Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of combustion chamber design on turbulence, cyclic variation and performance in an SI engine Tippett, Esther Claire 1989

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1989_A7 T56.pdf [ 10.07MB ]
Metadata
JSON: 831-1.0080823.json
JSON-LD: 831-1.0080823-ld.json
RDF/XML (Pretty): 831-1.0080823-rdf.xml
RDF/JSON: 831-1.0080823-rdf.json
Turtle: 831-1.0080823-turtle.txt
N-Triples: 831-1.0080823-rdf-ntriples.txt
Original Record: 831-1.0080823-source.json
Full Text
831-1.0080823-fulltext.txt
Citation
831-1.0080823.ris

Full Text

T H E E F F E C T S O F C O M B U S T I O N C H A M B E R D E S I G N O N T U R B U L E N C E , C Y C L I C V A R I A T I O N A N D P E R F O R M A N C E I N A N SI E N G I N E B y Esther Claire Tippett B . E . M e c h (Hons) University of Canterbury, New Zealand. 1983 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF MECHANICAL ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August, 1989 © Esther Claire Tippett , 1989 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of Br i t i sh Columbia , I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Mechanical Engineering The University of Br i t i sh Columbia Vancouver, Canada Date: A B S T R A C T A n experimental program of motored and fired tests has been undertaken on a single cylinder spark ignition engine to determine the influence of combustion chamber design on turbulence enhancement i n the achievement of fast lean operation. Flow field measurements were taken using hot wire anemometry i n the cylinder during motored operation. O n line performance tests and in-cylinder pressure data were recorded for the operation of the engine by natural gas at lean and stoichiometric conditions over a range of speed and loads. Squish and squish jet action methods of turbulence enhancement were investigated for six configurations, using a standard bathtub cylinder head and new piston designs incorporating directed jets through a raised wall , a standard bowl-in-piston chamber and an original squish jet design piston. A non squish comparison was provided by a disc chamber. Peak Pressure and Indicated Mean Effective Pressure ( I M E P ) , two parameters char-acterizing performance and cyclic variability, showed that enhanced turbulence by com-bustion chamber geometry is effective in improving performance at lean operation. The single jet action directed towards the spark was most effective i n improving the efficiency at high speed and lean mixtures. The addition of jets to the single jet, or jet chan-nels to the main squish action of the bowl- in-piston chamber, reduced performance and increased cyclic variability. Mass fraction burn analysis of the cylinder pressure data showed that squish action was most effective i n the main burn period. Configurations wi th large squish area and centrally located spark produced the greatest reduction in both the in i t i a l and main burn i i periods. The potential for the squish jet action to improve engine drivabil ity and increase the knock l imi t was exhibited in reduced coefficient of variance of I M E P and reduced ignition advance requirements. Directions for further research to exploit this potential for engines operated by alternative fuels are identified. i i i Table of Contents A B S T R A C T ii List of Tables viii List of Figures xii Nomenclature xxii Acknowledgments xxvi 1 Introduction 1 1.1 Introduction and Background 1 1.2 Objective of this study 4 1.3 Turbulent F low F ie ld i n an Engine 4 1.4 Discussion of Terminology i n Turbulence Studies 6 1.5 Scope of Work 8 1.6 Structure of thesis 10 2 Literature Review 11 2.1 Introduction 11 2.2 Turbulence studies in Engines 12 2.3 Combustion Studies in Engines 16 2.4 Combustion Chamber Design 24 3 Experimental Apparatus and Method 27 iv 3.1 Introduction 27 3.2 Experimental Apparatus 28 3.2.1 Introduction . . . 28 3.2.2 Engine Bed . 29 3.2.3 Combustion Chambers 30 3.2.4 Modif ied bathtub pistons 32 3.2.5 Instrumentation 34 3.2.6 Data Acquis i t ion 36 3.3 Motored Engine Tests 38 3.3.1 Introduction 38 3.3.2 Operational Procedures 39 3.3.3 Pressure Measurements 41 3.3.4 Hotwire Measurements 42 3.4 F i red Engine Tests 44 3.4.1 Introduction 44 3.4.2 Operational Procedures 45 3.4.3 Performance Measurements 46 3.4.4 Pressure Measurements 46 4 Data Analysis 48 4.1 Introduction 48 4.2 Motored Engine Tests 48 4.2.1 Ana ly t i ca l Procedure 48 4.2.2 Pressure Signal Processing 49 4.2.3 Anemometer Signal Processing 53 4.2.4 F low F ie ld Data Analysis 55 v 4.3 F i red Engine Tests 58 4.3.1 Ana ly t i ca l Procedure 58 4.3.2 Pressure Signal Processing 59 4.3.3 Performance Data Analysis 59 4.3.4 Combustion Analysis 62 5 Experimental Results and Discussion 66 5.1 Introduction 66 5.2 Motored Tests 66 5.2.1 Motored Pressure Results 66 5.2.2 Flow F ie ld Results 68 5.3 F ired Tests 72 5.3.1 General Performance Parameters 72 5.3.2 F ired Pressure Results 75 5.3.3 Mass Fraction Burned Results 80 5.4 Experimental Uncertainties and Technique 86 5.4.1 Flow Measurement 86 5.4.2 Performance Measurements 88 5.5 Turbulence, Combustion and Performance 91 6 Conclusions and Recommendations 94 6.1 Introduction 94 6.2 Conclusions 95 6.2.1 Turbulence Studies 95 6.2.2 Performance and Combustion Studies 96 6.3 Recommendations 97 vi Bibliography 99 Appendices 190 A Instrument Specification and Calibration 190 B Hot W i r e Anemometry Specification and Calibration 198 C B C Natural Gas Properties 202 D Pressure Filtering Methods 206 vii List of Tables 3.1 Ricardo Hydra Gasoline (or Gaseous fuel) Engine Specifications 104 3.2 Motored operating conditions for Pressure and Hotwire measurements at WOT 105 3.3 Fired operating conditions for Pressure measurements at MBT and Full Load (WOT) 105 3.4 Fired operating conditions for Pressure measurements at MBT and Part Load 105 4.1 Motored data Analysis program flow chart 106 4.2 Fired data Analysis program flow charts 107 5.1 Compression and Expansion coefficients for the motored condition. . . . 108 5.2 Engine performance as per SAEJ1349 for different piston geometries for stoichiometric and lean RAFR at MBT, WOT, and 20.0 rps 109 5.3 Engine performance as per SAEJ1349 for different piston geometries for stoichiometric and lean RAFR at MBT, WOT, and 33.3 rps 110 5.4 Engine performance as per SAEJ1349 for different piston geometries for stoichiometric and lean RAFR at MBT, WOT, and 50.0 rps 110 5.5 Engine performance as per SAEJ1349 for different piston geometries for stoichiometric and lean RAFR at MBT, 2.5 bmep, and 33.3 rps Ill 5.6 Engine performance as per SAEJ1349 for different piston geometries for stoichiometric and lean RAFR at MBT, 3.5 bmep, and 50.0 rps Ill viii 5.7 Ignition Advance and Brake Thermal Efficiency for different piston geome-tries for R A R F w l . 0 0 - 1 . 3 5 at M B T , W O T , and 20.0 rps 112 5.8 Ignition Advance and Brake Thermal Efficiency for different piston geome-tries for R A R F « 1 . 0 0 - 1 . 3 5 at M B T , W O T , and 33.3 rps 113 5.9 Ignition Advance and Brake Thermal Efficiency for different piston geome-tries for R A R F « 1 . 0 0 - 1 . 3 5 at M B T , W O T , and 50.0 rps 113 5.10 Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , W O T , and 20.0 rps 114 5.11 Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , W O T , and 33.3 rps 115 5.12 Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean operation at, M B T , W O T , and 50.0 rps 116 5.13 Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , 2.5 bmep, and 33.3 rps 117 5.14 Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , 3.5 bmep, and 50.0 rps 118 5.15 Init ia l (0-01% massburned) and M a i n (01-90% massburned) combustion durations for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 20.0 rps 119 IX 5.16 Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 20.0 rps 120 5.17 Init ial (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 33.3 rps 121 5.18 Init ial (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 33.3 rps 122 5.19 Init ia l (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 50.0 rps 123 5.20 Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 50.0 rps 124 5.21 Init ial (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 2.5 bmep, and 33.3 rps 125 5.22 Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean at R A F R M B T , 2.5 bmep, and 33.3 rps 126 5.23 Init ial (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 3.5 bmep, 50.0 rps 127 x 5.24 Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 3.5 bmep, and 50.0 rps 127 A . l Pressure transducer specifications for Kist ler model 6121 191 A . 2 Pressure transducer calibration data for Kist ler model 6121. No. 317205 . 192 B . l Hot wire anemometry equipment and specifications 199 B . 2 Hotwire calibration data for wire No. 1, yielding the calibration constants: A=0.1575; £ = 0 . 4 9 0 8 ; and n=0.360 200 C. l Composit ion of B C Natura l Gas 202 C.2 Molecular weight of B C Natura l gas 203 C.3 Higher and Lower Heating values of B C Natura l gas 203 C.4 Viscosity calculations for B C Natural gas 204 xi List of Figures 3.1 Ricardo H y d r a engine, dynamometer and control systems layout 128 3.2 Ricardo H y d r a M K I I I Gasoline (or gaseous fuel) Engine cross-sectional and longitudinal views 129 3.3 Standard bathtub combustion chamber geometry 130 3.4 Single slot and castellated piston geometries 131 3.5 Bowl-in-piston and squish jet piston geometries 132 3.6 Instrumentation layout for the Ricardo engine test cell 133 3.7 Hot wire probe location through the spark plug entry for the bathtub and flat cylinder heads 134 3.8 Acquis i t ion hardware arrangement wi th fast pressure data acquisition hook up 135 4.1 Comparison of the effect of window size on the turbulent intensity profile for the bathtub chamber at W O T , 33.3 rps 136 4.2 Polytropic coefficent calculated from the ensembled pressure for single slot chamber at M B T , W O T and 33.3 rps for R A F R = 1 . 2 7 137 5.1 Motored pressure profiles for different chamber geometries at W O T , 20.0 rps. 138 5.2 Motored pressure profiles for different chamber geometries at W O T , 33.3 rps. 138 5.3 Motored pressure profiles for different chamber geometries at W O T , 50.0 rps. 139 5.4 Motored pressure profiles for different chamber geometries at W O T , 66.7 rps. 139 5.5 Motored pressure profiles for the single slot piston at W O T , for three speeds; 20.0, 33.3, and 50.0 rps 140 xii 5.6 Motored temperature profiles for different chamber geometries at W O T , 33.3 rps 140 5.7 Window ensembled and cycle ensembled mean velocity and turbulent in-tensity profiles for the single slot piston at W O T , 33.3 rps 141 5.8 Mean velocity profiles for different chamber geometries at W O T , 33.3 rps. 142 5.9 Turbulent intensity profiles for different chamber geometries at W O T , 33.3 rps • 142 5.10 Mean velocity profiles for the 'bathtub' group of chambers at W O T , 20.0 rps. 143 5.11 Turbulent intensity profiles for the 'bathtub' group of chambers at W O T , 20.0 rps 143 5.12 Mean velocity profiles for the 'bathtub' group of chambers at W O T , 33.3 rps. 144 5.13 Turbulent intensity profiles for the 'bathtub' group of chambers at W O T , 33.3 rps 144 5.14 Mean velocity profiles for the 'disc' group of chambers at W O T , 20.0 rps. 145 5.15 Turbulent intensity profiles for the 'disc' group of chambers at W O T , 20.0 rps 145 5.16 Mean velocity profiles for the 'disc' group of chambers at W O T , 33.3 rps. 146 5.17 Turbulent intensity profiles for the 'disc' group of chambers at W O T , 33.3 rps 146 5.18 Mean velocity profiles for the bathtub chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 147 5.19 Turbulent intensity profiles for the bathtub chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 147 5.20 Mean velocity profiles for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 148 xiii 5.21 Turbulent intensity profiles for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 148 5.22 Mean velocity profiles for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 149 5.23 Turbulent intensity profiles for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 149 5.24 Mean velocity profiles for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 150 5.25 Turbulent intensity profiles for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 150 5.26 Mean velocity profiles scaled wi th mean piston speed for the bathtub cham-ber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 151 5.27 Turbulent intensity profiles scaled wi th mean piston speed for the bathtub chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 151 5.28 Mean velocity profiles scaled with mean piston speed for the single slot chamber at W O T , for two speeds; 20.0 and 33.3 rps 152 5.29 Turbulent intensity profiles scaled with mean piston speed for the single slot chamber at W O T , for two speeds; 20.0 and 33.3 rps 152 5.30 Mean velocity profiles scaled wi th mean piston speed for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 153 5.31 Turbulent intensity profiles scaled wi th mean piston speed for the castel-lated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 153 5.32 Mean velocity profiles scaled with mean piston speed for the bowl-in-piston chamber at W O T , for two speeds; 20.0 and 33.3 rps 154 5.33 Turbulent intensity profiles scaled wi th mean piston speed for the bowl-in-piston chamber at W O T , for two speeds; 20.0 and 33.3 rps 154 xiv 5.34 Mean velocity profiles scaled with mean piston speed for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 155 5.35 Turbulent intensity profiles scaled wi th mean piston speed for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps 155 5.36 Mean velocity profiles scaled with mean piston speed for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 156 5.37 Turbulent intensity profiles scaled with mean piston speed for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps 156 5.38 Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub' group of chambers at M B T , W O T and 20.0 rps 157 5.39 Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub' group of chambers at M B T , W O T and 33.3 rps 157 5.40 Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub' group of chambers at M B T , W O T and 50.0 rps 158 5.41 Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 20.0 rps 158 5.42 Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 33.3 rps 159 5.43 Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 50.0 rps 159 5.44 Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 20.0 rps 160 5.45 Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 33.3 rps 160 5.46 Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 50.0 rps 161 xv 5.47 Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 20.0 rps 161 5.48 Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 33.3 rps 162 5.49 Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 50.0 rps 162 5.50 Ensembled fired and motored pressure profiles over four strokes for the single slot chamber at W O T and 33.3 rps. F ired trace for M B T and R A F R = 1 . 0 0 163 5.51 F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 164 5.52 F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 2 7 164 5.53 F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 0 0 165 5.54 F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 . 165 5.55 F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 166 5.56 F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 166 5.57 F ired pressure profiles for the 'disc' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 167 5.58 F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 2 7 167 xvi 5.59 Fired pressure profiles for the 'disc' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 0 0 168 5.60 F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 168 5.61 F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 169 5.62 F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 169 5.63 F ired pressure profiles for the 'bathtub' group of chambers at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 0 0 170 5.64 F i red pressure profiles for the 'bathtub' group of chambers at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 2 7 170 5.65 F ired pressure profiles for the 'disc' group of chambers at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 0 0 171 5.66 F ired pressure profiles for the 'disc' group of chambers at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 2 7 171 5.67 F ired pressure profiles for three different chamber geometries M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 0 0 172 5.68 F i red pressure profiles for three different chamber geometries at M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 2 7 172 5.69 Mass fraction burned curve for the bathtub chamber at M B T , W O T and 33.3 rps for R A F R = 1 . 2 7 173 5.70 Mass fraction burned curves for different chamber geometries at M B T , W O T and 33.3 rps for R A F R = 1 . 0 0 174 5.71 Mass fraction burned curves for different chamber geometries at M B T , W O T and 33.3 rps for R A F R = 1 . 2 7 174 x v i i 5.72 Mass fraction burned curves for the bathtub chamber at M B T , W O T and R A F R = 1 . 2 7 for five speeds; 20.0, 33.3, 40.0, 50.0 and 66.7 rps 175 5.73 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 176 5.74 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 2 7 176 5.75 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 0 0 177 5.76 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 177 5.77 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 178 5.78 Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 178 5.79 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 179 5.80 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 2 7 179 5.81 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 0 0 180 5.82 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 180 5.83 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 181 5.84 Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 181 x v i i i 5.85 Mass fraction burned curves for five different chamber geometries at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 0 0 182 5.86 Mass fraction burned curves for five different chamber geometries at M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 2 7 182 5.87 Mass fraction burned curves for three different chamber geometries M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 0 0 183 5.88 Mass fraction burned curves for three different chamber geometries at M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 2 7 183 5.89 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 184 5.90 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T and 20.0 rps for R A F R = 1 . 2 7 184 5.91 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T , and 33.3 rps for R A F R = 1 . 