 0 1 2 3

a0n _______

a1n -Kc -2(2H+KaKc) -(4HK1+Ka2Kc) _______

a2n _______ 4 8K1 4K12

H2
_______ _______

where: i = mole number of species i, (mol of species i / kg mixture) H, O, N, C = mole numbers of elements

Table N.2 Constant coefficients fmn and g0n.

n 0 1 2

f0n 2K2 2K2(Ka+1)+Kb 2c+ H - o + (Ka+1)Kb+2K2Ka

f1n

g0n 1 Ka+1 Ka

-1

3 4

c+ H - o +
(2c - o + Kb)Ka (c+ o)Ka

Ka 2K1 - 1 Ka 2K1

458

3.) Iteration loop starts here

a m = a mn n
n=0
4

m = 0, 1, 2 m = 0, 1

(N.49) (N.50) (N.51)

f m = f mn n
n =0 2

g 0 = g 0 n n
n =0

10 = K d [1 + 8 N 2 / K d ] 1 / 4 2
1/ 2

(N.52) (N.53)

b1 = f1 b0 = f0 + g0 10 2
Corrected (original was: b0 = f0 - g0 10 )
2

(N.54) (N.55)

F() = a2 b02 a1 b1 b0 + a0 b12 = 0

F() = b02 a2 -b1b0 a1 + b12 a0 + (2a0 b1 - a1 b0 ) b1 + (2a2 b0 - a1 b1 ) b0 (N.56)

' am = namn n1 3

m = 0, 1, 2
2

(N.57)

n =1
4

' b0 = nf 0 n n1 + 2 N g 0 /(1 + 8 N 2 / K d )1 / 2 + 10 2 ng 0 n n1

(N.58)

n =1

n =1

N.58 is the corrected equation, the original equation was:

' b0 = nf 0 n n 1 2 N g 0 /(1 + 8 N 2 / K d )1 / 2 10 2 ng 0 n n 1 n =1 n =1 4 2

b1' = f1' = nf1n n1

n=2

(N.59) (N.60) (N.61) (N.62)

l+1 = l - F(l) / F(l)

|(yi)l+1 (yi)l| ERROR 1- ((l - F(l) / F(l) ) / l) ERROR
Iteration loop finishes here

459

4.) Compute final composition

1 = -b0 / b1 4 = K2 / 2

(N.63);

2 = C / ( +1) (N.64); 5 = K1 1
(N.67);

3 = C / ( +1) 6 = (N - 10)/2 8 = (K2K5)1/2 /

(N.65)

(N.66);

(N.68)

7 = (K4/K2K3)1/2 / 9 = (H - 2(1 + K1) 1 ) / (Ka + 1)

(N.69);

(N.70)

(N.71)

5.) For all species the thermodynamic properties such as specific heat and enthalpy are in the form of the Shomate Equation [164]. The accuracy of this format is equivalent to the more popular NASA polynomials. It is important to note the expression of specific heat when the composition changes through a series of equilibrium states.

c p = xi c p +hi
i

x T x

(N.72)

where: xi = mole fraction of specie

hi = enthalpy of specie

The second term hi T takes into account the dissociation of molecules at higher temperatures. It approaches zero if the composition is frozen and the components in the mixture are ideal gases (resulting in an ideal gas mixture).
8 7.5 7 6.5 6 Cp/R 5.5 5 4.5 4 3.5 3 0 500 1000 1500 Temperature (K) 2000 2500 3000
N2 CO H2 O2 CO2 H2O

Figure N.5: Specific heat at constant pressure cp/R, as a function of temperature for species CO2, H2O, O2, N2, H2, and CO [164].

460

29
1750 K

Molecular weight (g/mol)

28.5 28 27.5 27 26.5 26 0.2 0.3 0.4 0.5 0.6 0.7 0.8

2250 K 2750 K

0.9

1.1

1.2

1.3

(1/Lambda) 2.5
2750 K

2.25 Cp Burned (KJ/kg K) 2 1.75 1.5 1.25 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 (1/Lambda) 1.3 1.28 Gamma Burned ( b) 1.26 1.24 1.22 1.2 1.18 1.16 1.14 1.12 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 (1/Lambda)
2750 K 2250 K 1750 K 1750 K 2250 K

Figure N.6: Fuel Isooctane, equilibrium burned gases as a function of equivalence ratio at T = 1750 K, 2250 K, and 2750 K, at 30 bar. (Upper): Molecular weight. (Middle): Specific heat. (Lower): Ratio of specific heats. Results from E-CoBRA.

461

1
O2 N2

1750K
CO2 CO

H2O

0.1 Mole fraction

0.01
NO

H2

0.001
OH

0.0001 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (1/Lambda) 0.9 1 1.1 1.2 1.3

1
N2 O2

2250K
CO2

H2O CO

0.1 Mole fraction

0.01

NO H2 OH

0.001
CO O OH H

0.0001 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (1/Lambda) 0.9 1 1.1

1.2 H 1.3

1
N2 O2

2750K
CO2 NO

H2O CO H2

0.1 Mole fraction

0.01

OH O CO H

OH O2

0.001

0.0001 0.2 0.3 0.4 0.5 0.6 0.7 0.8 (1/Lambda) 0.9 1 1.1 1.2 1.3

Figure N.7: Fuel Mole fraction of equilibrium combustion products of isooctane-air mixture as a function of equivalence ratio. (Upper): T = 1750 K, (Middle): T = 2250 K, (Lower): T = 2750 K. Results from E-CoBRA.

462

N.6.1 Chemical Equilibrium Solver Accuracy

The chemical equilibrium solver produced identical results when compared to packages developed by Olikara and Borman [170], and Eriksson [70], which used a similar number of chemical reactions and reacting species. The solver was also compared to the sophisticated NASA Lewis Chemical Equilibrium Code (CEC76), which was revised in 1976. Some of these results are published by Heywood [98] and Negus [166] and in general, the errors were less than 1.0% for the specific heat of burned mixtures at the pressures and temperatures of interest. The maximum error of 2.9% occurred at 2750 K, 30 atm, and at a stoichiometric AFR. Based on the high accuracy and fast computational speed, it was concluded that the solver developed by Erickson and Prabhu [69] is more than adequate to satisfy the objectives of this research.

N.7 Flame Geometry

E-CoBRA utilizes a spherical flamelet model as shown in Figure N.8. In order to reduce computational time, the symmetry about section A-A and B-B from Figure N.8 must be realized, with parameters defined. Point of Contact:

cos =

2 Rc2 R 2 + Rh 2 RRh

(N.73)

cos =

2 R2 Rc2 + Rh 2 Rc Rh

(N.74) (N.75)

2 R = R2 f z

Burned Gas Volume:

h 0

Vb = R 2 + Rc2 Rc Rh sin dz
Flame Front Area:

(N.76) (N.77) (N.78) (N.79) (N.80)

A f = 2 R dz
0

Burned Wall Area:

Wall _ Area = Ac + Ah + Ap
Ac = 2 Rc dz
0 h

(N.43)

Ah = R 2 + Rc2 Rc Rh sin

2 2 Ap = R 2 f ( h Rv ) + Rc Rc Rh sin

463

The spherical flame geometry equations were validated against Poulos et al. [178]. The results for the equations are modeled for a disc combustion chamber, however the flat triangular approximations have been validated for pent roof and bowl in piston type combustion chambers [92, 148].

Burned Volume - Vb

Area of Flame Front - Af

Point of Ignition A

Rh

Area of Cylinder Wall - Ac

Area of Cylinder Head Ah

Rc

Rv
B Area of Piston Ap

Rf R

Figure N.8: Spherical flame geometry in a combustion chamber.

464

0.35 0.30 Burned Volume / Bore^3 0.25

70 60 50

0.20
40

0.15 0.10 0.05 0.00 0 0.1 0.2 0.3 0.4 0.5 0.6

30 20 10 0

0.7

Flame Radius / Bore

1.2 1.0 Flame Area / Bore^2 0.8 0.6 0.4 0.2 0.0 0 0.1 0.2 0.3 0.4

70 60 50 40 30 20 10 0

0.5

0.6

0.7

Flame Radius / Bore

Figure N.9: Flame geometry at varying crank angle (BTDC) for the test engine. Pent roof chamber, CR = 10. (Upper): Burned volume vs. flame radius. (Lower): Flame area vs. flame radius.

N.8 Laminar and Turbulent Flames

Before the calculation of flame speeds and turbulence levels is discussed, it is important to define laminar and turbulent flames and the turbulence structure, which have been discussed intensively in the literature [2, 30, 80, 98, 134, 149, 160]. The laminar flames surface is smooth and has a relatively thin reaction zone. The velocity at which the flame propagates into a non-turbulent, premixed, unburned mixture ahead of the flame is termed the laminar burning velocity (SL) [135, 211].

465

In IC engines, turbulent flames usually evolve from laminar flames. The smooth flame surface (laminar flame) can become wrinkled in the presence of turbulence (Figure N.10), which causes the reaction zone to become thicker. This wrinkled flame is known as the turbulent flame. the laminar flame speed [92]. Consequently, depending on the turbulence intensity, the turbulent flame speed (ST) is several times larger than

turbulent flame contour mean flame contour

burnt zone

laminar flame thickness

turbulent flame thickness Gibson-scale unburnt zone local curvature

Integral length scale (approx. 2mm)

Taylor microscale (approx. 0.7mm)

Low Reynolds Number

Kolmogorov scale (approx. 0.03mm)

High Reynolds Number

Figure N.10: (Upper): Schematic of wrinkled turbulent flame structure [99]. (Lower): Turbulent structure of jet during intake [98, 99].

N.8.1 Laminar Flame Speed

In theory, once the flame kernel is stabilized after ignition, it propagates away from the spark plug at the laminar flame speed (SL). In practice however, due to the expansion of the plasma and the conductive energy, the flame kernel needs to reach a radius of 10 mm before it can become fully independent of the spark

466

energy [34, 80, 134]. At this stage the flame kernel size is significant enough so that both large and small scale turbulence may distort the surface by wrinkling, thereby increasing the flame speed until it reaches the fully developed turbulent flame speed. James [110] has reviewed the most popular laminar burning velocity correlations. The most commonly used relationship in combustion modeling for laminar burning velocity is calculated by using the experimental correlations of Metghalchi and Keck [149]. The most advanced analytic approximations for n-heptane and isooctane were developed by Muller at al. [160]. The approximations are reproduced in the following form:

S L = FY
where:

m F ,u

T T T 0 exp(G / T ) u0 b T Tb Tu
0

(N.81)

F, m, G and n are fitting coefficients, Tb = adiabatic flame

temperature, T = inner layer temperature, Y = 1/AFR.
0

N.8.2 Actual Flame Speed

Predictions of SL and ST in the combustion chamber as a function of pressure, temperature and turbulence intensity have been outlined by Hamori [92]. However, the actual flame speed in the combustion chamber was calculated from experimental data using the following formula.