0 0 185 5.92 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 185 5.93 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 186 5.94 Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 186 5.95 Mass fraction burned ratio bar graphs, relative to the disc chamber, for M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 0 0 187 5.96 Mass fraction burned ratio bar graphs, relative to the disc chamber, for M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 2 7 187 5.97 Mass fraction burned ratio bar graphs, relative to the castellated chamber, M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 0 0 188 xix 5.98 Mass fraction burned ratio bar graphs, relative to the castellated chamber, for M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 2 7 188 5.99 Indicated mean effective pressure per cycle for the single slot chamber at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 : 200 cycles 189 5.100Indicated mean effective pressure per cycle for the bathtub chamber at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 : 44 cycles 189 A . l Kis t ler and laboratory calibration curves for pressure transducer Model 6121 No. 317205, (used for single slot castellated, bowl-in-piston and fired squish jet tests) 193 A . 2 Kis t ler calibration curve for pressure transducer Mode l 6121 No. 282737, (used for standard bathtub and disc chamber tests) 194 A . 3 Kis t ler calibration curve for pressure transducer Mode l 6121 No. 317125, (used for the motored squish jet test only) 195 A .4 Mer i am calibration curve for Mode l 50MW20-1.5 No. S-4875-l,( used for Natura l gas flow rate). 196 A . 5 Mer iam calibration curve for Mode l 50MC2-4F No. S-4875-2,( used for A i r flow rate) 197 B . l Hotwire calibration curve for W i r e No. 1 Ramb = ll.MQ.Rop = 11.95£2, (used for the bowl-in-piston chamber tests) 201 D . l Motored pressure trace at 50.0 rps low pass filtered at 6 k H z over the entire range and over an ini t ia l region 208 D.2 Motored pressure trace at 50.0 rps wi th the region between 150 and 60 degrees B T D C replaced wi th a section under end tension 209 xx D.3 Motored pressure trace at 50.0 rps with the region between 180 and 60 degrees B T D C averaged and smoothed over 6 and 12 degree windows. . . 210 D.4 Expanded fired pressure traces at 33.3 rps, with and without averaging and smoothing applied over 12 degree windows in the region between 150 and 55 degrees B T D C 211 xxi Nomenclature A Calibrat ion constant, hot wire calculation AS Area of flame front A B D C After bottom dead centre A T D C After top dead centre B Calibrat ion constant, hot wire calculation B D C Bot tom dead centre B B D C Before bottom dead center B T D C Before top dead centre B M E P Brake mean effective pressure (bar) B P Brake Power ( k W ) Bsfc Brake specific fuel consumption ( g / k W h r ) cP Constant pressure specific heat (k J /kg-K) cv Constant volume specific heat (k J /kg-K) ca Crank angle (degrees) CL Con rod length C O V Coefficient of variance D Piston bore d Bowl diameter E V O Exhaust valve open E V C Exhaust valve close h Convective coefficient ( W / m 2 - K ) i Number of record x x i i IMEP Indicated mean effective pressure (bar) Integer Analogue interger IVO Intake valve open IVC Intake valve close Lxg Lateral integral scale Lxf Longitudinal integral scale Lx Integral length scale LT Integral time scale L H V Lower Heating Value (kJ/kg) M B T Minimum ignition advance for best torque (degrees rhb Mass engulfment rate N Rotation speed (rps) N Number of cycles n Calibration constant Nu Nusselt Number Pbar Pressure (bar) PBDC Pressure at B D C (kPa) Pamb Ambient pressure (kpa) P{i) Pressure at i (kPa) APpiiton Change in pressure due to piston motion (bar) APcomb Change in pressure due to combustion (bar) A P ( i + 1) Change in pressure between point i and t + 1 (bar) Pcomb{total) Total pressure change due to combustion (bar) Qair Volumetric Air flow rate (m3/sec) Qtotal Total air and fuel flow supplied (m3/s) x x n i r Stroke R(x) Spatial autocorrelation coefficient RT Temporal autocorrelation coefficient Ramb Ambient resistance of wire Rop Operating resistance of wire R A F R Relative air fuel ratio Re Reynolds Number R P S / r p s engine speed revolutions per second %SAREA Squish area Sp Mean Piston speed (m/s) Sg Position of piston from B D C position t T i m e T T ime period T ( i ) Temperature at i ( K ) Tamb Ambient Temperature (k) Operating temperature ( K ) Tamb Ambient temperature ( K ) UT Turbulent flame speed (m/s) U(t) Instantaneous velocity at time t (m/s) u(t) velocity fluctuation at time t (m/s) u Turbulence intensity (m/s) U Mean velocity (m/s) U(t,i) Instantaneous velocity at time t in record i (m/s) U(tw,i) Average velocity at each window midpoint (m/s) Uw{t,i) Interpolated window averaged velocity (m/s) x x i v Uw(t) Window ensembled mean velocity (m/s) u\y{t,i) Fluctuating component of velocity (m/s) u(tw,i) Fluctuating velocity at each window midpoint (m/s) u(tw) Averaged fluctuating velocity at each window (m/s) u\y{t) Representative rms intensity (m/s) UE(t) Ensembled averaged mean velocity (m/s) !££;(£, i) Fluctuating component of velocity (m/s) UE(t) Rms intensity (m/s) V Volume (m3) VcBDC Volume at BDC before compression (m3) VcBDC Volume at BDC after expansion (m3) Vref Reference volume (m3) V, Swept volume (m3) WOT Wide open throttle a Thermal coefficient of resistance (1/K) r)th Brake thermal efficiency (%) 7/„ Volumetric efficiency (%) 7C Polytropic coefficient of compression 7e Polytropic coefficient of expansion A Ratio: crank length to conrod length AT Taylor microtime scale A z Taylor microlength scale pu Density of unburned fluid (kg/m3) 6 Crank angle from BDC (degrees) r Time (s) xxv Acknowledgments I am most grateful to Dr . R . L . Evans for his supervision and encouragement during this study. Further I would like to express my thanks for his concept of the 'squish jet' piston and for the opportunity to contribute to its evaluation. I also wish to express my appreciation to Professor P . G . H i l l for his interest, enlightening discussions and guidance during Dr . Evans' research leave at Cambridge University. I would like to thank the academic and technical staff of the Mechanical Engineering Department especially: A . Steeves for his help with the computing system; S. Oshika and J . Richards for their assistance in the operation and 'fault finding' of the test engine; T . Besic for the machining of the modified pistons; Dr . K . V . B u r y for his clarity in statistical matters; A . K a p i l and other recent graduates of the A F L Group for their support and legacy of information. Also I record my appreciation of the interest and financial support provided by the Canadian Gas Association for the Alternatives Fuels Laboratory. F ina l ly I would like to thank Professor J . D . W i l l m s and R. Love for the use of their facimile system and Professor Helen Tippett , V ic tor i a University of Well ington, for help i n the production of this thesis. xxvi Chapter 1 Introduction 1.1 Introduction and Background Since the first days of engine design, studies aimed at improving the combustion process to optimize spark ignition engine operation have been much in evidence. Initial ly it was sufficient to achieve smooth running at max imum power but the area of interest has expanded to meet specific requirements. The energy crisis of the 1970's promoted inves-tigations into alternative fuels. In the increased environmental concerns and restrictions on exhaust emissions of the present day, improvements in the combustion process have become a high priority. To improve engine operation, the engineer is concerned with fuel economy, knock propensity and emissions, which are al l combustion issues. Work output from the SI engine is obtained from a 'stepwise' cyclic production of power, based on the successive burning of a fixed amount of charge i n the combustion chamber. The requirements of maximum pressure development and repeatability of the torque output determines the acceptable range of operation of the system. As the operating characteristics are pushed to the extremes of charge dilution for economic and environmental reasons, the effect of cycle-to-cycle variations is increased. Cycle-to-cycle variations i n the combustion process from whatever cause are mani-fested in cyclic variations in the output torque and consequently affect drivability. Where these cyclic variations are large, the problems associated with operation at the extreme 1 Chapter 1. Introduction 2 l imits , such as misfire and knock, can seriously affect the engine performance. In balancing the requirements of modern engines, researchers and design engineers have aimed at the so called 'fast-lean' burn operation. Lean burn operation Lean operation, that is, at a relative air fuel ratio greater than stochiometric, has the advantage of increased fuel economy, higher thermal efficiency and lower levels of N O x emissions. In addition to efficiency from lower fuel consumption, the lower combustion temperatures developed and reduced pumping losses at part load associated with lean burn improve the efficiency. The disadvantages of lean operation result from the slower burning leading to in-creased cyclic variations, and from higher engine operating temperatures due to the longer burn period. A t the lean l imi t also, the level of unburned H C rises due to misfire and partial burn. A t a normally stable operating condition cyclic variation, may cause the engine to exhibit knock tendencies or to misfire due to an extreme cycle being at the lean l imi t . Fast burn operation Fast burn operation has the advantage of increasing thermal efficiency as the combustion process approaches constant volume combustion as represented i n the ideal Otto cycle, wi th higher peak pressure and hence greater indicated work. Fast burn also decreases the negative effects of lean operation, reducing cyclic variations and resulting in an extension of the lean l imi t and improved knock l imits . The disadvantages of fast burn are higher emissions of N O x from higher combustion temperatures, although more complete burning may reduce H C levels. Increased engine noise and roughness from the high rate of pressure rise can also be a l imit ing factor. Chapter 1. Introduction 3 Fast burn operation is particularly important with the use of slow burning fuels, such as natural gas. Stable fast-lean burn operation A balance of stable fast-lean burn operation, with minimal cyclic variation may be achieved by various means: optimization of mixture motion, increasing the flame speed through increased turbulence; optimization of the ignition system and geometry, selecting the number and position of spark points for shorter flame travel; and finally, through the chemical system by improving the reaction kinetics using additives for enhanced stability. Extensive research into the nature of cyclic variations in engines has been carried out through turbulence studies in bombs, rapid compression machines and SI and CI engines. While complicated by the interaction of the above variables, and the controversy of turbulence definitions in the engine environment, these studies have confirmed the early conclusions of Karim [1]: "By far the greatest variable element in engine operation is that of charge motion inside the cylinder, both in quality and direction." A review by Young [2] confirmed that cyclic variations in both scale and intensity of the flow near the spark plug at the time of ignition, would have a profound effect on the cycle-to-cycle combustion characteristic. In his study of the causal relation between engine variables and cyclic variations in combustion, Young [4] deduced that while engine and operating variables may significantly affect combustion variation, they do not necessarily cause it. He concluded that in-cylinder velocity variations near the spark were the major cause. With a predetermined engine system and operating fuel, the potential for optimization of the ignition and chemical systems is often limited. The influence of mixture motion on Chapter 1. Introduction 4 cyclic variations is therefore the area of greatest interest. In 1967, K a r i m [1] provided an early incentive for continued research in this area when he estimated economic increases through the reduction of gross variability as 10-20% at lean operation and 3-5% at opt imum operating levels. 1.2 Objective of this study The overall objective of this study was to investigate the influence of combustion chamber geometry on turbulence enhancement in the achievement of fast-lean operation of an SI engine. Before discussing the scope of the work undertaken, the remainder of this chapter provides a brief background on the turbulence flow field i n an engine and landmark studies in this field and discusses the terminology used. 1.3 Turbulent Flow Field in an Engine The in-cylinder flow i n an engine is determined by turbulence production influenced by geometric and dynamic considerations. Turbulence production occurs through the effect of piston motion on the fluid, shear flow past the intake valve and amplification of in-cylinder turbulence during the compression stroke due to the change in volume. The role of turbulence in affecting the flame speed was critically examined by Andrews et al [3] who showed that a higher level of turbulence increased the turbulent flame speed and burn rate. Various studies have shown that turbulent flame speed can be an order of magnitude greater than the laminar flame speed. Mechanisms for increasing the burning rate by turbulence, postulated i n several stud-ies, have been classified into both scale dependent and independent means.Damkohler's [5] flame wrinkl ing model is based on the effect of eddies, on a scale less than the thickness of Chapter 1. Introduction 5 laminar flame front, causing increased flame front area. Shchelkin's [6] model is based on the effect of eddies, on a scale larger than the flame thickness, higher gradients causing distortion of the flame front and hence increased area. The models developed by Tabaczynski [7] and Bl izard and Keck [8] are based on en-trapment of intermittent regions of activity in the flame front involving microconvection mixing of burned and unburned regions operating at all scales. The mass engulfment rate is determined by: rhb = UT * Af * pu where UT is the turbulent flame speed, Af is the area of the flame front and pu is the density of the unburned fluid. The turbulent flame speed being a function of laminar flame speed is also dependent on stoichiometry and gas properties. The turbulent flow field may be characterized by a set of parameters using the theory of isotropic turbulence. Two distinct features of engine turbulence noted by Tabaczynski [9] are that it was periodic and of the order of the mean. The study of turbulence in engines is further complicated by real engine flow parameters involving highly unsteady, inhomogeneous, non isotropic conditions and changing thermodynamic and geometric conditions. The results of turbulence studies i n engines from the pioneer work of Semenov [10] using Hot W i r e Anemometry ( H W A ) techniques to more recent Laser Doppler Anemom-etry ( L D A ) work by Fraser and Bracco [11] are discussed in Chapter 2. The literature review also discusses the engine and operating variables studied by Young [2] and others, and their relative influence on cyclic variation. Chapter 1. Introduction 6 1.4 Discussion of Terminology in Turbulence Studies As noted previously, there is some controversy in turbulence definitions in the engine environment. Turbulence is denned as a quantity that shows random variation wi th respect to time and space from which statistically distinct averages can be discerned [12]. Separation of such averages wi th engine velocity measurements pose problems of separating cycle-to-cycle variations from turbulence velocity fluctuations. Mean velocity is determined from the instantaneous velocity U(t) at any time t. Turbulence is then characterized by the following means. Mean velocity is determined: where T is a sufficiently long time period to contain the turbulent fluctuations. The velocity fluctuation is then separated by: The periodic nature of engine turbulence requires non-stationery methods for distin-guishing the mean from the instantaneous velocity determined for each record t . The definition of turbulence used by Lancaster [13] was defined by the fluctuating component determined by: u(t) = U(t) - U Turbulence intensity is defined as the root mean square, rms, fluctuation u: u{t,i) = U(t,i) - U(t) - U(i) where U(t,i) is the instantaneous velocity at time t in record i, U(t) is the ensemble averaged mean at time t over record i's, and U(i) is a time averaged mean after U(t) has been subtracted from the instantaneous velocity for each record i. Chapter 1. Introduction 7 Using a similar approach Heywood [14] questioned whether the turbulent fluctuating components of the instantaneous velocity could be separated from the cycle-to-cycle variations. The non-stationary method used in this study is discussed in detail in the analysis of test data, i n Chapter 4. W i t h the selection of an appropriate window, the non-stationary method uses a window averaging technique in place of the above equations. Turbulence is distributed over a continuous range of eddy sizes. Scales of turbulence, a representative measure of the average values of quantities wi th respect to time and or space are calculated by correlation techniques. The spatial integral length scale is defined: pothesis [15] relating spatial and time scales is applied. The time scales are determined from the temporal autocorrelation: where t is the correlation time and T is the period of measurement. App ly ing Taylor's hypothesis the spatial integral length scale is calculated: where R(x) is the spatial autocorrelation coefficient. In general practise, Taylor's hy-1 fT u{r)-u{t + T) dt U*LT The Taylor microtime scale can also be calculated from this autocorrelation: K = '\d2R{r)y 2 dr 2 and similarly the microlength scale: Xx = U * K Chapter 1. Introduction 8 For homogeneous isotropic turbulence the longitudinal integral scale is related to the lateral by: Lxf = 2 * Lxg Generally, the integral length scale in an engine is of the order of chamber height, 10mm and seen as a typical mixing length. The Taylor microscale is of the order of 0.1 m m and the Kolmogorov Scale at which dissipation occurs is of the order 0.01 mm. The validity of the assumptions of homogeneous and isotropic turbulence used by Taylor i n engine operation is discussed in the literature review. Of the characteristic turbulence parameters described above, Andrews and Bradley [16] determined that the role of turbulence scale was important in the in i t i a l stages of combustion only, and that turbulence intensity was primari ly responsible for increased flame speed. 