ST , a =
where:

m u A

(N.82)

= unburned gas entrained into the flame front, A = equivalent m

spherical flame surface area, u = unburned gas.

The flame speed due to the expansion of burned gas can be calculated with the following formula:

ut = u g + ST
where the gas expansion velocity ug behind the flame front is:

(N.83)

ug =

u b m u A u
467

(N.84)

N.8.3 Flame Speed Ratio

The most accepted explanation for the increase in flame speed due to turbulence considers the effect of turbulent eddies on a scale larger than the thickness of the flame front. These eddies are assumed to have no effect on the local flame velocity (laminar flame speed) but do distort the flame front, increasing its area (Figure N.10). Consequently, the increase in flame speed is proportional to the increased area of the flame [134], as first suggested by Damkohler in the following form:

FSR =
where:

ST A = T S L AL

(N.85)

FSR = flame speed ratio, ST = turbulent flame speed, SL = laminar

flame speed, AT = wrinkled flame area, AL = area of the smooth laminar flame.

N.8.4 Turbulence Intensity

Turbulence is a complex phenomenon that greatly affects the surface of the flame. The magnitude of turbulence intensity inside the combustion chamber is dependent on many variables including the chamber geometry, the intake ports and runners and the number and size of the intake valves. Data derived from experimental measurements are therefore of preference. Lancaster [132] has previously documented turbulence measurements, with his experimental data used extensively for modeling and comparison [2, 30, 86, 211]. Due to the strong linear relationship between turbulence intensity, MAP, and engine speed, a simple linear regression model was used to estimate the turbulence intensity at 45 BTDC, assuming the same behavior exists in the test engine.

u ' 0 = 1.7 + (((((0.0012 N 0.2267) 1.57 / 1.57) + (((0.0083264 MAP + 0.94107069) 1.7) / 1.7) + (((0.03783582 CR + 1.1319092) 1.46) / 1.46)) 1.7)
where: (N.86)

u0 = turbulence intensity at 45 BTDC, N = engine speed (rev/min),

MAP = manifold pressure (kPa), CR = compression ratio.

468

N.9 Knock Amplitude

A second order high pass Butterworth digital filter developed by Rogers [190] was used to extract the knocking signal from the raw pressure trace. The core filter module was adopted, however I/O functions were changed to suit the combustion model. The input data sampling frequency was 20 kHz, which allowed the monitoring of first and second order pressure oscillations in the engine, occurring in the region of ~5-10 kHz respectively. The Butterworth filter has been used extensively in the literature to describe the severity of the pressure oscillation [40, 51, 92, 125, 137]. is defined as the peak knock amplitude. Hence, the knock severity can be quantified using the maximum value of the high pass filter, which The peak knock amplitude is used extensively in the literature to define the knock intensity, with limits used to describe and predict abnormal phenomena [16, 192].

469

470

Appendix O
P O W E R E D B Y W A T T A R D

Error Analysis

O.1 Measurement and Calculation Error

Table O.1 comprises the independent measurements that were undertaken during experiments together with test ranges and associated absolute errors [28]. The errors displayed are a combination of the human error in measurement and the precision of the measurement device. Table O.2 shows the error propagation when calculating experimental results from independent measurements.
Table O.1: Independent measurements, test ranges and associated errors. Independent Measurement
(A) Engine swept volume (B) Ambient pressure (Po) (C) Wet bulb temperature (Twb) (D) Dry bulb temperature (Tdb) (E) Engine speed (N) (F) Raw brake torque (G) MAP (H) MAT (I) Fuel mass (J) Time (K) Lambda ( ) Emissions: (L) CO (M) (N) (O) CO2 HC NOX Table M.1 Table M.1

Range
433.8 mL 100 - 102 kPa 9 - 32C 12 - 35C 2,000 - 11,000 rev/min 5 - 88 Nm 55 - 270 kPa 12 - 80C 20 - 150 g 30 - 300 s 0.7 - 1.2

Absolute Error across the Range

0.5 mL 0.07 kPa 0.5C 0.5C 10 rev/min 0.1 - 0.6 Nm 2 kPa 0.5C 1g 0.5 s 0.02

471

Table O.2: Error propagation from Table O.1. Calculated Result

Brake torque Brake power BMEP Fuel mass flow rate (fuel) Air mass flow rate (air) BSFC BSCO BSCO2 BSHC BSNOX

Dependent Measurement (Table O.1)

B, C, D, F B, C, D, E, F A, B, C, D, F I, J I, J, K B, C, D, E, F, I, J B, C, D, E, F, L B, C, D, E, F, M B, C, D, E, F, N B, C, D, E, F, O

Average Uncertainty across the Range

2.0% 2.3% 2.1% 1.5% 3.5% 3.8% 5.3% 5.2% 6.0% 5.3%

Figure O.1 displays in-cylinder pressure traces over 40 consecutive cycles with average pressure and standard deviation overlaid. The quality of the pressure data is highlighted, with near zero standard deviation in areas which are largely unaffected by combustion. The maximum standard deviation in pressure reading reaches 0.4 MPa due to the combustion variability, corresponding to a low 1.1% CoV of IMEP. This gives high confidence in the pressure data used throughout the thesis.

7 Average Pressure 6 Pressure (MPa) 5 4 3 2 1 0 -160 Standard Deviation

1.0 0.8 0.6 0.4 0.2 0.0 160

CoV of IMEP = 1.1%

-120

-80

-40

0 40 3 ) Volume (cm Crank Angle (deg.)

80

120

Figure O.1: In-cylinder pressure traces for 40 consecutive cycles with average pressure and standard deviation highlighted. NA - Carburetion, 10500 rev/min, WOT, CR = 13.

472

Standard Deviation (MPa)

Appendix P
P O W E R E D B Y W A T T A R D

Drawing Registry

473

Registry Index

WATTARD: REGISTRY
A A

DESCRIPTION
B

DRAWING #
0000 0001 - 0099 0101 - 2299
B

WATTARD ENGINE TEST MODES ASSEMBLIES PARTS

REGISTRY INDEX
TEST MODE & ASSEMBLIES ASSEMBLY & PARTS

SHEET #
Index 1 Index 1-2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

REGISTRY
5 6

A4

WATTARD: INDEX
A A

TEST MODE & ASSEMBLIES INDEX

SHEET # WATTARD ENGINE Mode A (NA - CARBS): Exploded View 00 Mode B (NA - PFI): Exploded View Mode D (TC - PFI): Exploded View Test Mode & Assemblies
B

SHEET # CLUTCH A (NA): Exploded View B (TC): Exploded View Assembly & Parts 05 CRANKCASE Exploded View Assembly & Parts 06 CRANKCASE BOLT PATTERN Top / Bottom View Assembly & Parts 07 CRANKSHAFT Exploded View Assembly & Parts 08 CYLINDER HEAD Exploded View Assembly & Parts 09 CYLINDER HEAD COVERS Exploded View Assembly & Parts 10 ENGINE COVERS Exploded View Assembly & Parts 4.1 4.2 4.3 5.1 5.2 6.1 6.2 7.1 7.2 8.1 8.2 9.1 9.2 10.1 10.2
C B

0.1 0.2 0.3 0.4

04

ASSEMBLY & PARTS INDEX

SHEET # 01
C

BARREL / PISTON Exploded View Assembly & Parts BREATHER COVER / OIL PAN A (WET SUMP): Exploded View B (DRY SUMP): Exploded View Assembly & Parts CAMSHAFT / TENSIONER Exploded View Assembly & Parts

1.1 1.2 2.1 2.2 2.3 3.1 3.2

02

03

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY Index 1 of 2
5 6

A4

WATTARD: INDEX
A A

SHEET # EXHAUST SYSTEM A (PLENUM MANIFOLD - NA): Exploded View 11 B (Y MANIFOLD - NA): Exploded View C (PLENUM MANIFOLD - NA): Exploded View D (Y MANIFOLD- TC): Exploded View
B

SHEET # 16 IGNITION SYSTEM Exploded View Assembly & Parts LUBRICATION SYSTEM A (WET SUMP): Exploded View B (DRY SUMP): Exploded View Assembly & Parts SPECIAL JIGS & TOOLS Parts STARTER SYSTEM Exploded View Assembly & Parts 20 TRANSMISSION Exploded View Assembly & Parts WATER COOLING SYSTEM Engine: Exploded View A (DYNAMOMETER 2004): Exploded View 21 B (MUR-04 VEHICLE): Exploded View C (DYNAMOMETER 2003): Schematic Diagram D (MUR-03 VEHICLE): Schematic Diagram Assembly & Parts 16.1 16.2 17.1 17.2 17.3 18.1 19.1 19.2 20.1 20.2 21.1 21.2 21.3 21.4 21.5 21.6
C B

11.1 11.2 11.3 11.4 11.5 12.1 12.2 12.3 12.4 12.5 13.1 13.2 14.1 14.2 15.1 15.2

17

Assembly & Parts FUEL / INDUCTION SYSTEM RESTRICTOR: Exploded View 12 A (NA - CARBS): Exploded View B (NA - EFI): Exploded View C (TC - EFI): Exploded View Assembly & Parts 13 GEAR CHANGE DRUM / SHIFT FORKS Exploded View Assembly & Parts GEAR CHANGE MECHANISM Exploded View Assembly & Parts 15 GENERATOR Exploded View Assembly & Parts

18

19

14

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY Index 2 of 2
5 6

A4

WATTARD: 00 : WATTARD ENGINE: MODE A (NA - CARBS) (EXPLODED VIEW)

A A

0016

0001

0021

0008

0009

0012A
B

0003 0017A 0015

0019

0011A
C C

0020 0004A 0013 0014

D

0005

0002A

0007

0010
D

CAD FILE: REGISTRY 0.1 COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

0.1
6

A4

WATTARD: 00 : WATTARD ENGINE: MODE B (NA - PFI) (EXPLODED VIEW)

A A

0016

0001

0021

0008

0009

0012B

0003

0015

0019

0011A
C C

0020

0004A 0013 0014

D

0005

0002B

0007

0010

0017B
D

CAD FILE: REGISTRY 0.2 COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

0.2
6

A4

WATTARD: 00 : WATTARD ENGINE: MODE D (TC - PFI) (EXPLODED VIEW)

A A

0016

0001

0021

0008

0009 0012C

0003

0015

0019

0011C
C C

0020

0004B 0013 0014

D

0005

0002B

0007

0010

0017B
D

CAD FILE: REGISTRY 0.3 COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

0.3
6

A4

WATTARD: 00 : WATTARD ENGINE

A

(TEST MODE & ASSEMBLIES)

TEST MODE INDEX

MODE #
0000

DESCRIPTION
WATTARD engine

CAD FILE
0000A WATTARD engine Mode A (NA - CARBS) 0000B WATTARD engine Mode B (NA - PFI) 0000D WATTARD engine Mode D (TC - PFI)

DRAWING #
0000A 0000B 0000D
B

ASSEMBLIES INDEX
ASSY #
0001 0002 0003 0004
C

DESCRIPTION
Barrel / Piston Breather cover / Oil pan Camshaft / Tensioner Clutch Crankcase Crankcase bolt pattern Crankshaft Cylinder head Cylinder head covers Engine covers Exhaust system

CAD FILE
0001 Barrel / Piston Assy 0002A Breather cover / Oil pan Assy A (wet sump) 0002B Breather cover / Oil pan Assy B (dry sump) 0003 Camshaft / Tensioner Assy 0004A Clutch Assy A (NA) 0004B Clutch Assy B (TC) 0005 Crankcase Assy 0006 Crankcase bolt pattern Assy 0007 Crankshaft Assy 0008 Cylinder head Assy 0009 Cylinder head cover Assy 0010 Engine cover Assy 0011A Exhaust system Assy A (plenum manifold - NA) 0011B Exhaust system Assy B (y manifold - NA) 0011C Exhaust system Assy C (plenum manifold - TC) 0011D Exhaust system Assy D (y manifold - TC)

DRAWING #
0001 0002A 0002B 0003 0004A 0004B 0005 0006 0007 0008 0009 0010 0011A 0011B 0011C 0011D

0005 0006 0007 0008 0009 0010 0011

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

0.4A
6

A4

WATTARD: 00 : WATTARD ENGINE

A

(TEST MODE & ASSEMBLIES)

ASSEMBLIES INDEX
0012 Fuel / Induction Assy (restrictor) 0012A Fuel / Induction Assy A (NA - CARBS) 0012B Fuel / Induction Assy B (NA - PFI) 0012C Fuel / Induction Assy C (TC - PFI) 0013 Gear change drum / Shift forks Assy 0014 Gear change mechanism Assy 0015 Generator Assy 0016 Ignition system Assy 0017A Lubrication system Assy A (wet sump) 0017B Lubrication system Assy B (dry sump) 0018 Special jigs / Tools 0019 Starter Assy 0020 Transmission Assy 0021 Water cooling Assy (engine) 0021A Water cooling Assy A (dynamometer 2004) 0021B Water cooling Assy B (MUR-04 vehicle) 0021C Water cooling Assy C (dynamometer 2003) 0021D Water cooling Assy 0012 0012A 0012B 0012C 0013 0014 0015 0016 0017A 0017B 0018 0019 0020 0021 0021A 0021B 0021C 0021D

0012
B

Fuel / Induction system Gear change drum / Shift forks Gear change mechanism Generator Ignition system Lubrication system Special jigs / Tools Starter system Transmission

0013 0014 0015 0016 0017 0018 0019 0020

C

0021

Water cooling system

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

0.4B
6

A4

WATTARD: 01: BARREL / PISTON

A

(EXPLODED VIEW)

0103 2

0107 1

0108 1

0110 2 0109 2 0104 2 0106 8

0113 2

0116 8

0112 2
C

0105 8

0111 2

0118 4

0114 4 0117 4
D

0115 8

0102 2

0101 1
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

1.1
6

A4

WATTARD: 01: BARREL / PISTON

A

(ASSEMBLY & PARTS)

Barrel / Piston

ASSEMBLY INDEX

ASSY #
0001

DESCRIPTION

CAD FILE
0001 Barrel / Piston Assy

DRAWING #
0001

PART #

QTY
1 2 2 2 8 8 1 1 2 2 2

DESCRIPTION
Barrel-engine Dowell-hollow (barrel to case) Dowell-solid (barrel to head) Liner-cylinder Nut-flanged (barrel to case) Nut-flanged (barrel to head) O-ring (barrel to head oil ex) O-ring (barrel to head oil in) O-ring (barrel to head water) O-ring (liner to head) Pin-piston

CAD FILE
0101 Barrel-engine 0102 Dowell-hollow (barrel to case) 0103 Dowell-solid (barrel to head) 0104A Liner-cylinder (std) 0104B Liner-cylinder (+1.0) 0105 Nut-flanged (barrel to case) 0106 Nut-flanged (barrel to head) 0107 O-ring (barrel to head oil ex) 0108 O-ring (barrel to head oil in) 0109 O-ring (barrel to head water) 0110 O-ring (liner to head) 0111 Pin-piston 0112A Piston (std 12.9:1 CR Ring A) 0112B Piston (std 12.9:1 CR Ring B) 0112C Piston (std 12.9:1 CR Ring C) 0112D Piston (std 11:1 CR Ring D) 0112E Piston (std 9.6:1 CR Ring D) 0112F Piston (+1.0 10.0:1 CR Ring E) 0113A Rings-piston A (1.0 2.8) 0113B Rings-piston B (1.0 1.2 2.8) 0113C Rings-piston C (1.0 2.8) 0113D Rings-piston D (1.2 2.8) 0113E Rings-piston E (1.0 1.2 2.8) 0114 Ring-snap (piston pin) 0115 Stud (barrel to case) 0116 Stud (barrel to head) 0117 Washer-curved (inner barrel to case) 0118 Washer-flat (outer barrel to case)

DRAWING #
0101 0102 0103 0104A 0104B 0105 0106 0107 0108 0109 0110 0111 0112A 0112B 0112C 0112D 0112E 0112F 0113A 0113B 0113C 0113D 0113E 0114 0115 0116 0117 0118

PARTS INDEX
B

0101 0102 0103 0104 0105 0106 0107 0108 0109 0110 0111

0112
C

Piston

0113

Rings-piston

0114 0115 0116 0117 0118

D

4 8 8 4 4

Ring-snap (piston pin) Stud (barrel to case) Stud (barrel to head) Washer-curved (inner barrel to case) Washer-flat (outer barrel to case)

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

1.2
6

A4

WATTARD: 02: BREATHER COVER / OIL PAN A (WET SUMP)

(EXPLODED VIEW)
0203 1

0211 1

0213 1

0206 1
B

0208 1 0207 1

0205A 1

0210A 1
C

0202 10

0209 1 0212 1
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

2.1
6

A4

WATTARD: 02: BREATHER COVER / OIL PAN B (DRY SUMP)

(EXPLODED VIEW)
0203 1

0211 1

0213 1

0206 1

0208 1
B

0201B 1

0207 1

0205B 1 0204 8 0210B 1 0202 10 0209 1 0212 1

D D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

2.2
6

A4

WATTARD: 02: BREATHER COVER / OIL PAN

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0002

DESCRIPTION
Breather cover / Oil pan

CAD FILE
0002A Breather cover / Oil pan Assy A (wet sump) 0002B Breather cover / Oil pan Assy B (dry sump)

DRAWING #
0002A 0002B

PARTS INDEX
PART #
0201 0202 0203 0204 0205

QTY
1 10 1 8 1 1 1 1 1 1 1 1 1 Baffle-oil

DESCRIPTION
Bolt-BHSS (oil pan) Bolt-flanged (breather housing) Bolt-SHCS (oil baffle) Gasket (oil pan) Housing-breather O-ring (breather housing) O-ring (breather housing bolt) O-ring (sump plug) Pan-oil Pipe-breather Plug-sump Washer-sealing (breather pipe)

CAD FILE
0201A Baffle-oil (wet sump) 0201B Baffle-oil (dry sump) 0202 Bolt-BHSS M6x12 (oil pan) 0203 Bolt-flanged M6x25 (breather housing) 0204 Bolt-SHCS M3x10 (oil baffle) 0205A Gasket (wet sump) 0205B Gasket (dry sump) 0206 Housing-breather 0207 O-ring 92x1.7 (breather housing) 0208 O-ring 6x1.7 (breather housing bolt) 0209 O-ring 9x2 (sump plug) 0210A Pan-oil (wet sump) 0210B Pan-oil (dry sump) 0211 Pipe-breather 0212 Plug-sump 0213 Washer-sealing (breather pipe)

DRAWING #
0201A 0201B 0202 0203 0204 0205A 0205B 0206 0207 0208 0209 0210A 0210B 0211 0212 0213

0206 0207 0208 0209 0210 0211 0212 0213

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

2.3
6

A4

WATTARD: 03: CAMSHAFT / TENSIONER

A

(EXPLODED VIEW)

0302 4

0312 2

0302 4
B

0305 1 0303 2

0312 2 0304 1 0306 1 0309 1

C

0313 1
C

0311 1 0307 1 0308 1 0310 1 0301 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

3.1
6

A4

WATTARD: 03: CAMSHAFT / TENSIONER

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0003

DESCRIPTION
Camshaft / Tensioner

CAD FILE
0003 Camshaft / Tensioner Assy

DRAWING #
0003

PARTS INDEX
PART #
0301 0302 0303 0304 0305 0306 0307 0308 0309 0310 0311 0312 0313

QTY
2 4 2 1 1 1 1 1 1 1 1 2 1

DESCRIPTION
Bolt-CHSS (chain guide #0310) Bolt-flanged (camshaft sprocket) Bolt-SHCS (chain tensioner) Camshaft-exhaust Camshaft-inlet Chain-camshaft Dowell-solid (chain guide #0308) Guide-chain (slack side) Guide-chain (tight side) Guide-chain (tight side lateral) O-ring (chain tensioner) Sprocket-camshaft Tensioner-chain

CAD FILE
0301 Bolt-CHSS M3x10 (chain guide #0310) 0302 Bolt-flanged M7x12 (camshaft sprocket) 0303 Bolt-SHCS M6x20 (chain tensioner) 0304 Camshaft-exhaust 0305 Camshaft-inlet 0306 Chain-camshaft 0307 Dowell-solid 8x40 (chain guide #0308) 0308 Guide-chain (slack side) 0309 Guide-chain (tight side) 0310 Guide-chain (tight side lateral) 0311 O-ring 18x2 (chain tensioner) 0312 Sprocket-camshaft 0313 Tensioner-chain

DRAWING #
0301 0302 0303 0304 0305 0306 0307 0308 0309 0310 0311 0312 0313

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

3.2
6

A4

WATTARD: 04: CLUTCH A (NA)

A

(EXPLODED VIEW)