1.5 Scope of Work The main means of turbulence generation in an engine is by shear flow past the intake valve, which decays during the compression stroke. Turbulence enhancement is achieved by the break down of large scale organized motion either from ' s w i r l ' - a tangential rotat-ing flow generated during the intake stroke through inlet port geometry or by shrouded valves, or from 'squish'- a radial injecting flow generated during the compression stroke before T D C : or a combination of these two flows. As defined i n section 1.2, the overall objective of this study was to achieve fast-lean burn operation of an SI engine through turbulence enhancement. The advantages of an increased burning rate to improve efficiency and reduce cyclic variations inherent in lean operation has already been discussed. This is of particular importance in slow burning alternative fuels. The specific focus of the work was to improve understanding of squish and enhanced squish effects in promoting such operation. Chapter 1. Introduction 9 A comparative study of chamber design, restricted to squish type combustion cham-ber configurations, was performed in two areas. First , a qualitative and quantitative comparison of turbulence concentrated on the intensity parameter was carried out, using hot wire measurements at the spark location. Second, the effect of chamber design on the engine performance was assessed, with a subsequent correlation of the turbulence and combustion measurements. Analysis and evaluation of the experimental data was directed towards characterizing the cyclic variation through a statistical analysis of indicated mean effective pressure ( I M E P ) , and peak pressure variation. Investigation of the combustion process in i t ia l and main burn periods was aimed at determining the primary area of influence of each chamber design, and the l ink wi th the turbulence effects. The area of particular interest in the study was the enhancement of the squish effect using squish jet pistons. However, consistency and accuracy of the test base for com-parison was also a specific interest. The engine operation variables therefore involved a range of speed, loads and air fuel ratios. One of the difficulties associated wi th engine studies has been the multitude of inter-acting engine variables and dependence on the conditions of operation. Earlier engine studies, restricted to low operating speeds and cumbersome data acquisition systems were inclined to draw conclusions from single sets of data. More recent studies have qualified their results for specific engines or specific sets of conditions. The range of test conditions and measurement parameters used i n this study were selected wi th the aim of providing the broadest practical comparison with previous studies i n this field, and to give a consistent basis for comparison wi th other current work. Hence a standard bowl-in-piston was used as a comparison for the squish jet bowl-in-piston previously studied by Cameron [17]and Dymala-Dolesky [18]. A standard bathtub cylinder head chamber was used as a basis for comparison wi th studies using the new Chapter 1. Introduction 10 modified 'jet' bathtub pistons. Disc piston measurements were also taken as a non-squish case. 1.6 Structure of thesis Chapter 1 has discussed the study objectives and provided a background to the continued interest in fast-lean burn operation as a means of improved engine operation, particularly i n the use of natural gas and other alternative fuels. Chapter 2 presents a review of the extensive literature on turbulence and combustion studies in engines, wi th a final section on combustion chamber design. Chapter 3 describes the experimental apparatus and data acquisition systems used in this study, the analysis of which are described in Chapter 4. Chapter 5 discusses the experimental results in detail , the constraints and limitations, and relates the present findings to previous studies of turbulence and combustion. Chapter 6 presents the conclusions from this investigation and suggest directions for further research aimed at improving the combustion process to optimize spark ignition engine operation. The list of references cited in this thesis precedes the tables, figures and graphed results. Appendices A and B contain further details on instrumentation specifications and calibration procedures. Appendix C presents data on the properties of B C natural gas and lower heating value calculations. The final Appendix contains details of the filtering methods examined and used in the pressure data analysis. Chapter 2 Literature Review 2.1 Introduction The review of the literature and background of engine studies has been divided into three sections. The first section follows the successive progression of engine investigations of turbulence from early bomb work in the 1970's to present day computer simulation models. The second section covers combustion studies in the same period. The last section involves both turbulence and combustion studies in combustion chamber design and includes an evaluation of previous research on squish jet pistons. Studies of turbulence in engines, including the related studies on flow visualization bombs and rapid compression machines, have been concerned wi th characterizing turbu-lence, determining its source, achieving a reliable means of measurement and assessing its effects on cyclic variations in combustion. A n outline of the major stages in turbulence studies is given wi th a summary of the findings together with a review of the techniques used i n measurement. The combustion studies review summarizes the engine variables investigated in assess-ing the causes and effects of cycle-to-cycle variations. Emphasis has been placed on those variables controlled i n this study and those most likely to influence the test conditions. Characterization of cyclic variations through pressure measurements is also reviewed. More detailed reviews, on fluid motion within the cylinder of internal combustion 11 Chapter 2. Literature Review 12 engines, cyclic dispersion in spark ignition engine and the use of cylinder pressure mea-surement in combustion studies of engines, are contained i n papers by Heywood [14], Young, up to 1980 [2] and A m a n n [19] respectively. 2.2 Turbulence studies in Engines Semenov's [10] studies of turbulent gas flow in piston engines in 1963 is regarded as a fundamental work. Hot wire anemometry ( H W A ) studies in a motored C F R engine with disc piston showed that shear flow through the intake valve was the primary source of turbulence generation followed by decay during the compression stroke and relaxation towards isotropic turbulence at top dead centre ( T D C ) . Tri-axial H W A measurements were carried out by Lancaster [13] in a motored C F R engine, wi th and without swirl , generated by a shrouded intake valve. His studies con-firmed isotropy and showed that the mean and rms velocity increased linearly with engine speed. Similar investigations by others [20, 21, 22, 23] also showed isotropy and scaling wi th engine speed. Bopp, Vafidis and Whitelaw [22] estimated scaling of turbulence at T D C with mean piston speed in the range of 0.47 to 0.60 comparable wi th the findings of 0.5 for a disc chamber by H a l l and Bracco [24]. Most recently, laser homodyne measurements in an engine wi th a flat topped piston and varying swirl conditions by intake geometry by Ikegani et al [25], confirmed that in no-swirl cases the turbulence field near the end of compression is almost uniform. Hot wire anemometry The major method of turbulence measurement i n the 1970's was with hot wire anenom-etry, and extensive investigations were made into the accuracy of such methods [21]. Chapter 2. Literature Review 13 Dent and Salama [26] and Haghgooie et al [27] were among the first to measure tur-bulent time scales using H W A techniques wi th high speed random signal analysis. Using the above isotropic and homogeneous assumptions microlength scales were estimated to be of the order of 1/10 m m . Laser Doppler anemometry Ear ly laser Doppler anenometry ( L D A ) measurements such as those by Rask in 1979 [28] encountered problems with seeding, vapourization and visualization requirements of engine design. Advances in L D A techniques have been used by many researchers to verify H W A results and obtain more direct scale measurements. Fraser et al [29] measured integral length scales in a motored disc engine with and without swirl . They confirmed early measurements of Lancaster [13] and Hey wood [14], that is, integral length scale was of the order of the chamber clearance height at T D C , a l / 5 t h ratio was proposed. Fraser and Bracco [11] used a two-point, single probe-volume Laser Doppler Velocime-try system ( L D V ) to measure length scales directly. Their results appear to indicate that for the specific chamber investigated, the lateral integral length scale from both cycle-resolved and ensemble averaging did not scale with the clearance height, but was independent of it around T D C . A correlation between the larger spatial scales and the lower frequencies of fluid flow was also demonstrated. These scale studies have overshadowed research based on Lancaster's [13] and Se-menov's [10] conclusions that the turbulent intensity was the single parameter of impor-tance. Later studies noted however, that the Taylor approximation for length scales stil l had value due to its ease of calculation. Similarly, H W A methods are st i l l in use due to their relative simplicity and inexpense. Analyz ing techniques suggested by Witze [30] from the comparison of hot wire anenome-try and laser Doppler velocimetry used i n IC engines, provide the basis for the continued Chapter 2. Literature Review 14 use of H W A as selected for this study. The final area of turbulence studies reviewed concerns those l inking the turbulent flow field and its effect on the flame speed, with combustion studies. Matsuoka et al [31] compared flame arrival times by ion gap measurements in a fired engine, to mean velocities measured at the spark location while motored, using swirl and non-swirl conditions. They concluded that the mean velocity did not influence the flame speed, but that the small scale turbulence, produced by the breakdown of large scale motion, ie., swirl , d id . Schlieren photography techniques used by Gatowski et al [32] showed a wrinkled flame front developed after formation of the spherical flame kernel at spark, indicating laminar like burning immediately following the spark. Similarly, work by Keck, Heywood and Noske [33] on early flame development using schlieren techniques, supported the wrinkled laminar flame model of turbulent structure. The applicabilityof turbulence measurement in motored engines to fired engine oper-ation was originally supported by L D A measurements in a fired engine at 800 R P M by Rask [28]. These showed that motoring and firing results were quite similar unt i l the time of ignition and differed only when the flame was near the location of the velocity measurement. This work was also consistent wi th the rapid distortion theories of Wi tze and M a r t i n et al [34, 35] of turbulent intensity increasing ahead of the flame from one dimensional of compression of the unburned charge. Barton et al's [36] autocorrelation of peak pressure and cycle, showed no correla-t ion, leading to the conclusion of 'no history' in the pressure fluctuations. This study found that cyclic variations depended on the intake, compression and combustion strokes of each individual cycle, allowing comparisons of fired cycles wi th motored intake and compression stroke analysis. Advances in computer systems and analysis methods have shown promise of analytical Chapter 2. Literature Review methods for prediction of flow behaviour. However Gosman [37] cautions that, "The accurate prediction of the flow behaviour during individual cycles, which is of interest in correlation with cycle-to-cycle variations in combustion performance is believed to be outside the capabilities of the present method-ology.". Summary of turbulence study findings The general findings of the turbulence studies since Semenov's pioneering research in the early 1960's have been: • Turbulence approaches a homogeneous and isotropic state near T D C of compression. • In the absence of turbulence production generated through chamber ge-ometry, shear flow past the intake valve is the major source of turbulence in the engine cylinder. • Mean velocity and turbulent intensity are proportional to engine speed, wi th a near linear relation. • Cycl ic variations increase with increased engine speed. • In chambers without swirl or squish turbulent intensity at T D C scales wi th mean piston speed, in the range 0.47 to 0.60. • Turbulent length and time scales are inversely proportional to engine speed. • The integral length scale is of the order of chamber height at T D C . • Turbulent microtime scales are of the order of 0.1 millisecs. • Turbulent microlength scales are of the order of 0.1 m m Chapter 2. Literature Review • Turbulent energy is concentrated in the low frequencies; below 1000 H Z . • Flame travels as a laminar flamelet immediately after spark. • Hot wire measurements in the motored engine may be approximated to conditions in the fired engine prior to combustion. The last point forms the basis of motored engine turbulent studies use in combustion studies. 2.3 Combustion Studies in Engines Combustion engine studies have been used to establish which engine variables are responsible for cyclic variations, methods to measure this variation and the means to control cyclic variations through the most influential variable. Combustion experiments in engines were ini t ia l ly concerned with deter-mining the source of the cycle-to-cycle variations and influencing factors. Ear ly investigations covered both CI diesel and SI engines wi th l imited appli-cations of information between valve and ported engines. Factors investigated were, mixture type and preparation, the ignition system, engine speed, com-pression ratio, combustion chamber geometry and mixture motion. A t the same time, the best means of characterizing the cycle-to-cycle variation mani-fest in torque output and related to the pressure development, was evaluated. From the recognition of the importance of in cylinder mixture motion, investigations continued in the direction of increasing turbulence, through combustion chamber and inlet geometry, and correlations of combustion phe-nomena wi th turbulence. Barton et al's [36] statistical analysis of performance data in a fired C F R engine confirmed Kar im's [1] early theory that mixture Chapter 2. Literature Review motion was the primary cause of combustion pressure variations. The impor-tance of each engine operating variable was then determined by its affect on flame speed, flame distance travelled, variations of the mixture motion or a combination of these. The difficulty encountered with much of the engine research is the inter-action of engine variables and operating conditions. Isolation of the various effects has often been impractical . Steps to overcome these problems in this investigation are discussed in Chapter 3. Chemical variables Chemical factors of fuel air mixing , equivalence ratio, residuals and fuel type were investigated by skip firing tests by Soltau [38] and similar methods by others [1, 39]. M i n i m u m variation was found at the equivalence ratio giving the max imum power and shortest combustion duration. Hence, fuel air mix-ture and residuals were found to be important to the extent they affected the equivalence ratio; generally richer mixtures and faster burning fuels had the least fluctuation. A semi-quantitative study by Hansel [40] verified that leaner mixtures produced larger variations. Ignition variables Ignition variables, type of system, spark electrode geometry, and spark j i t-ter were shown to have negligible effects on cyclic variation in early studies [2], aside from any change in the number or location of spark points which decreased the maximum flame travel distance and hence decreased burning duration. Recent investigators [41, 42] have found that the arc duration and gap have significant effects on the combustion quality and stability in lean Chapter 2. Literature Review mixtures, wi th the effects becoming more pronounced as ignition t iming is ad-vanced and load reduced. Kalghett i [43] argued that, regardless of the system, once an ignition criterion and init iat ion criterion were met cyclic variations could not set in once the flame kernel was larger than a crit ical size, with the exception of some cyclic variation in the local mixture strength near the spark. This does not necessarily contradict the early investigator's results which were conducted principally at full load stochiometric conditions. Engine speed Winsor and Patterson [20] concluded that there was an overall weak relation between engine speed and cyclic variations. A n increase in engine speed, while producing an increase in flame speed also showed an increase i n variation of peak pressure at stochiometric conditions, related to increased combustion variations from the increased turbulence variations. Kar im's [1] work with iso-octane in the lean range showed a decrease in the coefficient of variance wi th increased speed. Similarly, compression ratio effects have generally been shown to be weak. Barton et al [36] showed an increase i n compression ratio slightly decreased the cycle-to-cycle variation through increased flame speeds. The higher tem-peratures and lower residuals inherent in higher compression ratios also affect the reduction in cyclic variations [13]. Combustion chamber design Optimizat ion of combustion chamber design for the desired turbulence gen-eration, is reviewed in Section 2.4. There is a general consensus that the best combustion chamber to affect a reduction in cycle-to-cycle variations provides Chapter 2. Literature Review the shortest combustion duration, ie., more open and more compact chambers combined with a central spark location are preferred [4]. From his review of the results of early engine combustion research Pat-terson [39] reiterated the importance of variations in mixture motion. More recent studies have confirmed that temporal variation in velocity gradients and turbulent mixture motion is very influential, causing variation in growth rate, location of the flame kernel and flame speed at the spark location [33]. Pressure measurements In the assessment of cyclic variations and combustion characteristics, piezo-electric pressure transducer measurement of the in-cylinder pressure has been the fundamental tool [19, 39, 44]. Results of early piezo-electric pressure transducer selection and operation are discussed in the apparatus section. Various characterizing parameters have been used: • M a x i m u m cylinder pressure within each cycle. • M a x i m u m rate of pressure rise. • Crank angle at which peak occurs. • I M E P covariance calculations. • Burning times — combustion duration calculation. • Flame arrival times. Many researchers have used peak pressure [1, 20, 36, 38, 45, 46] as being strongly related to the combustion rate however, Matekunas [45] cautioned that peak pressure was not a good indicator of drivability. Similari ly, crank angle occurrence of peak pressure studied by Barton et al [36] was found to Chapter 2. Literature Review be less useful than peak pressure. Nagayana's [47] studies on disc, squish and swirl chambers showed a measurable relation between the cycle-to-cycle fluctuations in I M E P and the vehicle surge l imi t . Amann's [19] review of pressure characterization also gave a poor correlation between I M E P and peak pressure. Kuroda's [48] fast burn studies showed that variation of the order of 10% in I M E P coefficient of variance seriously affected vehicle drivability. Combustion durations have been determined from flame arrival times using ionization gap methods and heat release analysis methods using in-cylinder pressure data. The combustion models have ranged from a simple pressure ratio model developed by Rasswieler and Withrow [49] using com-parison wi th flame pictures, to more complex heat release models detailed by Gatowski [50] based on Krieger and Borman '6 [51] two zone model for spark ignition engines. Combustion process The combustion process can be commonly divided into three or four phases: 1. ignition and kernel development. 