0404 5 0419A 5

0409 1
A

0412A 1 0402 1 0415A 1 0410 1 0420 1

0411A 7

0413A 6

0417 1

0401A 1 0405 1

0406 1
D

0408A 1

0418 1

0407 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

4.1
6

A4

WATTARD: 04: CLUTCH B (TC)

A

(EXPLODED VIEW)

0409 1 0404 5
A

0416 5 0412A 1

0419B 5

0415B 1
B

0402 1

0411B 8 0413B 7 0417 1

B

0410 1 0420 1

0403 12

0414 1 0401B 1 0405 1

0406 1
D

0408B 1

0418 1

0407 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

4.2
6

A4

WATTARD:
A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0004 Clutch

04: CLUTCH

DESCRIPTION

CAD FILE
0004A Clutch Assy A (NA) 0004B Clutch Assy B (TC)

DRAWING #
0004A 0004B

PART #
0401 0402 0403 0404 0405 0406 0407 0408 0409 0410 0411 0412
C

QTY
1 1 12 5 1 1 1 1 1 1 AR 1 AR 1 1 5 1 1 5 1 Basket

DESCRIPTION
Bearing (operating rod) Bolt-BHSS (stiffening ring) Bolt-flanged (clutch spring) Bush (basket to input shaft) Chain-primary Gear (oil pump drive) Hub-centre Lever-clutch Nut-hub Plate-friction Plate-pressure Plate-steel Ring-stiffening (basket) Rod-operating Shim (clutch spring) Spacer (hub to basket) Spacer (oil pump drive gear) Spring-clutch Washer (hub nut)

CAD FILE
0401A Basket (clutch A) 0401B Basket (clutch B) 0402 Bearing 28x12x8 (operating rod) 0403 Bolt-BHSS M3x8 (stiffening ring) 0404 Bolt-flanged M6x18 (clutch spring) 0405 Bush (basket to input shaft) 0406 Chain-primary 0407 Gear (oil pump drive) 0408A Hub-centre (clutch A) 0408B Hub-centre (clutch B) 0409 Lever-clutch 0410 Nut-hub 0411A Plate-friction (clutch A) 0411B Plate-friction (clutch B) 0412A Plate-pressure 0412B Plate-pressure (billet) 0413A Plate-steel (clutch A) 0413B Plate-steel (clutch B) 0414 Ring-stiffening (basket) 0415A Rod-operating (clutch A) 0415B Rod-operating (clutch B) 0416 Shim (clutch spring) 0417 Spacer (hub to basket) 0418 Spacer (oil pump drive gear) 0419A Spring-clutch (standard stiffness) 0419B Spring-clutch (1.5 x stiffer) 0419C Spring-clutch (2 x stiffer) 0420 Washer (hub nut)

DRAWING #
0401 0402 0403 0404 0405 0406 0407 0408A 0408B 0409 0410 0411A 0411B 0412A 0412B 0413A 0413B 0414 0415A 0415B 0416 0417 0418 0419A 0419B 0419C 0420

PARTS INDEX

0413 0414 0415 0416 0417 0418 0419 0420

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

4.3
6

A4

WATTARD: 05: CRANKCASE

A

0525 1

0503 1

0504 2 0524 5 0511 1

A

(EXPLODED VIEW)

0513 2 0513 2

0522 2 0515 2 0509 1 0505 3 0508 1 0521 1 0520 1 0507 2 0519 2 0512 2 0506 2
C

0515 2 0523 2 0510 1

0517 1 0507 2

0506 2

0518 1

0502 1

0501 1 0524 5
D

0516 1

0514 1

0525 1

0512 2
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

5.1
6

A4

WATTARD: 05: CRANKCASE

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0005 Crankcase

DESCRIPTION

CAD FILE
0005 Crankcase Assy

DRAWING #
0005

PARTS INDEX
PART #
B

QTY
1 1 1 2 3 2 2 1 1 1 1 2 2 1 2 1 1 1 2 1 1 2 2 5 1

DESCRIPTION
Bolt-banjo (clutch oil feed line to case) Bolt-banjo (clutch oil feed line to clutch) Bolt-banjo (head oil feed line to case) Bolt-banjo (head oil feed line to head) Bolt-flanged (chain guide) Dowell-solid (crankcase alignment) Dowell-solid (transmission endfloat alignment) Guide-chain (primary slack side) Guide-chain (primary tight side) Line-oil (clutch feed) Line-oil (cylinder head feed) Mount-engine (exhaust side) Mount-engine (intake side) O-ring (main oil gallery plug) O-ring (nylon plug) Plug (main oil gallery) Plug-nylon (balance shaft cyl 1) Plug-nylon (balance shaft cyl 2) Ring (transmission alignment) Seal-lip (transmission output shaft) Seal-rubber (transmission input shaft) Washer-sealing (10mm banjo bolt) Washer-sealing (6mm banjo bolt) Washer-sealing (8mm banjo bolt) Crankcase

CAD FILE
0501 Bolt-banjo M8 (clutch oil feed line to case) 0502 Bolt-banjo M6 (clutch oil feed line to clutch) 0503 Bolt-banjo M10 (head oil feed line to case) 0504 Bolt-banjo M8 fine (head oil feed line to head) 0505 Bolt-flanged M6x16 (chain guide) 0506 Dowell-solid 10x12 (crankcase alignment) 0507 Dowell-solid 6x16 (transmission endfloat alignment) 0508 Guide-chain (primary slack side) 0509 Guide-chain (primary tight side) 0510 Line-oil (clutch feed) 0511 Line-oil (cylinder head feed) 0512 Mount-engine (exhaust side) 0513 Mount-engine (intake side) 0514 O-ring 11x2 (main oil gallery plug) 0515 O-ring 27x2 (nylon plug) 0516 Plug (main oil gallery) 0517 Plug-nylon (balance shaft cyl 1) 0518 Plug-nylon (balance shaft cyl 2) 0519 Ring (transmission alignment) 0520 Seal-lip 52x40x8 (transmission output shaft) 0521 Seal-rubber (transmission input shaft) 0522 Washer-sealing (10mm banjo bolt) 0523 Washer-sealing (6mm banjo bolt) 0524 Washer-sealing (8mm banjo bolt) 0525 Crankcase

DRAWING #
0501 0502 0503 0504 0505 0506 0507 0508 0509 0510 0511 0512 0513 0514 0515 0516 0517 0518 0519 0520 0521 0522 0523 0524 0525
B

0501 0502 0503 0504 0505 0506 0507 0508 0509 0510 0511 0512 0513 0514 0515 0516 0517 0518 0519 0520 0521 0522 0523 0524 0525

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

5.2
6

A4

WATTARD: 06: CRANKCASE BOLT PATTERN

A

(TOP / BOTTOM VIEW)

0605 4

0606 8

0601 5

0604 1

0602 3

0601 5

0603 4

0605 4

0606 8

LOWER CASE

UPPER CASE

0601 5

0602 3

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

6.1
6

A4

WATTARD: 06: CRANKCASE BOLT PATTERN

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0006

DESCRIPTION
Crankcase bolt pattern

CAD FILE
0006 Crankcase bolt pattern Assy

DRAWING #
0006

PARTS INDEX
PART #
0601 0602 0603 0604 0605 0606

QTY
5 3 4 1 4 8

DESCRIPTION
Bolt-flanged (crankcase) Bolt-flanged (crankcase upper to lower) Bolt-flanged (crankcase upper to lower) Bolt-flanged (crankcase lower to upper) Bolt-flanged (crankcase upper to lower) Bolt-flanged (crankcase upper to lower)

CAD FILE
0601 Bolt-flanged M6x40 (crankcase) 0602 Bolt-flanged M6x60 (crankcase upper to lower) 0603 Bolt-flanged M6x80 (crankcase upper to lower) 0604 Bolt-flanged M8x55 (crankcase lower to upper) 0605 Bolt-flanged M8x75 (crankcase upper to lower) 0606 Bolt-flanged M8x100 (crankcase upper to lower)

DRAWING #
0601 0602 0603 0604 0605 0606

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

6.2
6

A4

WATTARD: 07: CRANKSHAFT

(EXPLODED VIEW)
A A

0703 4 0706 2 0708 4

B

0702 4

0702 4

0703 4 0705 2

0701 4

0701 4
C

0707 1
C

0708 4 0704 4 0704 4 0709 1 0703 4

D

0703 4

0702 4 0702 4
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

7.1
6

A4

WATTARD: 07: CRANKSHAFT

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0007 Crankshaft

DESCRIPTION

CAD FILE
0007 Crankshaft Assy

DRAWING #
0007

PARTS INDEX
PART #
0701

QTY
4

DESCRIPTION
Bearing (connecting rod shell)

CAD FILE
0701A Bearing (connecting rod shell- std) 0701B Bearing (connecting rod shell- +0.25) 0701C Bearing (connecting rod shell- +0.50) 0702A Bearing-main (inner shell- BLUE) 0702B Bearing-main (inner shell- BLACK) 0702C Bearing-main (inner shell- BROWN) 0703A Bearing-main (outer shell- BLUE) 0703B Bearing-main (outer shell- BLACK) 0703C Bearing-main (outer shell- BROWN) 0704 Bolt-ARP 5~16UNF (connecting rod) 0705 Bolt-SSS 5~16UNFx8 (crank oil gallery) 0706 Connecting-rod (assy) 0707A Crankshaft (100%rot 0%rec) 0707B Crankshaft (100%rot 40%rec) 0707C Crankshaft (100%rot 50%rec) 0707D Crankshaft (100%rot 50%rec even fire) 0708 Dowell-hollow (connecting rod) 0709 Key-woodruff

DRAWING #
0701A 0701B 0701C 0702A 0702B 0702C 0703A 0703B 0703C 0704 0705 0706 0707A 0707B 0707C 0707D 0708 0709

0702

Bearing (main inner shell)

0703
C

4 4 2 2 1 4 1

Bearing (main outer shell) Bolt-ARP (connecting rod) Bolt-SSS (crank oil gallery) Connecting-rod (assy) Crankshaft Dowell-hollow (connecting rod) Key-woodruff