2. flame development. 3. fully developed. 4. termination. The first two phases are commonly termed 'the early combustion period' and the third referred to as 'the main combustion period'. To date, most interest has been taken in the early and main combustion periods, however, recent remarks by A m a n n [52] suggest that the continued restrictions and Chapter 2. Literature Review requirements of engine operation wi l l require investigation and optimization of the complex termination phase. Effect of turbulence on combustion Lancaster et al [44] conducted one of the first correlation studies of turbu-lence and combustion data using H W A and pressure measurements, wi th heat release model techniques. Comparing a disc chamber for a swirl and, a non-swirl case, Lancaster et al found that the normalized flame speed was a linear function of turbulence intensity with l i t t le influence of turbulent scale within the l imits of the measurements. Using L D A techniques Cole and Swords [53] found a strong correlation between the mean velocity and peak pressure wi th a weak relation wi th turbulence intensity. Some other researchers, notably Nagayana et al [47] and Matekunas [45] have examined the effect of turbulent generation by squish and swirl chambers on combustion, using pressure combustion data only. Nagayama's comparison of plain, swirl , squish, and swirl-squish chambers showed that squish was active in the main phase of combustion and swirl in the early phase. From the higher performance of the swirl-squish chamber Nagayana argued that the squish generated before T D C was important in breaking up the swirl motion generated during intake. Matekunas's comparison of three levels of swirl with a disc chamber, suggested that in the absence of a large scale flow motion, such as swirl , to provide energy to the smaller scales a lower mass burn rate resulted. The effect of early flame development was also seen to be important near the lean l imi t wi th negligible effect away from this l imi t . Belmont et al's [54] extensive statistical analysis of cyclic variability in an SI engine showed a strong memory element under certain conditions. Memory Chapter 2. Literature Review was shown to be inversely proportional to the degree of cyclic variability. One final point noted by Dai ly [55] on the inherent chaotic nature of combustion, was that the longer the burn duration, the more control non-linear combustion kinetics exerted over cyclic variation. Operating variables Although there is a plethora of engine performance data on various com-bustion chambers and engine configurations, there has been l i t t le attempt to standardize or isolate the variables examined. Ear ly investigations of the effect of swirl on SI and C I engines by M a [56] showed negligible effects of swirl in an SI engine. This was contradicted by Witze's studies [57, 58] which included the effects of other parameters, ie., engine geometry and operat-ing conditions. W i t z e also showed that cyclic variation was not necessarily decreased by increasing the burn rate. Matekunas [45] found that very high swirl caused pre-mature detachment of the flame kernel or possible quenching of the flame by convection against the cylinder walls. The negative effects of partial or complete quenching caused by excessive turbulence have also been considered a l imit ing factor in improvement in lean burn operation in more recent work by Sheppard and Bradley [59] and Saxena and Rask [60]. Summary of combustion study findings The general findings from combustion studies on the influence of turbulence on cyclic variation are: • The greatest variable element i n engine operation is the charge motion within the cylinder on the micro-turbulence level. Chapter 2. Literature Review 23 • Cycl ic variation presents a major obstacle to implementation of lean burn. • Cycl ic pressure variation is a direct consequence of variation in the com-bustion process and the rate of heat release. • The strength of the relationship between peak pressure and indicated work is dependent on the operating conditions. • The maximum rate of change of peak pressure is less sensitive to indi-cated work than peak pressure. • Near the lean l imit variation in the early flame development, and hence fluctuation i n the combustion duration, is the main cause of cyclic vari-ation. • M a x i m u m engine stability with min imum I M E P variation is that with the fastest and steadiest combustion. • Reduction of the main combustion time mitigates the effects of early variations and the converse. • Longer main combustion duration allows greater influence of non-linear combustion kinetics. • Increase i n cyclic variation reduces the degree of memory. • Reduction of the coupling between I M E P and combustion through op-t i m u m phasing reduces the cyclic effects. • Squish is most effective in the main part of combustion. • Swir l is effective in the early part of combustion. Chapter 2. Literature Review 2.4 Combustion Chamber Design The final section of this review deals with the studies of optimizing turbulence by combustion design wi th an emphasis on squish. Studies have shown [14, 61] that the optimum chamber design should provide fast repeatable burn for high efficiency; good emission control through large valve effective areas; and low wall surface areas to minimise heat loss and avoid quenching and also must be geometrically practical for manufacture. Gruden's [62] investigation of combustion chamber layout in modern pas-senger cars showed that a combustion chamber located in the piston crown was the simplest way to comply wi th these requirements. In line wi th Young's [4] optimization criterion of an open and compact chamber, Gruden concluded that the dimensions and the positions of the quench areas and the quench dis-tance, ie., maximum flame travel, were important. Overington and Thring's [63] work on a Ricardo hydra engine wi th variable compression ratio and combustion chambers in the head and piston showed a 2-5% improvement in fuel economy for the chamber i n the piston crown over that of the head chamber, arguably from higher turbulence during combustion induced by squish. Squish effects Ear ly studies of the squish effect reviewed by Young i n 1980, were predomi-nantly in diesel configurations and gave conflicting or negligible results. Whi le more recent studies in diesel engines [64, 65] provide information on squish and swirl interactions and the effects of combustion chambers shape on fluid motion, only l imited comparisons can be made with the SI engines. Chapter 2. Literature Review Evans [66] proposed a variation of the standard bowl-in-piston squish de-sign i n 1985. He used channels in a bowl-in-piston chamber to enhance the squish effect through developed jets. This design was evaluated analytically and with H W A and combustion pressure measurements in a C F R engine by Cameron [17, 67]. Init ial results showed an increase in peak pressure and re-duction i n combustion duration when compared with standard bowl-in-piston squish pistons. Subsequent investigations were carried out by Dymala-Dolesky [18] to evaluate the nature of the jets developed. The jets in general were shown to diminish the effect of the main squish motion. I M E P covariance calculations and mass fraction burned analysis indicated that the squish jet was most effective i n the latter half of combustion. The most promising chamber was wi th eight jets angled towards the centre of the bowl, at spark location. The effect of jet flow introduced in the cylinder was also investigated by Nakamura et al [68]. A piston operated jet valve was used to direct air or a super lean mixture towards the spark plug. A strong swirling flow was observed by ionisation flame arrival time measurements, and improvements in peak pressure were attributed to this. The optimum diameter and injected fluid rate was 6 mms and 0.97 1/s equivalent to a jet of approximately 34 m/s. The maximum jet velocities observed by Dymala-Dolesky at spark were 5-20 m/s suggesting stronger jets formed by blocking the main squish effect may have a more definite effect on the motion. As defined in the objective, the principle aim of this project was to achieve fast-lean burn operation of an SI engine through turbulence enhancement. The findings from the turbulence and combustion studies reviewed in this Chapter 2. Literature Review Chapter suggested several promising directions for further research. Chap-ter 3 describes the apparatus and data acquisition selected in this study to examine the effect of piston chamber geometry on turbulence performance. Chapter 3 Experimental Apparatus and Method 3.1 Introduction The experimental investigation of the effects of combustion chamber design on turbulence, cyclic variation and performance of a spark ignition engine, was conducted i n a Research Engine test cell facility of the Alternatives Fu-els Laboratory ( A F L ) , i n the Department of Mechanical Engineering, The University of Br i t i sh Columbia. The a im of the experimental measurement was to provide turbulence, combustion and performance data for a number of different combustion cham-bers. The ability of the Ricardo H y d r a research engine to be run at speeds representative of modern IC engines (20-90 rps), and the ease of conversion to a number of different configurations using alternative cylinder heads and pistons, has led to its use i n this work and at other research centers. Six combustion chamber configurations were studied using two types of cylinder head and six different pistons. The tests for each chamber configuration were divided into two phases: 1. Motored engine tests using a D C dynamometer, to yield phasing and flow measurement data. 2. F i red engine tests using B C natural gas as a fuel to yield combustion and performance data. 27 Chapter 3. Experimental Apparatus and Method The motored tests were run over a range of speeds at wide open throttle ( W O T ) . The fired test regime involved tests run at W O T full load, over a range of speeds, and part load at two speeds for stoichiometric to lean operation In this chapter the experimental apparatus, instrumentation and the data acquisition systems are first described, followed by a description of the two test procedures. Cal ibrat ion requirements, tests and data are also detailed. 3.2 Experimental Apparatus 3.2.1 Introduction The general arrangement of the Ricardo Hydra Test Ce l l with an I B M P C l ink setup i n an adjacent control room is illustrated in Figure 3.1. The test cell con-sists of a base mounted Single cylinder engine with electrical dynamometer, oi l and coolant modules, converter cabinet,transformer and control console. The engine, fully instrumented with electronic transducers as supplied by Ricardo Consulting Engineers, was equipped wi th a Cussons electronic con-trol unit . This unit with instrumentation and data acquisition systems added by the A F L Group enabled control and measurement of the main engine pa-rameters, such as torque, speed, flow rates, throttle position, pressures and temperatures through signals fed into an I B M personal computer. Measure-ment of cylinder pressure and hotwire signals was also possible. The basic engine system components relevant to this experimental work are briefly described i n the next sections. The instrumentation for monitoring and obtaining performance data is described in section 3.2.4. More detailed information is available in the A F L and Ricardo operation manuals [69, 70]. er 3. Experimented Apparatus and Method 3.2.2 Engine Bed The Ricardo Hydra is a single cylinder, four stroke, water cooled spark igni-tion engine with vertical valves operated by an overhead camshaft. The 0.45 litre capacity engine has an 80.2 m m bore, 88.9 m m stroke and a compression ratio of approximately 8.93:1. Engine specifications are given in Table 3.1. The engine can be operated to a maximum speed of 90 rps developing 15kW power. In this study, however l imitations on the modified combustion cham-ber materials resulted in the operation in mid range only. The standard bathtub cylinder head and flat piston configuration for gaso-line (or gaseous fuel) operation, cross section is shown i n Figure 3.2. The engine was coupled to an electrical D C trunnion mounted dynamome-ter which uses a regenerative load system and was operated through a K T K thyristor converter unit. The dynamometer's output, capable of absorbing 44 kw of engine power up to 5500 rpm was fed into the three phase 400 volt mains supply. The M c C l u r e dynamometer was also used to motor the engine where the power is drawn from the A C supply through the thyristors. The engine ignition system used a conventional coil and spark plug ar-rangement with the primary coil circuit operated by a 'Lumenit ion ' unit picking up speed and T D C reference from the flywheel. This system enabled manual control of the ignition t iming from the control consol. A standard champion N 6 Y spark plug wi th 0.6mm gap was used. The cooling and lubrication oi l systems are mounted within the engine bed. These comprised a closed circuit pressurised coolant system of a water and antifreeze rust inhibitor mixture; oil and coolant circulation pumps; an Chapter 3. Experimental Apparatus and Method integral oil filter and oil and coolant heat exchanger, oil and coolant tempera-ture control valves, and mains water bypass valve. Temperature control, and temperature maintenance during motored tests, was provided through sepa-rate oil and coolant heaters with Spirac Sarco l imit sensors set in the range 60-85 ° C . Low level and low pressure sensors in the coolant and oil system respectively, and thermocouples in each system, provide safety overheating trips together wi th temperature indication at the control consol. Fuel supply for operation wi th natural gas were provided by regulated flow from B C Hydro mains, with needle valve control in the control room. Flow measurement was provided by a M i r i a m laminar flow (50MW20-1.5) with a differential pressure transducer and signal demodulator mounted between the regulator and needle valve. The pressure drop across the laminar flow element was also sensed by an inclined differential manometer i n the control room. The air intake system consisted of an air filter, a 1 k W heater, throttle body assembly with servo motor controlled throttle and inlet manifold. Sim-ilar to fuel flow, a M i r i a m laminar flow element (50MC2-4F) mounted ahead of the intake filter was used for flow measurement. 3.2.3 Combustion Chambers The standard gasoline Ricardo Hydra engine configuration has a bathtub com-bustion chamber located in the cylinder head and uses a flat piston wi th solid skirt , two compression rings and one oil control ring. Six different combustion chamber configurations were obtained by using the standard bathtub cylin-der head wi th three piston shapes and a flat cylinder head with an additional three piston shapes. Chapter 3. Experimental Apparatus and Method These configurations were separated into two groups. The 'bathtub' group consisted of the standard bathtub cylinder head wi th a flat piston configura-tion and two modified pistons termed 'single slot' and 'castellated' pistons. The 'disc' group comprised the flat cylinder head and non-squish disc piston configuration and two types of bowl-in-piston pistons termed 'bowl-in-piston' and 'squish jet ' pistons. These configurations are shown in Figures 3.3 to 3.5. Development and manufacture of the flat a luminium cylinder head with extended pistons was carried out by R. Dymala-Dolesky and is detailed in his M A S c report [18]. To allow for a more central position of the spark plug, the valve position was moved from the centre of the chamber. Exhaust port and cooling manifolds were redirected and a smaller 12mm spark plug was specified. The engine rebuild for this series of configurations required several modi-fications: • Installation of a longer cylinder liner. • Installation of a modified connecting rod. • Modifications to the cylinder block and t iming drive system. • Add i t ion of packing plates under the cylinder block. Pistons The extended pistons to accommodate the combustion chamber bowl were cast from a luminium alloy A356, heat treated to T6 , this being the same material as that used in the flat cylinder head. The new bathtub type pistons were manufactured from the pattern of the extended bowl type pistons. Sand casting and heat treatment were carried out by a local foundry, then the piston Chapter 3. Experimental Apparatus and Method blanks were machined in the machine shop of the Mechanical Engineering Department. Consequently these pistons were heavier than the original flat piston used i n the bathtub configuration. The l imitations of the sand cast piston material and process compared to a more standard forged or cast process with fast cooling in a permanent mould, resulted in a reduction of material strength estimated to be 1/3 to 1/2 that of the original. This and the hybrid nature of the bathtub type piston were considered in the choice of maximum operating speeds. The modified bathtub pistons were designed to have the same compression ratio as the bathtub wi th flat piston and the bowl i n piston configuration. L iqu id displacement volumetric checks on the inverted cylinder head and appropriate pistons with the valves in position and a pressure transducer blank inserted were carried out using a thin o i l . W i t h i n the accuracy of this technique the combustion chambers were found to have the same clearance volume. 3.2.4 Modified bathtub pistons The desired effect of the modified pistons was to block off the plain squish action occurring near T D C , and direct the flow through one or more narrow slots to form jet action. In his evaluation of the jet action produced by a new 'squish jet' piston Dymala-Dolesky measured a weak jet effect in the piston bowl. He therefore recommended the closure of the squish area to force flow through the jet forming channels. In this study it was proposed to create the forced jet action through slots in a raised wall on the piston surface, wi th the intention that these slots would Chapter 3. Experimental Apparatus and Method form channels as the wall entered the cylinder head cavity. The most practical design configuration within the l imits of the equipment available was the use of the standard bathtub cylinder head and a modified piston. The modified pistons were designed to have the same compression ratio and be capable of operating over the same range of speed and loads as the standard bathtub and bowl-in-piston chambers. A 5mm wall thickness was used equal to the min imum thickness of the cylinder head. The combus-tion chamber volume was maintained by the addition of a shallow bowl to compensate for the raised wall . To produce the maximum blockage effect at the earliest point before T D C the max imum wall height possible without fouling of the intake or exhaust valves was desired. The piston position per crank position was determined by: Se = r / 2 * (1 + cosO) + CL*(1- - (\sin6)2) where Sg is the distance from B D C , X = r/2*CL,rie the stroke and CL is the con rod length. The position of the valves were determined from their cam profiles and a maximum wall height of 12 m m chosen to give a min imum of 1 m m clearance. This clearance was verified by measurements i n the actual configuration. The 12 m m wall was calculated to produced a blocking effect after 36 degrees B T D C . Squish area A final note on the combustion chamber designs deals wi th the difference in squish area between the bathtub group of pistons and the bowl-in-piston types. Squish area is determined from the percentage of top squish surface Chapter 3. Experimental Apparatus and Method area of the piston bore: 7r * £ ) 2 / 4 . For the simple bowl-in- piston this is determined by: D 2 - d 2 %SAREA = D 2 * 100 where D is the piston bore and d is the bowl diameter. The bowl-in-piston pistons used in this study had a squish area of 70%. From similar calculations the bathtub type piston squish area was approximately 27%-29% allowing for the spark entry point. The effect of this difference in squish area is discussed i n Chapter 5. 3.2.5 I n s t r u m e n t a t i o n A schematic of the instrumentation layout is shown in Figure 3.6. Pressure measurements in the combustion chamber were made with a Kist ler 6121 piezo-electric pressure transducer mounted within an extended sleeve i n the cylinder head of the engine. The charge output signal was fed to a Kist ler model 5004 charge amplifier to yield voltage data proportional to the cylinder pressure. The pressure data were digitalized at a rate of 1 sample per 0.2 crank angle degree for the motored tests and at 1 sample per crank angle degree for the majority of the fired engine tests. The pressure signal was also displayed on a Textronic oscilloscope during operation. The transducer was recessed from the combustion chamber by half its diameter to protect it from the effects of thermal shock, as recommended by Brown, Benson and Pick [71, 72]. The selection, mounting and operation of the pressure transducer and amplifier system used in the Ricardo test cell were consistent with the 'ideal transducer specifications and recommended operation' of earlier investigators reviewed Chapter 3. Experimental Apparatus and Method in Chapter 2. The transducer specification and calibration curve are given in Appendix A . Hotwire voltage measurements were taken wi th a TSI 1226 high tempera-ture probe, suitable for high temperature environments, i n conjunction with a D I S A M - 1 0 anemometer bridge circuit i n the constant temperature mode. The signal was low pass filtered at 20 K H z with a D I S A 55D26 signal condi-tioner before being digitalized at a rate of 1 sample per crank angle degree. Hot wire sensor The sensor used a Plat inum-Ir idium wire, 6.3 micrometers i n diameter and 1.5 m m in length. The probe was inserted in the cylinder head through the spark plug hole using a two part specially machined swaged fitting sealing around the probe shaft. Two adapter fittings were required for the two different spark plug sizes used in the cylinder head. The sensor wire was positioned approximately 2 m m below the spark point as shown in Figure 3.7. The anemometer system specifications are given in Appendix B . A n A V L model 360c/600 optical pickup, mounted on the crank shaft was used to generate clock pulses every 0.2 degrees of crank angle and a trigger signal at B D C for the data acquisition circuitry. Volumetric air and fuel measurements were obtained from the differential pressure sensors across their respective laminar flow elements, displayed on inclined differential manometers in inches of water gauge. The manufacturers calibration curves and constants at standard normal conditions are given in Appendix A . The airflow meter was calibrated and compensated for pulsating flows. Engine speed was measured by a tachometer attached to the dynamometer Chapter 3. Experimental Apparatus and Method shaft and controlled by a set-speed potentionmeter on the control consol. A difference in the input setting and true engine speed sensed in the converter cabinet resulting in appropriate correcting control signals sent to the thyristor triggering circuits. The engine speed was therefore maintained at a constant, ± 0.2%, irrespective of the operator adjustments to throttle settings, fuel/air ratio or spark advance. Torque measurements were made by strain gauge load cells mounted on the torque arm. The output of these cells was provided i n continuous display at the control consol. Static calibration of the torque measuring system was carried out using a 20 N m calibration weight provided in the dynamometer pedestal. These calibration tests are discussed in the method Chapter 4. Ignition control and measurement were provided by a mult i turn dial with a range of 70 degrees B T D C to 20 degrees A T D C , sending a set point signal to the Lumenit ion electronic ignition unit . Addi t iona l pressure transducers and thermocouples mounted on the en-gine rig provided operating condition information and safety trips for the air intake and exhaust system, and the engine coolant and oi l modules. Control over the engine motoring and fired conditions was carried out by the Cussons electronic control system. Control over the acquisition process was carried out by the I B M P C , using signals from a majority of the above instruments fed through to the control room via a milspec connector. 3.2.6 Data Acquisition Two data acquisition systems were employed for this study. A slow data translation system was used primari ly for engine monitoring and performance Chapter 3. Experimental Apparatus and Method 37 information, while a fast, I S A A C 2000 acquisition system was used for cylin-der pressure, B D C signal and Hotwire voltage data collection. The 'slow' data translation system operated at 27.5 K H z and collected signals from the engine instrument analogue transducers, eg; airflow, fuel flow, speed, torque, and ignition advance. These signals were optically isolated and low-pass filtered at 60 Hz before passing the A D converter inside the P C . The hardware arrangement detailed in Figure 3.8 shows the circuit box containing the screen terminal boards DT752 and DT709 which connect to the data translation DT2801 analogue to digital converter board installed in the P C . The 'fast' acquisition system consisted of an I S A A C 2000 unit with 64 K of buffer memory, off-line block transfer of data and a flexible control system. Sampling rates of 200 k H z were available on four channels, using labsoft II software and the I B M P C . Data acquisition programs The data acquisition programs used i n conjunction wi th the test cell instru-mentation system were ini t ia l ly developed for the Alternatives Fuel Labora-tory by A.Jones. Further details of the system circuitry and programs may be found in the operations manual [69]. The slow data acquisition system was run using the ' D A T A Q ' program which scans the channels of the DT2801 board for sensor information. One hundred values were read i n from each channel and averaged. The signal volt-ages were then converted to the appropriate units and calculations made for engine performance parameters; Brake Power, Brake Mean Effective Pressure ( B M E P ) , Thermal efficiency, Brake Specific Fuel Consumption and relative Chapter 3. Experimental Apparatus and Method air fuel ratio. These data were displayed to the P C screen and updated every 5 seconds providing operating control information. The original programs were modified to provide ignition advance display also. Collected data for a desired set point were stored on disc drive B at a rate of 400 samples per channel. Similar conversion calculations and performance calculations are made on this data using the ' C R U N C H ' program. Data from the fast acquisition system were obtained using an interac-tive program, (a modified version of ' H O T W I R E ' [69]), allowing a variable number of channels, acquisition speeds and number of cycles to be obtained. Consecutive or nonconsecutive acquisition was also controlled integrally. Data trigger signal O n init iat ion of the system at the desired test conditions, the phasing of the acquisition was controlled by trigger B D C and clock pulses from the DT2801A board. The I S A A C unit was triggered to start acquisition from the 'next' B D C signal after an expansion stroke. The data stored by the I S A A C were transferred to the I B M P C prior to storage on floppy discs. The data collected on disc were then transferred to the Department's mini-computer, a V A X 11/750 unit using an 'Ethernet ' communication package. 3.3 Motored Engine Tests 3.3.1 Introduction Flow measurement tests were conducted in the engine, while motored by the D C dynamometer for each chamber configuration at the conditions given in Table 3.2. Al though most of the flow measurements were taken at three Chapter 3. Experimental Apparatus and Method different speeds for the wide open throttle condition ( W O T ) , gaps in the data set were the result of unstable operating conditions and associated difficulties wi th the measuring equipment. Hot wire anemometry techniques were used to yield velocity and tur-bulence information for the flow near the spark plug location. Digitalized hotwire voltages and cylinder pressure acquired at the rate of one sample per 0.2 crank angle degree, were recorded for 44 nonconsecutive cycles using the fast acquisition system and ' H O T W I R E ' program. During the data acquisi-tion procedure, engine operating conditions such as airflow rate and ambient pressure and temperature conditions were noted. The repeatable nature of the motored pressure trace made it possible to record single channel data on each floppy disc, thus the hotwire and pressure measurements were taken from two separate sets of 44 cycles wi th the engine operating at the same conditions. The motored pressure data required for the turbulence analysis were also used as a powerful check on the phasing of the trigger signal and acquisition and the operation of the engine. Operational procedures used in running the engine, specific to the mea-surements made are herewith described followed by details of the pressure and hotwire measurement methods. 3.3.2 Operational Procedures The extent of the engine changes required for each combustion chamber configuration, and the subsequent extended period over which the experi-ments were made, emphasized the need for standard operation procedures and checks to minimize or eliminate extraneous engine variables. The impor-tance of maintaining accurate calibration of the engine instrumentation, used Chapter 3. Experimental Apparatus and Method i n establishing the test condition, to obtain a reliable set of data was also recognized. Engine rebuilds were carried out in accordance wi th the manufacturers' recommendations [70]. These changes included fitting the appropriate piston, conrod, cylinder liner, spaces and cylinder head assembly using piston rings and head gasket reserved for each, checking of the belt drives, valve t iming, and oil and flushing of the coolant system. The newly machined pistons, modified bathtub and replacement bowl-in-piston, were fitted with new piston rings and run in as detailed in the method section 3.4. A l l motored tests were run at wide open throttle indicated by a leveling off of the maximum air intake flow. This was confirmed by a visual check of the throttle butterfly valve position. It was particularly important to reset the manual control of the throttle position for maximum gain when the engine was rebuilt from the diesel configuration with injector rack control. Pr ior to any measurements, the system was powered for a min imum of one hour to warm up the electronic circuits and then zeroed. The water heater and Senco l imi t valve 6 e n s o r were used to maintain the coolant temperature in the range 65-80 ° C during the motored tests. Init ial problems encountered in reaching this temperature resulted in a rectified valve position and resetting of the Senco l imits . Manual override was also used to ensure correct operating temperature. When using previously fired pistons, operation at low load and speed could also be used to heat the system. The standard start up procedure was then used with preheat of the lubrication oil to 40 ° C , partly open throttle and low speed settings. Chapter 3. Experimental Apparatus and Method Static calibration checks performed on the dynamometer torque arm sys-tem, using the 20 Nm calibration weight, rezeroing and resetting the gain as required, gave an acceptable error of 1%. Examination of the dynamic behaviour however, revealed a possible 'sticktion' problem in the trunnion bearings. Care was taken to repeatedly check the torque readings during testing for cold and hot operation, allowing settling time, or jiggling, prior to zeroing. This resulted in an non elimitable error of ± 0.6 Nm. 3.3.3 Pressure Measurements Motored pressure measurements were taken primarily to provide operating pressure and temperature data for calculation of the gas properties used in the analysis of the hotwire signal. The sampling rate of 0.2 crank angle degrees was therefore chosen to be compatible with the hotwire rate of acquisition. For each test speed 44 nonconsecutive cycles of data were taken. Each cycle was triggered from BDC covering the 720 degrees of the exhaust, intake, compression and expansion strokes. The fast acquisition system was used with the program 'HOTWIRE' and the data stored on floppy discs. Static calibration of the pressure transducer and amplifier system, using a dead weight tester over a limited range, was carried out prior to the engine tests, confirming the manufacturers' calibration constants. Calibration curves for the pressure transducers and testing details are given in Appendix A. The amplifier was operated with the sensitivity, mechanical units to volt-age dial, set to the manufactures calibration constant, gain set to 5.0 and on medium response. The pressure trace was also observed on an oscilloscope and monitored for correct transducer operation, eg., overload error evidenced by a faulty trace. During the course of the engine tests damage from engine Chapter 3. Experimental Apparatus and Method seizures and faulty pressure transducers led to extensive checking of the B D C signal, phasing and transducer units. 3.3.4 Hotwire Measurements The hotwire measurement process involved three stages; 1. Preparation and welding of the probe sensor wire. 2. Cal ibrat ion of the probe and anemometer bridge system. 