0704 0705 0706 0707 0708 0709

D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

7.2
6

A4

WATTARD: 08: CYLINDER HEAD

A

(EXPLODED VIEW)
0805 8 0805 8 0803 1

0802 1

0804 1

0811 2

0806 4 0801 1 0809 4 0811 2 0807 8 0815 8

C

0810 0810 11 0807 8 0815 8 0808 16 0816 8 0814 8 0813 8 0814 8 0812 8 0816 8 0813 8
C

0808 16 0812 8

0817 4

0818 4

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

8.1
6

A4

WATTARD: 08: CYLINDER HEAD

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0008 Cylinder head

DESCRIPTION

CAD FILE
0008 Cylinder head Assy

DRAWING #
0008

PARTS INDEX
B

PART #
0801 0802 0803 0804 0805 0806 0807 0808 0809 0810 0811 0812 0813 0814 0815 0816 0817 0818

QTY
1 1 1 1 8 4 8 16 4 1 2 8 8 8 8 8 4 4

DESCRIPTION
Bearing-camshaft (A) Bearing-camshaft (B) Bearing-camshaft (C) Bearing-camshaft (D) Bolt-flanged (camshaft bearing) Bolt-flanged (oil line) Bucket (valve actuation) Collet Dowell-hollow (camshaft bearing) Head-cylinder Line-oil (camshaft bearing) Retainer (valve spring) Seal (valve stem) Seat (valve spring) Shim (tappet adjustment) Spring-valve Valve (exhaust) Valve (intake)

CAD FILE
0801 Bearing-camshaft (A) 0802 Bearing-camshaft (B) 0803 Bearing-camshaft (C) 0804 Bearing-camshaft (D) 0805 Bolt-flanged M6x45 (camshaft bearing) 0806 Bolt-flanged M6x40 (oil line) 0807A Bucket (valve actuation- std) 0807B Bucket (valve actuation- +1.0) 0808 Collet 0809 Dowell-hollow 8mm (camshaft bearing) 0810 Head-cylinder 0811 Line-oil (camshaft bearing) 0812 Retainer (valve spring) 0813 Seal (valve stem) 0814 Seat (valve spring) 0815 Shim (tappet adjustment) 0816 Spring-valve 0817 Valve (exhaust) 0818 Valve (intake)

DRAWING #
0801 0802 0803 0804 0805 0806 0807A 0807B 0808 0809 0810 0811 0812 0813 0814 0815 0816 0817 0818

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

8.2
6

A4

WATTARD: 09: CYLINDER HEAD COVERS

A

(EXPLODED VIEW)

0902 1

0913 4

0911 2

0905 6 0906 4 0906 4 0905 6 0908 4 0904 4 0901 4 0915 1 0912 2 0909 2 0910 2 0909 2

0911 2 0901 4 0913 4 0905 6 0906 4

0907 1

0917 3

0903 3

0916 1

0906 4 0908 4 0904 13 0914 1 0905 6

0905 6

0905 6

0912 2

0913 4

0910 2
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

9.1
6

A4

WATTARD: 09: CYLINDER HEAD COVERS

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0009

DESCRIPTION
Cylinder head covers

CAD FILE
0009 Cylinder head cover Assy

DRAWING #
0009

PARTS INDEX
PART #
0901 0902 0903 0904 0905 0906 0907 0908 0909 0910 0911 0912 0913 0914 0915 0916 0917

QTY
4 1 3 13 6 4 1 4 2 2 2 2 4 1 1 1 3

DESCRIPTION
Bearing-ball (side plate) Bolt-BHSS (cam sensor) Bolt-flanged (rocker cover) Bolt-SHCS (side plate to head) Bolt-SHCS (side plate to rocker cover) Bolt-SHCS (side plate to rocker cover) Cover-rocker Dowell-soild (side plate to head) O-ring (rocker cover spark plug hole) O-ring (rocker cover to head) O-ring (side plate to head for oil) O-ring (side plate to head for water) O-ring (side plate to head oil gallery) Plate-side (cylinder 1) Plate-side (cylinder 2) Sensor-hall (camshaft syncronisation) Washer-rubber (rocker cover bolt)

CAD FILE
0901 Bearing-ball 32x20x7 (side plate) 0902 Bolt-BHSS M6x40 (cam sensor) 0903 Bolt-flanged M7 (rocker cover) 0904 Bolt-SHCS M4x10 (side plate to head) 0905 Bolt-SHCS M3x10 (side plate to rocker cover) 0906 Bolt-SHCS M4x15 (side plate to rocker cover) 0907 Cover-rocker 0908 Dowell-soild 4x16 (side plate to head) 0909 O-ring (rocker cover spark plug hole) 0910 O-ring (rocker cover to head) 0911 O-ring (side plate to head for oil) 0912 O-ring (side plate to head for water) 0913 O-ring (side plate to head oil gallery) 0914 Plate-side (cylinder 1) 0915 Plate-side (cylinder 2) 0916 Sensor-hall (camshaft syncronisation) 0917 Washer-rubber (rocker cover bolt)

DRAWING #
0901 0902 0903 0904 0905 0906 0907 0908 0909 0910 0911 0912 0913 0914 0915 0916 0917

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

9.2
6

A4

WATTARD: 10: ENGINE COVERS

(EXPLODED VIEW)

1015 1

1006 1

1018 1

1003 11

1011 6 1004 8 1017 1

1013 1

1001 2
B

1001 2

1007 1
C

1008 1

1005 1 1016 1 1009 1 1014 1

1011 6

1012 1

1020 1

1002 6

1019 1

1021 1

1010 1

1011 6

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

10.1
6

A4

WATTARD: 10: ENGINE COVERS

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0010 Engine covers

DESCRIPTION

CAD FILE
0010 Engine cover Assy

DRAWING #
0010

PARTS INDEX
B

PART #
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021

QTY
2 6 11 8 1 1 1 1 1 1 6 1 1 1 1 1 1 1 1 1 1

DESCRIPTION
Bearing-needle (clutch lever) Bolt-flanged (shift cover) Bolt-SHCS (clutch cover) Bolt-SHCS (generator cover) Cap-centre (generator cover) Cap-oil (clutch cover) Cap-timing (generator cover) Cover-clutch Cover-generator Cover-shift Dowell-hollow (cover alignment) Gasket (clutch cover) Gasket (generator cover) Gasket (shift cover) O-ring (clutch cover oil cap) O-ring (generator cover centre cap) O-ring (generator cover timing cap) Seal (clutch lever) Seal (gear selector shaft) Sensor-hall (crankshaft syncronisation) Switch (neutral position) 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021

CAD FILE
Bearing-needle (clutch lever) Bolt-flanged M6x20 (shift cover) Bolt-SHCS M6X20 (clutch cover) Bolt-SHCS M6X25 (generator cover) Cap-centre (generator cover) Cap-oil (clutch cover) Cap-timing (generator cover) Cover-clutch Cover-generator Cover-shift Dowell-hollow 8mm (cover alignment) Gasket (clutch cover) Gasket (generator cover) Gasket (shift cover) O-ring 32.5x2.5 (clutch cover oil cap) O-ring 40x3 (generator cover centre cap) O-ring 17x2.5 (generator cover timing cap) Seal (clutch lever) Seal (gear selector shaft) Sensor-hall (crankshaft syncronisation) Switch (neutral position)

DRAWING #
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

10.2
6

A4

WATTARD: 11: EXHAUST SYSTEM A (PLENUM MANIFOLD - NA)

A

(EXPLODED VIEW)

1115 1 1114A 1

1104A 3

1109 1

1117 3

1111 3 1101 2

1110A 1 1102 2 1103 2 1106 2 1108 2

C

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

11.1
6

A4

WATTARD: 11: EXHAUST SYSTEM B (Y MANIFOLD - NA)

A

(EXPLODED VIEW)

1115 1 1114A 1

1104A 3

1109 1

1117 3 1111 3 1110C 1 1101 2 1103 2 1102 2 1106 2

B

1108 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

11.2
6

A4

WATTARD: 11: EXHAUST SYSTEM C (PLENUM MANIFOLD - TC)

A

(EXPLODED VIEW)
1104B 4 1115 1

1117 7 1116 1
B

1114B 1 1109 1 1117 7 1111 3 1101 2 1110B 1 1102 2

B

1112 1 1113 1 1107 1 1105 2

1103 2 1106 2 1108 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

11.3
6

A4

WATTARD: 11: EXHAUST SYSTEM D (Y MANIFOLD - TC)

(EXPLODED VIEW)
A A

1115 1 1104B 4 1117 7 1116 1

B

1114B 1 1109 1 1117 7 1111 3 1101 2 1110C 1 1102 2 1103 2 1106 2

C B

1112 1 1113 1 1107 1 1105 2

1108 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

11.4
6

A4

WATTARD: 11: EXHAUST SYSTEM

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0011 Exhaust system

DESCRIPTION
0011A Exhaust 0011B Exhaust 0011C Exhaust 0011D Exhaust system system system system

CAD FILE
Assy Assy Assy Assy A (plenum manifold - NA) B (Y manifold - NA) C (plenum manifold - TC) D (Y manifold - TC)

DRAWING #
0011A 0011B 0011C 0011D

PARTS INDEX
PART #
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117

QTY
2 2 2 AR 2 2 1 2 1 1 3 1 1 1 1 1 AR

DESCRIPTION
Adaptor-EGA (exhaust gas sampling) Bolt-HEX (exhaust manifold to top head) Bolt-SHCS (exhaust manifold to bottom head) Bolt-SHCS (outlet pipe) Bolt-SHCS (turbo oil flange) Flange (cylinder head to exhaust manifold) Flange (turbo oil) Gasket (cylinder head to exhaust manifold) Gasket (exhaust manifold outlet) Manifold-exhaust Nut-plain (outlet pipe) O-ring (turbo oil drain) O-ring (turbo oil supply) Pipe-outlet Sensor-oxygen Turbocharger-GT12 Washer-spring (outlet pipe)

CAD FILE
1101 Adaptor-EGA (exhaust gas sampling) 1102 Bolt-HEX M8x25 (exhaust manifold to top head) 1103 Bolt-SHCS M8x25 (exhaust manifold to bottom head) 1104A Bolt-SHCS M8x20 (outlet pipe- NA) 1104B Bolt-SHCS M8x15 (outlet pipe- TC) 1105 Bolt-SHCS M6x12 (turbo oil flange) 1106 Flange (cylinder head to exhaust manifold) 1107 Flange (turbo oil) 1108 Gasket (cylinder head to exhaust manifold) 1109 Gasket (exhaust manifold outlet) 1110A Manifold-exhaust (plenum- NA) 1110B Manifold-exhaust (plenum- TC) 1110C Manifold-exhaust (Y- NA TC) 1111 Nut-plain M8 (outlet pipe- NA) 1112 O-ring 13x1.5 (turbo oil drain) 1113 O-ring 5x1.5 (turbo oil supply) 1114A Pipe-outlet (NA) 1114B Pipe-outlet (TC) 1115 Sensor-oxygen 1116 Turbocharger-GT12 1117 Washer-spring M8 (outlet pipe)

DRAWING #
1101 1102 1103 1104A 1104B 1105 1106 1107 1108 1109 1110A 1110B 1110C 1111 1112 1113 1114A 1114B 1115 1116 1117