3. Insertion in the engine and fast data acquisition. The harsh operating environment i n the motored engine combustion cham-ber resulted in frequent breakage of the probe sensor wire. The probe was subjected to rapidly changing pressures and high temperatures, stray oil and residue particles, engine vibration and a turbulent flow field. Unfortunately no wire was able to survive the vibration on shutdown or subsequent removal of the probe, frequently wire breakage also occurred during speed changes from test point to point or during the test. Preparation The sensor wire was prepared by spot welding the 6.3 micrometer P la t inum Ir id ium wire to the probe support needles using a D I S A 55A12 welding unit and microscope system. After welding and visual inspection for a clean weld, the sensor was annealed, through the anemometer bridge circuit at the oper-ating temperature of 600 degrees Celsius for 6-8 hours unt i l a stable resistance was recorded. The 600 degree overheat temperature was compatible with rec-ommendations of Witze [30] and the limitations of the plat inum i r id ium wire reviewed by Vines [73]. Chapter 3. Experimental Apparatus and Method The resistance of the wire and compensation for the probe and cable was determined using the bridge circuit and the operating resistance of the system obtained from: Rop — -Ramb(l + "(Top — Tamb)) where a was the thermal coefficient of resistance is specified by the manufac-turer. Calibration Calibrat ion of each wire sensor, with the anemometer bridge and low pass filter, was performed against a pitot tube in a small wind tunnel, range 0.5-16 m/s, at ambient temperature and pressure. The hotwire system registers change i n the heat transfer from the sensor due to the cooling effect of the flow. Being a function of the temperature differential between the sensor and the fluid flow, this heat transfer process is sensitive to the operating temperature and pressure and gas properties of the flow. Ideally calibration at al l operating pressures and temperatures should be made, this was however impractical . A n analytical model involving a Nusselt-Reynolds relation was therefore used to obtain empirical calibration coefficients of the form: Nu = A + B * Ren This approach enabled extension of the calibration data to the operating conditions. The techniques and theory behind this method are described in Chapter 4. Detailed information on the welding and calibration procedures, and a representative calibration data set and curve are given in Appendix B . Chapter 3. Experimental Apparatus and Method 44 To minimize the probe time in the engine and reduce possible breakage occurrences, the engine was first warmed up and stabilized at the test con-dition. On shut down the probe was inserted and the engine returned to the set condition as smoothly as possible. Calibration and operation of the anemometer was carried out with a 20 kHz low pass frequency filter to pre-vent aliasing of the signal. The acquisition rate was dependent on the engine speed ranged from 36 kHz to 90 kHz. 3.4 Fired Engine Tests 3.4.1 Introduction Performance and combustion pressure measurements were conducted in the fired engine fueled by natural gas, at the conditions and configurations given in Tables 3.3 and 3.4. The highest operating speed for the pressure measure-ments was reduced from 66.7 rps to 50.0 rps, after siezure of the bowl-in-piston piston. An examination of the engine indicated this was probably due to a combination of the weaker materials in the non-standard piston and failure of the water cooling system. Gaps in the data set for part load conditions were due to the difficulty of maintaining stable engine operation at those conditions. Engine performance parameters were also taken using the slow data ac-quisition system over these ranges. All tests were run at optimum ignition, minimum spark advance for best torque, MBT, determined by the operator. The fired tests were performed directly after the motored engine tests without altering the configuration. New pistons were the exception to this Chapter 3. Experimental Apparatus and Method procedure, where the cylinder head was removed and the piston surface in-spected for damage after the run in tests. 3.4.2 Operational Procedures In this section the fired operational procedures are described followed by details of the performance and pressure measurements. System checks and motored operation were carried out in similar manner to the motored tests and the recommended start up procedure at low load, partial open throttle employed. The ignition was initially set to 25 degrees BTDC and the gas flow controlled via the needle valve in the control room. The desired operating condition was then reached by slowly increasing the speed and coordinating fuel and throttle openings. Part load conditions were achieved by reducing the throttle and fuel in tandem. Stable engine conditions at the desired load and MBT, with coolant and lubricating oil held between 70-90 °C were maintained prior to any measurements. The new pistons, ie; replacement bowl-in-piston and single slot modified bathtub, were run in under a combined part and full load schedule to a max-imum speed of 70 rps over 20 hours. The procedure followed Ricardo man-ufacturer recommendations and previous piston experience, to ensure proper seating and sealing of the new rings [70]. The slow data acquisition system previously detailed, was used to moni-tor engine operation with intermittent checks on the torque calibration and updating of the ambient conditions made during the test process. Chapter 3. Experimental Apparatus and Method 46 3.4.3 Performance Measurements A t each test point, after stable operation for a min imum of ten minutes, en-gine operating data, processed v ia the data translation board from instrument analogue signals, were taken using the slow acquisition system. Similar perfor-mance parameters to those displayed on the P C monitor, were then available using the ' C R U N C H ' program on the stored converted signals which had been averaged over 400 samples. The performance parameters stored were, speed, Brake power, B M E P , Bsfc, thermal efficiency, torque, ignit ion advance, air-flow rate, natural gas flow rate, relative air fuel ratio and ambient pressure and temperature data. 3.4.4 Pressure Measurements The fired engine cylinder pressure data obtained at each test point was taken for use in combustion measurement and cycle-to-cycle analysis. For this rea-son consecutive cycle acquisition was desired. The pressure transducer and charge amplifier system of the motored tests were used wi th the amplifier gain set to 10.0 to prevent overloading of the acquisition system. The analogue signal from the amplifier was digitalized at a rate of one sample per crank angle degree. During consecutive cycle acquisition the Bot tom Dead Centre ( B D C ) signal was used to trigger the fast acquisition at the start of the first data set only. Subsequently the B D C signal was also recorded on a second channel for use as a phasing check on the pressure signal. The pressure signal processing method provided a check on the position of B D C , relative to the pressure signal, and rejected out of phase cycles. For Chapter 3. Experimental Apparatus and Method data acquired per crank angle degree this resulted in a maximum of 3 cycles being rejected. However it was found that data acquired per 0.2 crank angle degree were liable to major phasing errors, therefore only nonconsecutive data were taken at this rate. The original 'HOTWIRE' program was modified for the acquisition of consecutive data and expanded for 200 cycles. These were stored on two floppy discs per test condition. Ambient pressure and temperature and the air and fuel flow rates from the differential pressure manometers were also recorded for use in the analysis. Extensive data were obtained from each phase of the tests for the six chamber configurations over a period of 20 months. The schedule for this stage of the investigation was greatly extended due in part to the difficulties encounted with probe breakage but also to the range of operating conditions studied for each configuration. This however, allows for comparison with studies over a broad range. The techniques used to analyse the data are discussed in Chapter 4. Chapter 4 Data Analysis 4.1 Introduction The techniques used i n the analysis of the collected data have been divided into two areas, first those used in the processing of the raw analogue signals to render raw velocity and pressure information, and second those used in the interpretation of the raw data through the selection of characterizing parameters. The method used on the motored engine test data to provide data acquisition phasing checks and flow field information is first detailed followed by the fired engine test data analysis methods used on both the 'fast' and 'slow' aquired performance data. Whi l e the selection of the various techniques and choice of input parameters of previous research studies, have been primari ly covered in the introduction, some further general comments are included in this chapter. 4.2 Motored Engine Tests 4.2.1 Analytical Procedure The sequence of events required in the analysis of the motored engine test data is given i n Table 4.1. The programs and methods used for this por-t ion of the study were based on similar studies of turbulence measurement in 48 Chapter 4. Data Analysis spark ignition engines and a rapid compression machine carried out by grad-uate students for the Alternatives Fuels Laboratory Group [17, 18, 74, 75]. Modifications to the base programs have been made for successive experi-ments, specific to the equipment dimensions and experimental variables, in this investigation for example, in the acquisition rate and number of cycles acquired. The calculations were carried out on a V A X 11/750 following transfer of the motored pressure and hotwire data from the P C stored floppy discs. Pro-cessing of the pressure analogue signal was carried out in the P R E S S - A N A L program, after a conversion program, I S A A C 2 V A X , was used to condense the indiv idual cycle data into one file and strip the header information. A smoothing-filtering program was then used on the second half of the data set prior to calculation of the cylinder temperature, the area of interest being the compression and expansion strokes. The ensembled cylinder pressure and temperature trace data were then used i n H W - A N A L with the raw hotwire data, also reduced, and the calibration constants from H W - C a l , to yield raw velocities. Flow field analysis parameters were then calculated in the T U R B U L E N C E program. A n additional program was used for logarithmic plots of pressure versus volume to check phasing of the pressure signal data with the cylinder volume assignment, using the crank angle position from the optical pickup. 4.2.2 Pressure Signal Processing The digital data analogue signal was converted to a relative pressure by a scaling factor, a function of the gain and charge amplifier setting, ie., me-chanical unifs/volts . Following recommendations by Lancaster et al [76] the Chapter 4. Data Analysis cylinder pressure at B D C after the intake stroke was then equated to a refer-ence condition, a simple equivalent to the pressure in the intake manifold, to obtain the absolute cylinder pressure. The average manifold intake pressure was assumed equal to the ambient pressure times the volumetric efficiency. PBDC = Pamb * f]v The pressure in bars was determined by: * - = < ^ - i o ) . . . o where the analogue signal is digitalized 0-4096 for -10 to +10 volts wi th the gain set to 5.0 for the motored tests. The volumetric efficiency for a four stroke engine is defined as the ratio of the actual mass of air supplied per cycle to the theoretical mass of air required at standard conditions ( 15 ° C , 101 kPa) . Volumetric efficiency varied with combustion chamber and speed, and was calculated using measured flow rates by: 2 * Qair Vv = V. * RPS where QaiT is the total amount of air m 3 / s e c supplied and Vt is the volume swept. The scaled pressure data were ensemble averaged over the 44 nonconsec-utive cycles prior to use in the temperature calculations. Information on indiv idual pressure traces was used in a separate program for phase checking. The importance of accurate total volume calculations with pressure phasing is highlighted in its use in the assessment of the work inventory of an cy-cle, ie: I M E P calculation. Logarithmic plots of pressure and volume provide information on the phasing accuracy. Chapter 4. Data Analysis The main inaccuracies in pressure assignment and their evidence may be summarized • Improper reference pressure, exhibited as curvature in the compression stroke, which is a scaling problem only. • Improper clearance volume, exhibited as curvature in the end of the compression stroke. • Improper phasing, exhibited by the crossing of the compression and ex-pansion slopes, this may also indicate a faulty transducer. Extensive checks were carried out on the pressure transducer system used in this work by the author and concurrent users of the equipment [77], re-sulting in the removal and replacement of suspect transducers and data. The occurrence of peak pressure in the motored traces were found to be within 2 degrees of T D C , the discrepancy being due to irreversibilities due to heat transfer, confirming correct phasing. Maintenance of the same clearance volume, and subsequent compression ratio, i n the different combustion chambers, checked by l iquid displacement measurements of the cavities, was confirmed by the similarity of the motored pressure traces. It was noted, however, other effects such as varying heat loss to the cylinder wall and head, can cause differences in the traces. A comparison of the motored pressure data was also useful in assessing the reliability and repeatability of tests carried out over an extended period. A partial smoothing-filtering program was used on the reduced ensem-bled pressure data, in the in i t ia l region, B D C to 60 degrees B T D C , prior to temperature calculation. Details of the filtering method used, and the op-tions examined, are given in Appendix D . The smoothing was done to remove Chapter 4. Data Analysis 52 transient effects in the region of inlet valve closing, possibly due to vibration, without affecting the critical region near TDC. The gas cylinder temperature was determined indirectly through its poly-tropic relation with pressure. Previous investigations [74] have examined various methods assuming: • Adiabatic compression and expansion. • Polytropic compression and expansion using calculated values of fc and • Adiabatic compression to IVC, application of the perfect gas law prior to EVO and adiabatic expansion after EVO. Witze [30] showed that adiabatic computation was appropriate during the compression stroke and increasingly poor after 50 degrees after TDC, through comparison with actual measurements. The last assumption, using the ambient temperature as the reference temperature at intake BDC, was taken in this study, and the temperature determined by: P(i) ^ intake BDC to IVC T{i) = Tamb * PB DC IVC to EVO 7/(0 = T{i - 1) * p . P ( t ) * V ^ P(i-1) V(i-1) P(i) ^ EVO to expansion BDC T(i) = T(i - 1) * v The polytropic compression and expansion coefficients differ from isen-tropic for air for a variety of reasons. The compression coefficient (7,.) is less than isentropic due to residual gas and increased temperature. The expan-sion coefficient (7 e ) is less than isentropic due to residual reactions and higher temperature. Chapter 4. Data Analysis In the motored engine residual gas reactions are not in question, how-ever the polytropic coefficient does vary with engine speed and combustion chamber. The isentropic coefficient for air at standard conditions was 1.365. In tr ia l runs 7 C and 7 e were both set to 1.35 consistent with previous work by Dymala-Dolesky [18] and current bathtub chamber work by K a p i l [77] who used values of; 7 C : 1.35 to 1.352 and 7 e : 1.38. In the final analysis the motored pressure compression coefficient was calculated from the logarithmic slope of the pressure from the fired cases, averaged over speed and stoichiometric ratio for each configuration. The expansion coefficient of 1.35 was used for all cases. The compression values ranged from 1.27 to 1.37. 4.2.3 Anemometer Signal Processing The analogue hotwire signal registered change in the electrical resistance due to the heat transfer from the wire sensor due to the cooling effect of the flow. This heat transfer is a function of the temperature differential between the sensor and the fluid flow and is sensitive to the gas properties, themselves de-pendent on the pressure and temperature in the measurement regime. Hence it was possible to determine the gas velocity at the point of measurement by applying an analytical correction to the anemometer output for the separately determined gas properties at the particular operating conditions. A n empirical Nusselt-Reynolds relationship formed from the calibration data has been used. A n outline of the techniques and theory involved in this analysis follows. Heat transfer from the wire occurs through radiation, bouyant and forced convection and conduction along the wire to the supports. The equations of Lancaster [13] have been used i n this analysis. These were based on work Chapter 4. Data Analysis by Collis and Wil l i ams [78] and Davies and Fisher [79] on heat transfer from electrically heated cylinders applying modifications to Kings Law, which re-lates the Nusselt number to a Reynolds number based on the wire diameter and gas properties: Nu — a + bRei These studies made the assumptions that, conduction to the supports may be neglected for sufficiently long wires, ie., with length to diameter ratios in excess of 200, radiation may be neglected and that the convective bouyancy was important i n low speed flows only. The early methods solved the one dimensional heat balance equation to obtain an expression for the mean wire temperature wi th the convective heat transfer coefficient, h, solved iteratively. In a different approach Lancaster used a I D energy equation involving the gas temperature, resistance of the wire at zero condition and the current, for the heat distribution along the wire. This equation was then integrated to find an expression for the mean temperature with h unknown. The coefficient h, found by an iterative method was used to evaluate the Nusselt number. App ly ing the calibration coefficients to the analytical model the Reynolds number was obtained and hence raw velocity data were calculated. The selection and calculation of the free stream gas temperature for the gas property determination, described in the pressure section 4.2.2, was suggested by Witze [30] as giving the best agreement wi th L D A studies. Chapter 4. Data Analysis 4.2.4 Flow Field Data Analysis Returning to the definition of turbulence as a "fluctuating velocity component superimposed on the bulk velocity of a flow" [24], the means of defining the turbulent intensity, a characterizing parameter of the turbulent flow field, is subject to debate. Many different definitions have been used i n both H W A and L D A analysis [13, 22, 35, 80, 81]. Generally these methods may be divided into ensemble averaging methods and cycle resolved methods. The l imitations of ensemble averaging a nonstationary process, where the ergodic definition of the ensemble average being equal to the time average does not hold, is evident in the analysis of engine turbulence. The periodic nature of the mean flow causes cyclic variations in the mean to be included i n the estimation of turbulence. In contrast the nonstationary description of Lancaster [13] divides the instantaneous velocity into a mean, time averaged mean velocity and a turbulent component. A cycle-by-cycle nonstationary time averaging method developed by Cata-nia and M i t t i c a [82] was used to extract turbulence data without including the cyclic fluctuations of the mean flow. This method involved the ensemble time averaging of the raw velocity in a defined period, or window size, and the fitting of a smoothed velocity curve similar to the method of Rask [80]. The following outline describes the procedure employed in the turbulence analysis and the velocity component definition. The separation of a mean velocity per cycle was obtained by evaluating the mean from the