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

11.5
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM (RESTRICTOR)

A

(EXPLODED VIEW)
1232 1

1225 1

1209 4

1237 1

1208 2

1220 2

1207 2

1211B 1

1214B 1

1220 2

1207 2

1238 1

1210 1

1226 1

1201 1

1204 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.1
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM A (NA - CARBS)

A

(EXPLODED VIEW)

1206A 4

1234A 1

1223A 2 1214A 2

1211A 2

1215A 2

1212 2

1206A 4

1233A 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.2
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM B (NA - PFI)

A

(EXPLODED VIEW)
1212 2 1234B 1 1206B 4 1224 2 1227A 1

1236 1

1228 2

1222 2

RESTRICTOR ASSY
B

1206B 4

1235 1 1203 1 1219 AR

1229 1

1217 AR 1230 1 1213 2 1217 AR 1231 1 1221 4 1215B 2

1219 AR 1233B 1 1223B 2 1221 4

1205 4

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.3
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM C (TC - PFI)

A

(EXPLODED VIEW)
1239 1 1218 AR 1212 6

1219 AR

1202 1

1212 6

1212 6 1234C 1 1206B 4

B

RESTRICTOR ASSY

1206B 4

1212 6 1218 AR 1227B 1

1229 1
C

1203 1 1217 AR 1230 1 1213 2 1205 4 1217 AR 1233C 1 1216 2 1223C 2 1219 AR 1219 AR 1235 1
C

1236 1

1231 1

1221 4

1215B 2

1221 4

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.4
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM

A

(ASSEMBLY & PARTS)

DESCRIPTION

ASSEMBLY INDEX
ASSY # CAD FILE
0012 Fuel / Induction Assy (restrictor) 0012A Fuel / Induction Assy A (NA - CARBS) 0012B Fuel / Induction Assy B (NA - PFI) 0012C Fuel / Induction Assy C (TC - PFI)

DRAWING #
0012 0012A 0012B 0012C

0012

Fuel / Induction system

PARTS INDEX
PART #
1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216

QTY
1 1 1 2 4 4 2 2 4 1 1 AR 2 AR 2 2

DESCRIPTION
Actuator-throttle Adaptor-vacuum (plenum to bleed valve) Adaptor-vacuum (plenum to MAP to regulator) Bolt-pan (butterfly) Bolt-SHCS (fuel rail to manifold) Bolt-SHCS (inlet runner to head) Bolt-SHCS (throttle start stop) Bolt-SHCS (TP sensor) Bolt-SSS (restrictor to intake) Butterfly Clamp-jubilee (air filter) Clamp-jubilee (manifold connections) Clip (injector to rail) Filter-air Fuel-metering Gasket-tufnol (insulator plate)

CAD FILE
1201 Actuator-throttle 1202 Adaptor-vacuum (plenum to bleed valve) 1203 Adaptor-vacuum (plenum to MAP to regulator) 1204 Bolt-pan M3x6 (butterfly) 1205 Bolt-SHCS M4x10 (fuel rail to manifold) 1206A Bolt-FHSS M6x20 (inlet runner to head- carbs) 1206B Bolt-SHCS M6x20 (inlet runner to head- PFI) 1207 Bolt-SHCS M3x16 (throttle start stop) 1208 Bolt-SHCS M4x20 (TP sensor) 1209 Bolt-SSS M4x6 (restrictor to intake) 1210 Butterfly 1211A Clamp-jubilee (air filter- carbs) 1211B Clamp-jubilee (air filter- PFI) 1212 Clamp-jubilee (manifold connections) 1213 Clip (injector to rail) 1214A Filter-air (carbs) 1214B Filter-air (PFI) 1215A Fuel-metering (carburettor) 1215B Fuel-metering (fuel injector) 1216 Gasket-tufnol (insulator plate)

DRAWING #
1201 1202 1203 1204 1205 1206A 1206B 1207 1208 1209 1210 1211A 1211B 1212 1213 1214A 1214B 1215A 1215B 1216

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.5A
6

A4

WATTARD: 12: FUEL / INDUCTION SYSTEM

A

(ASSEMBLY & PARTS)

PARTS INDEX
1217 1218 1219 1220 1221 1222 1223 AR AR AR 2 4 2 2 2 1 1 1 2 1 1 1 1 1 Hose-braided (#10 with fittings) Hose-silicon (manifold connections) Hose-vacuum Nut-plain O-ring (fuel injector) O-ring (injector plug) O-ring (inlet runner to head) O-ring (inlet runner to plenum) O-ring (restrictor to intake) Pin-roll (butterfly shaft) Plenum Plug-injector Pump-fuel Rail-fuel Regulator (fuel pressure) Restrictor Runner-inlet (cyl 1) 1217 Hose-braided (speedflow #10 with fittings) 1218 Hose-silicon (manifold connections) 1219 Hose-vacuum 3~16 1220 Nut-plain M3 1221 O-ring 8x3.5 (fuel injector) 1222 O-ring 10x2 (injector plug) 1223A O-ring 40x2.5 (inlet runner to head- NA CARBS) 1223B O-ring 38x1.5 (inlet runner to head- NA PFI) 1223C O-ring 42x2 (inlet runner to head- TC PFI) 1224 O-ring 38x1.5 (inlet runner to plenum) 1225 O-ring 46x2 (restrictor to intake) 1226 Pin-roll 3~32 (butterfly shaft) 1227A Plenum (NA PFI) 1227B Plenum (TC PFI) 1228 Plug-injector 1229 Pump-fuel 1230 Rail-fuel 1231 Regulator (fuel pressure) 1232 Restrictor 1233A Runner-inlet (cyl 1- NA CARBS) 1233B Runner-inlet (cyl 1- NA PFI) 1233C Runner-inlet (cyl 1- TC PFI) 1234A Runner-inlet (cyl 2- NA CARBS) 1234B Runner-inlet (cyl 2- NA PFI) 1234C Runner-inlet (cyl 2- TC PFI) 1235 Sensor-MAP 1236 Sensor-MAT 1237 Sensor-TP 1238 Shaft-butterfly 1239 Valve-bleed (turbo actuator) 1217 1218 1219 1220 1221 1222 1223A 1223B 1223C 1224 1225 1226 1227A 1227B 1228 1229 1230 1231 1232 1233A 1233B 1233C 1234A 1234B 1234C 1235 1236 1237 1238 1239

1224 1225 1226 1227 1228 1229 1230 1231 1232

C

1233

1234 1235 1236 1237 1238 1239

D

1 1 1 1 1 1

Runner-inlet (cyl 2) Sensor-MAP Sensor-MAT Sensor-TP Shaft-butterfly Valve-bleed (turbo actuator)

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

12.5B
6

A4

WATTARD: 13: GEAR CHANGE DRUM / SHIFT FORKS

A

(EXPLODED VIEW)
1306 1 1307 1

1303 3
B

1301 1

1304 1 1305 1 1302 1

1303 3
C C

1310 1 1309 1 1308 2

1303 3
D

1308 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

13.1
6

A4

WATTARD: 13: GEAR CHANGE DRUM / SHIFT FORKS

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0013
B

DESCRIPTION
Gear change drum / Shift forks

CAD FILE
0013 Gear change drum / Shift forks Assy

DRAWING #
0013
B

PARTS INDEX
PART #
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310

QTY
1 1 3 1 1 1 1 2 1 1

DESCRIPTION
Bearing-ball (selector drum) Bolt-FHSS (gear change cam) Bolt-flanged (drum bearing) Cam (gear change) Contact (neutral position switch) Dowell-solid (gear change cam) Drum-selector Fork-shift Rod-shift Washer-flat (shift rod)

CAD FILE
1301 Bearing-ball 42x25x9 (selector drum) 1302 Bolt-FHSS M6x35 (gear change cam) 1303 Bolt-flanged M6x12 (drum bearing) 1304 Cam (gear change) 1305 Contact (neutral position switch) 1306 Dowel-solid 4mm (gear change cam) 1307 Drum-selector 1308 Fork-shift 1309 Rod-shift 1310 Washer-flat 6.4x24x1.2 (shift rod)

DRAWING #
1301 1302 1303 1304 1305 1306 1307 1308 1309 1310

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

13.2
6

A4

WATTARD: 14: GEAR CHANGE MECHANISIM

A

(EXPLODED VIEW)

1411 1

1402 1

1412 1

1413 1 1409 1 1410 1 1403 1 1404 1 1401 1 1405 1 1407 1

1406 1

1408 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

14.1
6

A4

WATTARD: 14: GEAR CHANGE MECHANISM

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0014

DESCRIPTION
Gear change mechanism

CAD FILE
0014 Gear change mechanism Assy

DRAWING #
0014

PARTS INDEX
PART #
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413

QTY
1 1 1 1 1 1 1 1 1 1 1 1 1

DESCRIPTION
Clip-E (gear change shaft) Collar (gear change shaft) Lever (gear position) Lever (pawl) Nut-nyloc (gear position lever) Shaft (gear change) Spacer (gear position lever) Spring-return (gear change shaft) Spring-return (gear position lever) Spring-return (paul lever) Stud (gear change shaft) Stud (gear position lever) Washer-plain (gear position lever)

CAD FILE
1401 Clip-E 12mm (gear change shaft) 1402 Collar 13x18.3x21.5 (gear change shaft) 1403 Lever (gear position) 1404 Lever (pawl) 1405 Nut-nyloc M6 (gear position lever) 1406 Shaft (gear change) 1407 Spacer (gear position lever) 1408 Spring-return (gear change shaft) 1409 Spring-return (gear position lever) 1410 Spring-return (paul lever) 1411 Stud M8 (gear change shaft) 1412 Stud M6 (gear position lever) 1413 Washer-plain 6mm (gear position lever)

DRAWING #
1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

14.2
6

A4

WATTARD: 15: GENERATOR

A

(EXPLODED VIEW)

1503 1 1501 1 1502 5 1504 1

1505 1

1502 5

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

15.1
6

A4

WATTARD: 15: GENERATOR

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0015
B

DESCRIPTION
Generator

CAD FILE
0015 Generator Assy

DRAWING #
0015
B

PARTS INDEX
PART #
1501 1502 1503 1504 1505

QTY
1 5 1 1 1

DESCRIPTION
Bolt-flanged (crankshaft snout) Bolt-SHCS (voltage regulator and stator) Flywheel Regulator-voltage Stator

CAD FILE
1501 Bolt-flanged M10LHx36 (crankshaft snout) 1502 Bolt-SHCS M6x25 (voltage regulator and stator) 1503 Flywheel 1504 Regulator-voltage 1505 Stator

DRAWING #
1501 1502 1503 1504 1505

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

15.2
6

A4

WATTARD: 16: IGNITION SYSTEM

A

(EXPLODED VIEW)

1601A 2

1604 2

1601A 2

1603 1

1604 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

16.1
6

A4

WATTARD: 16: IGNITION SYSTEM

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0016

DESCRIPTION
Ignition system

CAD FILE
0016 Ignition system Assy

DRAWING #
0016

PARTS INDEX
PART #
1601 1602 1603
C

QTY
2 AR 1 2 Coil-ignition

DESCRIPTION
Lead-ignition Module-CDI Plug-spark

CAD FILE
1601A Coil-ignition (on plug) 1601B Coil-ignition (external) 1602 Lead-ignition 1603A Module-CDI (50mJ) 1603B Module-CDI (100mJ) 1604A Spark-plug (NA) 1604B Spark-plug (TC)

DRAWING #
1601A 1601B 1602 1603A 1603B 1604A 1604B

1604

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

16.1
6

A4

WATTARD: 17: LUBRICATION SYSTEM A (WET SUMP)

A

(EXPLODED VIEW)
1745 1

1715 1 1714 1 1743A 1 1701A 4 1741A 1

1742 1

1748A 1

1752 1

1731 1 1718A 1

1747 1

1708 1

1713 1

1702A 3

1740 1

1735 2 1721 1 1734 1 1738 2 1738 2 1725 1

C B

1710 1

1751A 1 1753 1 1733 1 1735 2

1746A 1
D

1705A 2

1706 1

1704 4

1722A 1

1732 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

17.1
6

A4

WATTARD: 17: LUBRICATION SYSTEM B (DRY SUMP) (EXPLODED VIEW)

A

1716 1
B

1717 1

1752 2

1719 1

1709 2

1709 2

1749 1

OIL & SCAVENGE PUMP

1702B 3
B

1720 1 1744 1

1701B 1 1708 1 1713 1 1718B 1 1705B 4 1746B 1 1740 2 1723 1 1743B 1 1742 2 1742 2 1748B 1 1714 1 1715 1 1752 2 1705B 4
D

1740 2 1741B 1

1703 4

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

17.2A
6

A4

WATTARD: 17: LUBRICATION SYSTEM B (DRY SUMP) (EXPLODED VIEW)

A A

1707 1 1739 1 1750 1 1712 1

B

1724 1 1727 1 1731 1

B

1730 2 1737 1 1728 1 1706 3

1751B 1

1747 1

1732 1 1734 1

1733 1

1735 2 1721 1
C

OIL OUT

1726 1

OIL SCAVENGE PUMP

OIL IN

1710 1

1735 2 1738 2
D

1725 1

1706 3

1738 2

1736 1

1730 2

1729 1

1711 1

1704 4

1722B 1
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

17.2B
6

A4

WATTARD: 17: LUBRICATION SYSTEM

A

(ASSEMBLY & PARTS)

A

ASSEMBLY INDEX
ASSY #
0017 Lubrication system

DESCRIPTION

CAD FILE
0017A Lubrication system Assy A (wet sump) 0017B Lubrication system Assy B (dry sump)

DRAWING #
0017A 0017B

PARTS INDEX
PART #
1701 1702
B

QTY
AR AR 4 4 AR AR 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

DESCRIPTION
Bolt (oil pump back plate) Bolt (pump to casing) Bolt-SHCS (filter screen) Bolt-SHCS (manifold oil line) Bolt-SHCS (pump inlet) Bolt-SHCS (pump oulet) Cap (oil tank) Circlip-external (pump shaft) Dowell-hollow (oil pump to scav pump) Filter-oil Fitting (oil in) Fitting (oil out) Gear (pump drive) G-rotor (oil pump inner) G-rotor (oil pump outer) G-rotor (scav pump inner) G-rotor (scav pump outer) Housing (oil pump) Housing (scav pump) Housing-extension (scav pump) Line-oil (manifold to head) Line-oil (manifold) Line-oil (oil in fitting to oil pump) Line-oil (oil out fitting to oil tank)

CAD FILE
1701A Bolt-PAN M6x16 (oil pump back plate- wet sump) 1701B Bolt-SHCS M6x60mod (oil pump back plate- dry sump) 1702A Bolt-flanged M6x25 (pump to casing- wet sump) 1702B Bolt-SHCS M6x90 (pump to casing- dry sump) 1703 Bolt-SHCS M3x10 (filter screen) 1704 Bolt-SHCS M3x20 (manifold oil line) 1705A Bolt-SHCS M6x20 (pump inlet- wet sump) 1705B Bolt-SHCS M6x25mod (pump inlet- dry sump) 1706 Bolt-SHCS M6x12 (pump oulet) 1707 Cap (oil tank) 1708 Circlip-external 11mm (pump shaft) 1709 Dowell-hollow 8x16 (oil pump to scav pump) 1710 Filter-oil 1711 Fitting (oil in) 1712 Fitting (oil out) 1713 Gear (pump drive) 1714 G-rotor (oil pump inner) 1715 G-rotor (oil pump outer) 1716 G-rotor (scav pump inner) 1717 G-rotor (scav pump outer) 1718A Housing (oil pump- wet sump) 1718B Housing (oil pump- dry sump) 1719 Housing (scav pump) 1720 Housing-extension (scav pump) 1721 Line-oil (manifold to head) 1722A Line-oil (manifold- wet sump) 1722B Line-oil (manifold- dry sump) 1723 Line-oil (oil in fitting to oil pump) 1724 Line-oil (oil out fitting to oil tank)

DRAWING #
1701A 1701B 1702A 1702B 1703 1704 1705A 1705B 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718A 1718B 1719 1720 1721 1722A 1722B 1723 1724

1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

17.3A
6

A4

WATTARD: 17: LUBRICATION SYSTEM

A

(ASSEMBLY & PARTS)

PARTS INDEX
1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1 1 1 1 1 2 1 1 1 1 2 1 1 2 1 AR 1 AR 1 1 1 1 1 1 1 1 1 AR 1 Line-oil (oil pump to oil filter) Line-oil (oil tank to oil in fitting) Line-oil (pressure relief valve to oil tank) Line-oil (scav pump to oil out fitting) Nut-plain (oil in fitting) O-ring (case to fitting) O-ring (main gallery plug) O-ring (manifold centre port) O-ring (manifold cyl head port) O-ring (manifold oil pump port) O-ring (manifold to head oil line) O-ring (oil in fitting) O-ring (oil out fitting) O-ring (oil pump to oil filter line) O-ring (oil tank cap) O-ring (pickup) Pickup-oil Pin (pump shaft) Plate-back (oil pump) Plate-back (scav pump) Plug (main gallery) Screen-filter (oil pickup) Sensor (oil pressure) Shaft-pump Spacer (scav pump to oil pump) Tank-oil Valve-relief (pressure) Washer (pump housing to G-rotor) Washer-sealing (wet sump pressure relief valve) 1725 Line-oil (oil pump to oil filter) 1726 Line-oil (oil tank to oil in fitting) 1727 Line-oil (pressure relief valve to oil tank) 1728 Line-oil (scav pump to oil out fitting) 1729 Nut-plain 7~8UNF (oil in fitting) 1730 O-ring 22x1.5 (case to fitting) 1731 O-ring 19x1.5 (main gallery plug) 1732 O-ring 12.5x1.8 (manifold centre port) 1733 O-ring 9x1.8 (manifold cyl head port) 1734 O-ring 10x2 (manifold oil pump port) 1735 O-ring 10x2.5 (manifold to head oil line) 1736 O-ring 16x1.5 (oil in fitting) 1737 O-ring 15x1.5 (oil out fitting) 1738 O-ring 10x2.5 (oil pump to oil filter line) 1739 O-ring 32.5x2.5 (oil tank cap) 1740 O-ring 23x2.5 (pickup) 1741A Pickup-oil (wet sump) 1741B Pickup-oil (dry sump) 1742 Pin 4mm (pump shaft) 1743A Plate-back (oil pump- wet sump) 1743B Plate-back (oil pump- dry sump) 1744 Plate-back (scav pump) 1745 Plug (main gallery) 1746A Screen-filter (oil pickup- wet sump) 1746B Screen-filter (oil pickup- dry sump) 1747 Sensor (oil pressure) 1748A Shaft-pump (wet sump) 1748B Shaft-pump (dry sump) 1749 Spacer (scav pump to oil pump) 1750 Tank-oil 1751A Valve-relief (pressure- wet sump) 1751B Valve-relief (pressure- dry sump) 1752 Washer 15x11.2x1 (pump housing to G-rotor) 1753 Washer-sealing (wet sump pressure relief valve) 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741A 1741B 1742 1743A 1743B 1744 1745 1746A 1746B 1747 1748A 1748B 1749 1750 1751A 1751B 1752 1753

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

17.3B
6

A4

WATTARD: 18: SPECIAL TOOLS AND EQUIPMENT

A

(PARTS)

PARTS INDEX
B

PART #
1801 1802 1803 1804 1805 1806 1807 1808 1809 1810

QTY
1 1 2 1 1 1 1 1 1 1

DESCRIPTION
Adaptor-flowbench Adaptor-spanner (barrel to case bolts) Bobweight-balancing (crankshaft) Jig-machining (camshaft bearing) Jig-welding (inlet manifold) Plate-torque (barrel to case) Plate-torque (head to barrel) Tool-removal (flywheel) Tool-removal (generator cover) Tool-removal (valve spring)

CAD FILE
1801A Adaptor-flowbench (65.5mm bore) 1801B Adaptor-flowbench (69mm bore) 1801C Adaptor-flowbench (74mm bore) 1802 Adaptor-spanner (barrel to case bolts) 1803 Bobweight-balancing (crankshaft) 1804 Jig-machining (camshaft bearing) 1805 Jig-welding (inlet manifold) 1806 Plate-torque (barrel to case) 1807 Plate-torque (head to barrel) 1808 Tool-removal (flywheel) 1809 Tool-removal (generator cover) 1810 Tool-removal (valve spring)

DRAWING #
1801A 1801B 1801C 1802 1803 1804 1805 1806 1807 1808 1809 1810

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

18.1
6

A4

WATTARD: 19: STARTER SYSTEM

A

(EXPLODED VIEW)

1903 2 1907 1 1908 1 1904 1 1913 1

A

1912 3

1909 3

1910 3

1905 1

1911 1

1902 3
D

1914 1

1901 2

1906 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

19.1
6

A4

WATTARD: 19: STARTER SYSTEM

A

(ASSEMBLY & PARTS)