instantaneous velocity in a given window by trapezoidal integration. A cubic spline curve fitting routine was applied to the mean value at the center of each window and then interpolated. These traces were then Chapter 4. Data Analysis ensemble averaged over the number of cycles to give a representative mean. The average at each window midpoint was defined by: 1 U(tw,i)=- U(t,i)dt Cubic spline curve fitting and interpolation yielded: Uw{t,i) and the window ensembled mean calculated from: i y t=i The fluctuating component of the velocity was defined by: uw(t,i) = U(t,i) - Uw(t,i) The squared average fluctuation at each window midpoint was then evaluated by: [«(*-,«)]' = 4 r ? \U{t,i) - Uw(t,i)]2dt These were ensemble averaged over N cycles and the square root taken: u(tw) 1 N Cubic spline curve fitting and interpolation was again applied to yield: uw{t) a representative rms intensity. Ensemble averaging was also carried out for comparison, where the en-semble mean is defined by: uE(t) = -Ni:u(t,i) The fluctuating component was separated by: uE{t,i) = U(t,i)-UE{t) Chapter 4. Data Analysis and the rms intensity defined by: Mt) = ^^Yt\u(t,i)-uE(t)i) In addition the relative turbulent intensities, referenced to the mean flow and representative kinetic energy terms were calculated. Scaling of point data with the mean piston speed defined by: Sp = 2rN has been used as a quantifying characteristic in some studies. However Lan-caster cautions that "quantitative interpretation of these single point mea-surements is unwarranted" [13]. This calculation has been performed for comparative purposes only. In selection of the window size to separate the 'turbulent' fluctuating velocity component from the mean and cyclic variations, a balance must be maintained. Too small a window size may result i n the loss of low frequency turbulence, ie., having been interpreted as mean flow, too large a window size and the mean flow .motion may be attributed to turbulence. Catania and M i t t i c a [82] showed that a window size i n the range 4 to 12 crank angle degrees had l i t t le effect on the mean, while the 8 to 20 range was insensitive for the intensity, giving consistent values. A 12 degree window was used for this work consistent wi th previous work by Dymala-Dolesky [18] and supported by a sensitivity check. Figure 4.1 shows the effect of a range of window sizes on the turbulence intensity. W i n -dow sizes between 4 and 18 degrees had a negligible effect on the mean while the window sizes between 8 and 15 degrees gave the most consistent value of turbulent intensity. Chapter 4. Data Analysis 4.3 Fired Engine Tests 4.3.1 Analytical Procedure Data from the two acquisition systems were analysed to yield engine per-formance information. The primary object of this work was to obtain in-formation of the effect of the various combustion chamber shapes on cyclic variations. This is one of the main l imit ing factors to the operation of an en-gine at the extended l imits of lean mixtures, particularly with slow burning fuels such as natural gas. Analysis of the fired pressure data was made to provide a characteristic measure of this variability, ie; C O V of I M E P , and to provide a comparison wi th engine performance, ie; fuel consumption and efficiency. The majority of the analysis was performed on the fired cylinder pressure data; a fundamental engine variable. The sequence of events followed in the analysis are shown in Table 4.2. After processing of the analogue pressure signal in a similar manner to that used in the motored engine tests, characterizing parameters were calculated, eg; I M E P and peak pressure, these were ensembled averaged and statistical analysis carried out to quantify the cyclic variations. Combustion, an inter-mediate phenomenon, was also investigated through a mass fraction burned analysis. Details of the techniques used, and the choice of the characterizing parameters are described in the following sections. The analytical procedure has been illustrated for the data acquired over 200 consecutive cycles at a rate of one sample per crank angle degree. The data obtained at the rate of one sample per 0.2 crank angle degree for 44 cycles were reduced to the lesser rate and minor modifications made to the Chapter 4. Data Analysis program. 4.3.2 Pressure Signal Processing The method for converting the digital analogue signal to absolute pressure data was similar to the motored pressure process using a gain factor of 10.0. The reference pressure was again taken as the pressure in the intake manifold at B D C calculated by: T PBDC = Pamb * V v * ~^ a m b _ 2 * Qtotal Vv ~ V. * RPS where Qtotal was the total amount of air and fuel in m3/sec supplied and Vt was the volume swept. A smoothing routine similar to that for the motored pressure data was applied to the data i n the inlet valve closing region up to 30 degrees be-fore spark, transient spikes in this region causing errors in the mass fraction burned analysis. This smoothing routine was performed within the F I R E R U N program. The individual cycle pressure data wi th volume assignment was then used to calculate performance parameters. Ensemble averaging over the total num-ber of cycles also provided a representative fired pressure trace. 4.3.3 Performance Data Analysis Two measurements of performance data taken: on line monitoring through the slow acquisition system, and cylinder pressure data v ia the fast acquisition system. The on-line monitoring was used primari ly as an indication of the Chapter 4. Data Analysis engine operating efficiency and stability over a wide range of speeds and fuel ratios, while the fast acquisition system was used in characterizing that performance at a given speed and fuel ratio. The analogue signals corresponding to engine speed, torque, natural gas and air flow rates and ignition advance taken at 27.5 k H z and averaged over 400 samples by D A T A Q were scaled using instrument calibration factors and the flows corrected for standard pressure and temperature given ambient con-ditions. The ambient temperature used was an average of the manifold intake temperature measured by thermocouple and the test cell room temperature. The C R U N C H program was then used to evaluate the performance pa-rameters. Brake Power, B M E P , Bsfc, relative air fuel ratio ( R A F R ) and brake thermal efficiency were calculated: Brake Power =Torque * Speed *2TT/1000 B M E P =Torque *4TT / (L /2) Bsfc =mg * 3600/Brake Power T]tn =Brake Power/rri f l * L H V A brake power ( B P ) correction for standard conditions assuming 85 % mechanical efficiency in accord wi th S A E J1349 requirements was then per-formed, being defined by: B P c o r r e c t e d = B P * (1.18 * ( ^ 9 - ) ( ^ ) ° B - 0-18) - T o m b -^ yo The composition of the fuel, B C Natural Gas, and the calculation of the Lower Heating Value ( L H V ) used in this analysis is given in Appendix C. The Indicated Mean Effective Pressure ( I M E P ) was calculated from both the individual pressure traces and the ensembled trace generated from the Chapter 4. Data Analysis analogue processing. Statistical analysis was performed to calculate the mean, standard deviation and the coefficient of variance ( C O V ) of I M E P from the ensemble average. where VCBDC was the volume at B D C before compression and VCBDC was the volume at B D C after expansion. The I M E P coefficient of variance was determined by: This statistical analysis was also applied to the peak pressure data and the angle of occurrence of the peak pressure. I M E P and its C O V was selected as the direct relation between engine drivability and fluctuations in the output torque ie. I M E P C O V [19]. Peak pressure and position were also used for comparison wi th previous squish jet work by Dymala-Dolesky [18] and with other researchers [1, 36, 45, 46]. The conditions at which the engine operates have been shown to influence the governing characteristics. Statistical studies by Belmont et al [54] using autocorrelation techniques over a wide range of conditions show, that i n the lean burn operating regime al l of the primary cycle features are strongly related. The preferred sample size used in the fired analysis was 200, in line wi th Amann's recommendations of a min imum of 100 and recent research by Matekunas [45] 120, Nakamura [68] 200, and Lancaster [76] 100-300 and wi th in the capabilities of the acquisition and computation systems. Very large sample sizes have generally been restricted to frequency information [76] or I M E P the characterizing parameter for cyclic variation from those reviewed, due to Chapter 4. Data Analysis with on line point data acquisition, ie., peak pressure measurements by Pat-terson [39] and Belmont et al [54] with 2048 cycles taken. Increased sampling error was evident in the smaller data set of 44 cycles. However theses data were useful in comparison with tests taken on the same pistons, using this number. A cautionary note is supplied by Brown [71] "Consistency is no in-dication of accuracy". Lancaster also maintained that the number of cycles of data required was dependent on the use of the data and the stability of the running condition. 4 . 3 . 4 Combustion Analysis The final phase of analysis of the individual pressure traces was used to inves-tigate the combustion process. The mass fraction burned was evaluated based on a simple procedure developed by Rassweiler and Withrow [49] from the correlation of cylinder pressures with flame front photographs. After extract-ing the pressure rise due to piston motion from the pressure records Rass-weiler and Withrow found that the fractional pressure rise due to combustion was approximately equal to the fractional mass burned at the corresponding instant i n the combustion period. Using this technique i n the current work, the effect of piston motion and combustion were evaluated as though taking place sequentially not simulta-neously. p(i) A P C O M B piston O Chapter 4. Data Analysis The pressure increase due to piston motion was calculated assuming poly-tropic compression and expansion by: which was extracted from the pressure record. The increase of pressure due to combustion alone was then evaluated as occurring at a constant volume at each stepped volume. The change from variable volume to constant volume was made by scaling to a reference volume, ie. Volume at T D C . Rassweiler and Withrow justified this from earlier work by G . Brown in 1925 showing "Pressure increase produced in a const ant-volume bomb by liberating a given amount of energy i n a given mass of charge is inversely proportional to the total volume of the bomb". v = v ( { + *) + y ( 0 * _L_ 2 VTDC APcomb = {AP(i + 1) - A P ^ ) * Vref The mass fraction burned was then calculated by dividing the summed pressure increase due to combustion at each point by the total pressure in-crease due to combustion: Mass fraction burned(i) = ^ eomb(') W Y,APcomb{total) B y definition, this simplified procedure scales the mass burn fraction be-tween 0 and 1. Studies by A m a n n [19] show good agreement of this method wi th more advanced computer techniques based on multiple flame zone heat release models. Start of combustion was taken at the crank angle of spark occurrence. E n d of combustion was determined where the pressure increase due to combustion was zero or negative. Chapter 4. Data Analysis 64 Several methods of polytropic exponent calculation were investigated: • Experimental determination of 7 C and 7 e and an average used as per Rassweiler and Withrow; 7 of 1.3 [57]. • Experimental determination of 7 C prior to spark and 7 e from end of combustion with a linear relation used in the intermediate region, as per Dymala-Dolesky [18]. • Arb i t ra ry choice of an appropriate 7, as per Boisvert; 7 C of 1.3 [74]. Rasswieler and Withrow found that changes between 1.25 and 1.35 had l i t t le effect on their analysis, and noted that for the equations for constant vol-ume bomb development to hold, the same polytropic exponent was required in the burned and unburned region. In this study the polytropic coefficient for compression was calculated from the average slope of the logarithmic pressure-volume curve i n the range ten degrees before spark. This value was used throughout. A sensitivity analysis showed that individual values of -yc calculated from each pressure trace were required in the massburn calculation wi th an average value of 7 C only suitable for the ensembled pressure calcula-tion. Figure 4.2 shows a typical plot of the polytropic coefficient versus crank angle. Characterizing combustion parameters of the in i t i a l burn period and burn duration were calculated using the ranges 0 to 1% and 0 to 5% and 1% to 90% and 5% to 90% respectively. The two ranges were calculated to provide a wider basis for comparison with others in the arbitrary selection of ini t ia l and main burn period. Statistical analysis of the mean, standard deviation and C O V of these parameters was carried out similarly to the pressure data calculations. Chapter 4. Data Analysis The findings generated in the motored engine and fired engine tests of the six chamber configurations are discussed in the next chapter. Tables 5.2 to 5.24 and Figures 5.1 to 5.100 present the test results and performance measurements obtained ove a period of 20 months using two data acquisi-tion systems. As noted in Chapter 1, the range of tests and measurement parameters selected for this investigation was designed to allow the broadest comparison with conclusions drawn from previous combustion and turbulence studies. Chapter 5 Experimental Results and Discussion 5.1 Introduction In this chapter the results obtained from the experiments conducted on the Ricardo research engine are presented and discussed. In the first two sec-tions the results from the motored and fired tests are presented. The third section evaluates the uncertainties of the measuring techniques and the final section discusses the relation between the flow field measurements and the performance measurements. The six combustion chamber geometries are compared over the range of speeds, load conditions and air fuel ratios given in the method section. Com-parisons are made between all six configurations and by separation into the two groups, 'bathtub' and 'disc' defined in section 3.2.3. 5.2 Motored Tests 5.2.1 Motored Pressure Results The motored pressure data taken by the fast acquisition system were used pr imari ly to determine the gas properties required in the hotwire analysis of the in-cylinder flow field. These data also provided a powerful check on the phasing of the analogue signal of interest, either pressure or hotwire, and the acquisition trigger signal. 66 Chapter 5. Experimental Results and Discussion The comparison of the motored pressure traces for the different chambers also gives an indication of the similarity of the clearance volume of the cham-bers. In their heat release analysis of engine pressure data Gatowski et al [50] confirmed that the cylinder pressure, in addition to combustion effects, responds to changes in the volume of the combustion chamber, heat transfer to the walls and mass leakage. Volume and compression ratio Figures 5.1 to 5.4 show the motored pressure trace for the various chambers at four speeds, for the 360 degrees of crank angle from B D C of the intake stroke to B D C of the expansion stroke. Zero crank angles is assigned to T D C of the compression stroke. These figures show that the maximum developed pressure, at each speed, was similar for the different chamber configurations, wi th the exception of the disc piston. Volumetric checks of the combustion chamber cavities and clearance heights indicated that the clearance volume was wi th in 5% for al l the chambers except the disc. The compression ratio for these configurations was approximately 9.0:1. In situ depth measurements of the disc configuration revealed a compression ratio closer to 10.0:1. Aside from any difference in the compression ratios, a difference in heat transfer from the larger surface area of the bowl-in-piston and bathtub type chambers may also produce a decrease in pressure. This is supported by the lower peak pressure of the squish jet piston. This chamber had the largest surface area, estimated to be 30% greater than the bowl-in-piston without channels [18]. Figure 5.5 shows the effect of engine speed on the motored pressure for the single slot chamber. This figure shows an increase i n peak pressure with increased speed due to the shorter time for heat transfer to the cylinder walls. Chapter 5. Experimental Results and Discussion O n al l the pressure plots a high frequency fluctuation in the pressure trace is seen at approximately 130 degrees B T D C , increasing in extent at the higher speeds. This is the point of inlet valve closing and the fluctuation is attributed to vibration of the valve. Whi l e this fluctuation had no effect on the overall pressure trace, it proved detrimental to the temperature calculations based on the pressure data. Changing the temperature calculation after valve closure from an assumption of adiabatic compression to the use of the ideal gas law, meant that a reference temperature could be taken from a spike. This section of the pressure trace was smoothed prior to the temperature calculation. Temperature Figure 5.6 shows the calculated temperature at 33.3 rps for al l chambers. The temperature profile for the disc st i l l shows some fluctuation at the end of the filtering region, 60 degrees B T D C . The filtering routine was restricted to the most crit ical region to prevent the loss of information in the area of greater interest,around T D C . Table 5.1 lists the compression and expansion coefficients averaged over speed and R A F R used in the temperature calculations. The choice of these coefficients is discussed in the analysis section 4.2.2. 