ASSEMBLY INDEX
ASSY #
0019

DESCRIPTION
Starter system 0019 Starter Assy

CAD FILE

DRAWING #
0019

PARTS INDEX
PART #
1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914

QTY
2 3 2 1 1 1 1 1 3 3 1 3 1 1

DESCRIPTION
Bolt-flanged (starter chain guide) Bolt-SHCS (sprag clutch) Bolt-SHCS (starter to casing) Chain-starter Clutch-sprag Guide (starter chain) Motor-starter O-ring (starter motor to casing) Pin (sprag clutch) Roller (sprag clutch) Spacer (sprag clutch) Spring (sprag clutch) Sprocket (sprag clutch) Sprocket (starter motor)

CAD FILE
1901 Bolt-flanged M6x20 (starter chain guide) 1902 Bolt-SHCS M8x20 (sprag clutch) 1903 Bolt-SHCS M6x25 (starter to casing) 1904 Chain-starter 1905 Clutch-sprag 1906 Guide (starter chain) 1907 Motor-starter 1908 O-ring 42x2 (starter motor to casing) 1909 Pin (sprag clutch) 1910 Roller (sprag clutch) 1911 Spacer (sprag clutch) 1912 Spring (sprag clutch) 1913 Sprocket (sprag clutch) 1914 Sprocket (starter motor)

DRAWING #
1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

19.2
6

A4

WATTARD: 20: TRANSMISSION

A

(EXPLODED VIEW)

2017 3 2010 1 2025 2 2003 1

2017 3

2021 1

2011 1

2022 1

2017 3

2019 1

2002 2 2023 1 2006 1 2014 1

2026 2

2018 2

2020 1

2004 1 2025 2 2012 1 2005 2 2026 2 2001 3 2007 1

C

2016 1

2024 1

2009 1

2015 1

2002 2

2013 1

2008 1

2018 2

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

20.1
6

A4

WATTARD: 20: TRANSMISSION

A

(ASSEMBLY & PARTS)

ASSY #
0020 Transmission

ASSEMBLY INDEX PARTS INDEX

DESCRIPTION

CAD FILE
0020 Transmission Assy

DRAWING #
0020

PART #
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026

QTY
3 2 1 1 2 1 1 1 1 1 1 1 1 1 1 1 3 2 1 1 1 1 1 1 2 2

DESCRIPTION
Ball-steel (nuetral finding) Bearing-ball (input and output shaft) Bearing-needle (input shaft) Bearing-needle (output shaft) Bolt-SHCS (sprocket plate) Bush (3rd output gear) Cog (1-3 output gear) Cog (2nd output gear) Collar (output shaft) Gear-input (2nd 18T) Gear-input (3nd 24T) Gear-output (1st 36T) Gear-output (2nd 32T) Gear-output (3rd 27T) O-ring (output shaft) Plate-sprocket Ring-snap (input shaft) Ring-snap (output shaft) Shaft-input Shaft-output Spacer (input shaft) Spacer (input shaft) Spacer (output shaft) Sprocket Washer-flat (needle bearing) Washer-toothed (output shaft)

CAD FILE
2001 Ball-steel 5~32 (nuetral finding) 2002 Bearing-ball (input and output shaft) 2003 Bearing-needle (input shaft) 2004 Bearing-needle (output shaft) 2005 Bolt-SHCS M6x8 (sprocket plate) 2006 Bush (3rd output gear) 2007 Cog (1-3 output gear) 2008 Cog (2nd output gear) 2009 Collar 40x25x12 (output shaft) 2010 Gear-input (2nd 18T) 2011 Gear-input (3nd 24T) 2012 Gear-output (1st 36T) 2013 Gear-output (2nd 32T) 2014 Gear-output (3rd 27T) 2015 O-ring 2x27 (output shaft) 2016 Plate-sprocket 2017 Ring-snap 24mm (input shaft) 2018 Ring-snap 25.9mm (output shaft) 2019 Shaft-input 2020 Shaft-output 2021 Spacer 29x25x27.5 (input shaft) 2022 Spacer 29x25x15 (input shaft) 2023 Spacer 37x30x10 (output shaft) 2024A Sprocket (14 tooth) 2024B Sprocket (15 tooth) 2024C Sprocket (16 tooth) 2024D Sprocket (17 tooth) 2025 Washer-flat 30x20x2 (needle bearing) 2026 Washer-toothed 33.5x28.3x1 (output shaft)

DRAWING #
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2022 2021 2023 2024A 2024B 2024C 2024D 2025 2026

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

20.2
6

A4

WATTARD: 21: WATER COOLING SYSTEM (ENGINE)

A

(EXPLODED VIEW)

2110B 2

2112 2

2118B 2
B

2112 2 2114 1 2111 2

2113 1 2110B 2

2102 6

2109B 1

2118B 2

2102 6
D

2111 2

2120B 1

ENGINE
D

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.1
6

A4

WATTARD: 21: WATER COOLING SYSTEM A (DYNAMOMETER 2004)

A

(EXPLODED VIEW)
2108D 1 2101B 1

2108G 2 2103 1 2121A 1

B

2108E 2 2108E 2

2108G 2 ENGINE 2108D 2

C

WATER OUT WATER IN

2117 1 2108D 2 2101E 2

2115B 1

2119 1

2108F 1

2101D 1

2105 1

2108H 1

2101C 1

2122 1

2107 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.2
6

A4

WATTARD: 21: WATER COOLING SYSTEM B (MUR-04 VEHICLE)

A

(EXPLODED VIEW)

2103 1 2108G 2 2121B 1 2108D 2

ENGINE

2108G 2 WATER OUT WATER IN 2108D 2

2106 2

2101E 2 2115B 1 2106 2 2116B 1 2101E 2 2108I 1

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.3
6

A4

WATTARD: 21: WATER COOLING SYSTEM C (DYNAMOMETER 2003)

A

(SCHEMATIC DIAGRAM)

2108E

2108E
B B

CYLINDER HEAD Water Temp 2120A 2108C Water out 2110A ADAPTOR 2101A 2101A ADAPTOR 2108B
C

OPEN COOLING TOWER 2122

C

Water in 2109A BLOCK

2108A

BOOSTER PUMP 2115A

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.4
6

A4

WATTARD: 21: WATER COOLING SYSTEM D (MUR-03 VEHICLE)

A

(SCHEMATIC DIAGRAM)

Fill point, Cap pressurized 1BAR 2103 2108B

B B

CYLINDER HEAD 2106 Temp 2120A COOLING FAN COOLING FAN RADIATOR 2116A Overflow bottle

Water out 2110A

2106

2108A Water in 2109A BLOCK BOOSTER PUMP 2115A

2108B

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.5
6

A4

WATTARD: 21: WATER COOLING SYSTEM

A

(ASSEMBLY & PARTS)

DESCRIPTION

ASSEMBLY INDEX
ASSY #
0021

CAD FILE
0021 Water cooling Assy (engine) 0021A Water cooling Assy A (dynamometer 2004) 0021B Water cooling Assy B (MUR-04 vehicle) 0021C Water cooling Assy C (dynamometer 2003) 0021D Water cooling Assy (MUR-03 vehicle)

DRAWING #
0021 0021A 0021B 0021C 0021D
B

Water cooling

PARTS INDEX

PART#
2101

QTY
AR

DESCRIPTION
Adaptor-water

CAD FILE
2101A Adaptor-water (22-40mm) 2101B Adaptor-water (25-40mm) 2101C Adaptor-water (40mm U) 2101D Adaptor-water (40mm) 2101E Adaptor-water (EWP) 2102 Bolt-SHCS M3x10 (water inlet line to barrel) 2103 Cap-pressure 2104 Clamp-jubilee (water connections) 2105 Exchanger-heat (cooling tower) 2106 Fan-thermatic 2107 Heater-water (cooling tower) 2108A Hose-water (16mm) 2108B Hose-water (19mm) 2108C Hose-water (22mm) 2108D Hose-water (25mm) 2108E Hose-water (convoluted 40mm) 2108F Hose-water (40mm transparent) 2108G Hose-water (16mm L) 2108H Hose-water (40mm) 2108I Hose-water (25mm transparent)

DRAWING #
2101A 2101B 2101C 2101D 2101E 2102 2103 2104 2105 2106 2107 2108A 2108B 2108C 2108D 2108E 2108F 2108G 2108H 2108I

2102 2103 2104 2105 2106 2107

6 1 AR 1 2 1

Bolt-SHCS (water inlet line to barrel) Cap-pressure Clamp-jubilee (water connections) Exchanger-heat (cooling tower) Fan-thermatic Heater-water (cooling tower)

2108

AR

Hose-water

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.6A
6

A4

WATTARD: 21: WATER COOLING SYSTEM

A

(ASSEMBLY & PARTS)

PARTS INDEX
2109
B

1 AR 2 2 1 1 1 1 1 2 1 1 1 1

Line-water (inlet) Line-water (outlet) O-ring (water inlet line to barrel) O-ring (water outlet plug) Plug (water outlet cyl1) Plug (water outlet cyl2) Pump-water Radiator Rotameter-water Seal (water outlet line) Sensor-pressure (water) Sensor-temp (water) Tank-header Tower-cooling

2110 2111 2112 2113 2114 2115 2116 2117 2118

2119 2120 2121 2122

2109A Line-water (inlet- 2003) 2109B Line-water (inlet- 2004) 2110A Line-water (outlet- 2003) 2110B Line-water (outlet- 2004) 2111 O-ring 16x1.5 (water inlet line to barrel) 2112 O-ring 20x2.5 (water outlet plug) 2113 Plug (water outlet cyl1) 2114 Plug (water outlet cyl2) 2115A Pump-water (booster) 2115B Pump-water (EWP) 2116A Radiator (MUR-03 vehicle) 2116B Radiator (MUR-04 vehicle) 2117 Rotameter-water 2118A Seal (water outlet line- O-ring 19x3) 2118B Seal (water outlet line- washer) 2119 Sensor-pressure (water) 2120A Sensor-temp (water- VDO) 2120B Sensor-temp (water- Bosch) 2121A Tank-header (translucent) 2121B Tank-header (aluminium) 2122 Tower-cooling

2109A 2109B 2110A 2110B 2111 2112 2113 2114 2115A 2115B 2116A 2116B 2117 2118A 2118B 2119 2120A 2120B 2121A 2121B 2122

CAD FILE: REGISTRY COPYRIGHT: WILLIAM ATTARD +61 3 9367 2008 DATE: DRN: W.A
1 2

5/1/06

Department of Mechanical and Manufacturing Engineering

TEL +61 3 8344 7982 FAX +61 3 9347 8784
3 4

WATTARD

SHEET

REGISTRY
5

21.6B
6

A4