5.2.2 Flow Field Results The flow field measurements, taken approximately 2mm below the spark plug position, have been represented by window ensembled mean velocities and intensities per crank angle for the compression and expansion strokes. The differences i n the velocity and intensity profiles calculated using a stationary, cycle-by-cycle ensembled method, and those calculated by non-stationary, Chapter 5. Experimental Results and Discussion window ensembled, techniques is shown in Figure 5.7. The window averaging and interpolating techniques used in the analysis have the affect of reducing the spikes in the mean velocity without changing the magnitude of the over-all profile. The difference in the analytical methods is more pronounced in the intensity plot. Without the removal of the cyclic fluctuation from the instantaneous velocity, the turbulent intensity may be grossly overestimated. The positions of the inlet valve closing, I V C , and the exhaust valve opening, E V O , are also shown on these plots. Figures 5.8 and 5.9 show the effect of the piston chamber geometry on the velocity and turbulent intensity at 33.3 rps. These and all flow field measurements were taken at W O T conditions. Although these figure show very different traces around T D C from the various chamber configurations, they also show the common high velocity and intensity generated through shear past the intake valve, relaxing after valve closure and decaying to T D C . A rapid increase i n velocity and intensity is also evident after the exhaust valve opens. Combustion chamber geometries A detailed description of the effect of the different combustion chamber ge-ometries on the in-cylinder flow field follows. Figures 5.10 to 5.17 compare the six configurations i n two groups at 20.0 rps and 33.3 rps. Measurements at a higher speed, either 50.0 rps of 66.7 rps obtained for four of the cham-bers are given in Figure 5.18 to 5.23. These later figures show that the mean velocity and intensity generally increase wi th engine speed. This tendency is clearly evident during the intake stroke though to T D C . However after T D C , this trend is less obvious, showing some crossing over of traces such as the Chapter 5. Experimental Results and Discussion castellated piston at 20.0 rps. Bathtub chambers Figures 5.10 and 5.12 show the mean velocity profiles for the bathtub and modified bathtub piston. These figures show a decrease in, or the lack of formation of, the high mean velocity prior to TDC, with a sustained relatively high mean after TDC, for the modified pistons. Figures 5.Hand 5.13 show a similar tendency in the intensity profiles. A possible reason for this tendency is the addition of a slotted ridge to the standard flat piston, used with the bathtub cylinder head, promotes early break down of the large scale intake generated turbulent motion. The ridge may prevent the formation of large scale rolhng motion within the chamber such as 'barrel' motion, a source of energy at a latter stage. These figures also show the predominant squish motion of the bathtub, evident by a 'hump' displaced after TDC. The essentially one sided squish area of the bathtub configuration, and the location of the hotwire probe at one side, has been suggested as a plausible cause for the displacement of this hump [74]. Disc and bowl-in-piston chambers The non squish disc configuration and the enhanced squish and squish jet bowl-in-piston type chambers are described next. The mean velocity and tur-bulent intensity profiles for these pistons are compared in Figures 5.14 and 5.17. All chambers showed a decrease in the mean velocity form IVC reach-ing a minimum at TDC. The squish action generated by the bowl-in-piston Chapter 5. Experimental Results and Discussion chamber configuration produced a visible local increase in the mean at T D C , however this was substantially lower than the other chambers. W hi l e the disc chamber phenomenon! may be attributed to the higher compression ratio, the disparity with the squish jet mean velocity profile was unexpected. The performance and combustion characteristics of this piston, discussed i n the following sections, do not support such low values. Exami-nation of the probe position i n the bowl-in- piston measurements, shown in Figure 3.7, suggests that the wire location of 4.7 m m below the cylinder head may be outside the region of the plane squish flow. The destruction or ab-sence of a main squish motion at T D C , ie a 'hump', by the added channels in the squish jet piston was consistent wi th previous evaluations of this design [17, 18]. scaling with mean piston speed Figures 5.26 to 5.37 give the mean velocity and turbulent intensity profiles scaled with the mean piston speed. The range of the scaled values at differ-ent crank angle positions, for example T D C compared to 30 degrees A T D C , illustrates the l imitat ion of using single point measurements to determine characteristic parameters from the flow field. The profiles for the mean ve-locity generally 6 h o w the scaled values in the region of T D C to be between 0.2 and 0.5. These are comprable wi th the findings of Bopp, Vafidis and White law [22]. However, the scaled turbulent intensity values in the region of T D C , ranging from 0.05 to 0.2 are substantially lower. This reduction may be partly due to the rms intensity evaluated using window averaging techniques being lower than ensemble averaged values. Differences in the geometry of the chambers used and in the position of flow measurement are also significant Chapter 5. Experimental Results and Discussion factors i n the range of values obtained. 5.3 Fired Tests 5.3.1 General Performance Parameters Engine performance parameters, calculated in accord with SAEJ1349, are given in Tables 5.2to5.9 for stoichiometric and lean operation. Full load performance Tables 5.2 to 5.4 compare the different piston geometries at the three speed, 20.0, 33.3 and 50.0 rps for W O T full load operation. These tables show marked differences in the brake power and thermal efficiency values for the various chambers. A t al l conditions the bowl-in-piston chamber gave the best performance and the squish jet or castellated pistons the poorest. These tables also show an improvement in the relative performance of the single slot piston with speed. The brake specific fuel consumption (Bsfc) displayed a similar trend, being lowest for the bowl-in-piston configuration and generally highest or least economical for the squish jet piston. Part load performance Table 5.5 details part load operation at 2.5 bar bmep at 33.3 rps. This measurement also shows that the bowl-in-piston chamber has the highest brake power and efficiency and the castellated piston the lowest. However the squish jet piston performance is now closer to that of the bowl without channels. Chapter 5. Experimental Results and Discussion 73 Table 5.6 comparing three piston geometries at the part load condition of 3.5 bmep at 50.0 rps again shows the bowl-in-piston produced the best performance, the single slot piston comparable, and the castellated piston gave the poorest. These part load cases also show that there is some variance in maintaining the part load condition at a pre set B M E P . For stoichiometric operation, Table 5.5 shows a 12% range. This however, does not appear to be the cause of the difference in performance between the chambers. The castellated piston exhibited relatively poor performance when operating at both higher and lower values of B M E P than the other pistons. In all cases the single slot piston performed as well as or better than the castellated piston. For the lean cases the single slot chamber approached or exceeded the standard bathtub chamber performance. Brake thermal efficiency A closer examination of the effect of piston chamber geometry on the brake thermal efficiency at different air fuel ratios for full load operation is made i n Tables 5.7 to 5.9, and illustrated in Figures 5.38 to 5.43. A discussion of these results follows. Figure 5.39 shows a continuous rise in efficiency with leaner mixtures for the single slot chamber, while the bathtub chamber displays constant or de-creasing efficiency after a relative air fuel ratio of 1.2. This same trend is shown at 20.0 rps, Figure 5.38 and 50.0 rps, Figure 5.39 with a pronounced improvement of the single slot chamber exhibited at the higher speed com-pared with the performance of the bathtub chamber. These figures also show an improvement in efficiency with the castellated piston at the higher speeds, though it was st i l l the lowest of the six chamber designs tested. The data Chapter 5. Experimental Results and Discussion for the bathtub chamber at 50.0 rps were taken from multispeed tests rather than from a single speed versus relative air fuel ratio set. These data however were consistent over the range examined. The next group of Figures 5.41 to 5.43 illustrate the results for the disc and bowl-in-piston type chambers. These figures show the same trend in brake thermal efficiency, at increasingly lean operation as discussed for the performance tables. A t all speeds the bowl-in-piston chamber had the highest efficiency, increasing with leaner mixtures, the disc chamber had a constant, lower efficiency and the squish jet piston the lowest. F ina l ly comparing the bathtub and disc group of chambers at 33.3 rps, shows the bathtub chambers to be similar to the disc chamber. Ignition advance The effect of chamber geometry on the amount of ignition advance required for M B T operation is also shown in Tables 5.7 to 5.9. This is a useful comparison because reduced ignition advance is an indication of faster burn rates. In general ignit ion advance shows less pronounced trends between the bathtub type chambers than the disc bowl-in-piston group. Figures 5.44 to 5.46 show a similar advance required for the single slot and the castellated pistons, increasing slightly wi th leaner mixtures, with the bathtub chamber requiring more advance at the higher speeds. The reduced advance required by the bowl-in-piston type chambers suggest a possible benefit of these types of chambers in the extension of the knock l imi t . Where slower burn rates require very early advance, the chance of preheating of the end gases leading to knock is more prevalent. Chapter 5. Experimental Results and Discussion 75 5.3.2 Fired Pressure Results Fired pressure traces Figure 5.51 to 5.68 show the ensembled fired pressure trace versus the crank angle position for the 360 degrees from B D C before the compression stroke to B D C after the expansion stroke. These plots are divided into the two piston groupings, 'bathtub type' and 'disc bowl-in-piston types', for three speeds and two air fuel ratios at full load. The part load cases for 2.5 B M E P at 33.3 rps are also shown in two groupings for clarity. The smaller data set for 3.5 B M E P at 50.0 rps for three pistons is shown in Figure 5.67 and 5.68. Figure 5.50 shows typical fired and motored pressure traces for the single slot piston at 33.3 rps and W O T stoichiometric operation. The pressure trace is given over two complete revolutions from B D C of the expansion stroke and through the exhaust, intake, compression and expansion strokes. Compres-sion T D C is taken as the reference of zero degrees. The motored pressure shows the peak within 2 degrees of T D C , confirming proper phasing ot the signal. The fired peak pressure is approximately 17 crank angle degrees after T D C for this condition. • A more quantitative comparison of the effect of combustion chamber con-figuration on the pressure developed is shown by the characteristic parameters of I M E P , peak pressure and angle of occurrence of peak pressure. The full load cases are given i n Tables 5.10 to 5.12 and the part load cases in Tables 5.13 and5.12 Chapter 5. Experimental Results and Discussion Full load operation In this section the trends are first highlighted in a comparison of the pressure traces and then discussed in relation to the cycle statistics. The two sets of Figures 5.51 to 5.56 and 5.57 to 5.62 for full load operation exhibit the same trends in peak pressure for the bathtub and disc groups. The chambers with multijets, ie., squish jet and castellated pistons, show a lower peak pressure than the equivalent non-jet pistons, at al l conditions. Figures 5.53 and 5.54 show that at the mid speed the single slot developed a comparable peak pressure to the standard bathtub chamber. However at the higher speed, Figures 5.55 and 5.56 show the bathtub developed the highest peak pressure. Figures 5.59 and 5.60 show that the standard disc and bowl-in-piston pistons were similar at the m i d speed. This was contrary to expectations for this non squish case but may be explained by the higher compression ratio. No measurements were taken at 50.0 rps, however a set of data taken at 66.7 rps showed very high pressures and intermittent knock at stoichiometric operation. Part load operation The fired pressure traces for the two modified bathtub pistons are compared at fixed part load operating conditions 2.5 B M E P at 33.3 rps in Figures 5.63 and 5.64. The standard bathtub chamber proved unstable for 'fast' data acquisition at this condition. However general performance parameters were taken by the 'slow' acquisition technique as shown in Table 5.3. Again the single slot piston gave a higher peak pressure than the multi jet, castellated Chapter 5. Experimental Results and Discussion piston at the both stoichiometric and lean operation. This trend is also displayed in the disc group, shown in Figure 5.65 and 5.66 where the plain squish, bowl-in-piston chamber exhibits the highest peak pressure. The non squish disc piston had the lowest peak pressure and broadest trace at lean operation. Comparing the chambers at stoichiometric operation, the standard disc, single slot and squish jet chambers produced similar traces. A marked dif-ference however is seen in the bowl-in-piston chamber with the highest peak pressure and the castellated piston wi th the lowest. The bowl-in-piston and squish jet chambers also gave a more peaked pressure trace suggesting faster burn rates. Part load operation at 3.5 B M E P at 50.0 rps shows the same ordering for stoichiometric and lean operation. A t both conditions the bowl-in-piston shows the highest peak pressure with the castellated piston exhibiting a smaller difference compared to the single slot piston at lean running. Full load I M E P The full load I M E P statistics for the various chambers given in Tables 5.10 to 5.12 exhibit pronounced trends in the mean I M E P and lesser trends in the coefficient of variance data. Comparing the bathtub chambers only, the single slot chamber consistently gave the highest I M E P and the castellated piston the lowest. The differences in I M E P ranged from 2-5% at stoichiometric operation to 6-8% at lean operation over the speed range examined. These tables show a less obvious trend i n the coefficient of variance ( C O V ) of I M E P . A t lean operation the single slot piston displayed the least variability wi th the castellated piston having a lower C O V than the bathtub at the Chapter 5. Experimental Results and Discussion lowest speed, 20.0 rps, only. A t stoichiometric operation the single slot piston displayed the least variation at the higher engine speed of 50.0 rps only. Over the speed range tested the differences in C O V between the chambers ranged from 7-30% at stoichiometric and 6-50% at lean operation. Similarly comparing the I M E P for the disc group, the disc piston had the highest mean I M E P and the squish jet the lowest. The C O V information produced a mixed result. The disc chamber gave the lowest C O V for stoi-chiometric operation and the highest for lean operation at 20.0 rps. A t 33.3 rps the disc piston again displayed the least variation at stoichiometric wtih comparable values at lean operation. The tables generally indicate the highest C O V for the squish jet piston at lean operation and high speed with a 50% difference between the six pistons. Part load I M E P Table 5.11 compares the pistons at the slower part load condition and shows an increased difference in the C O V for the castellated and disc pistons from the single slot and bowl-in-piston types. The modified bathtub chambers show increased C O V ' s at part load wi th the differences between the pistons being less marked at lean operation. I M E P was again highest for the single slot chamber. The largest difference is shown in the disc chamber C O V at stoichiometric operation where it is an order of magnitude greater than the bowl-in-piston types. Comparing the I M E P values for al l the chambers at all conditions shows that the bathtub type chambers generally developed higher I M E P s than the disc or bowl-in-piston type chambers. Chapter 5. Experimental Results and Discussion Angle of occurrence of peak pressure The angle of occurrence of peak pressure is also given in Tables 5.10 to 5.14. For al l conditions of speed, load and relative air fuel ratio, the maximum spread of peak position for al l chambers was eight crank angle degrees. The latest peaks occurred approximately 20 degrees A T D C . Barton et al [36] used the angle of occurrence as an indication of flame speed. His studies however, were based on the same ignition advance for all tests. This does not hold for the different values of ignition advance required for M B T t iming, and hence the position of the peak occurrence based on T D C is an indirect indication of burn rate only. The coefficient of variance of peak position provides a measure of the phasing variation. This variance is dependent on the reference position, eg., measuring the peak pressure angle of occurrence from T D C the C O V ranges from 15% to 70%. Hence the significance of the C O V is debatable. The stan-dard deviation of peak position however, does give an indication of phasing variation. These tables show that trends in the peak position are not consis-tent wi th I M E P or performance trends. This last result confirms Matekunas' [45] view that the position of peak pressure occurrence is less useful than I M E P as a measure of engine drivability. The remainder of this section discusses the general trends found in the pistons with and without jet action in relation to the findings from previous studies. Chapter 5. Experimental Results and Discussion Previous studies The full and part load results for the peak pressure in this study indicate higher values for the plain squish or single slot piston chambers. These results are in line with comparative squish and squish jet results of Dymala-Dolesky [18]. A t all operating speeds and conditions Dymala-Dolesky found that the chambers without slots had the highest peak pressures. However, the same consistency was not evident for the I M E P values. Dymala-Dolesky concluded that the best I M E P was obtained for the piston with eight holes directed upwards, (the piston termed 'squish jet' in this study), and that the bowl-in-piston had the lowest I M E P . The current results show the bowl-in-piston to have a similar or higher I M E P than the squish jet. In examining the effect of compression ratio, Dymala-Dolesky tested an alternative bowl-in-piston chamber with a marginally higher compression ra-tio. He obtained I M E P values with this chamber similar to the squish jet chamber wi th eight channels. However, the effect of minor changes in the chamber volume noted by Dymala-Dolesky, has not been confirmed in the combustion geometry or compression checks performed in this work. Other factors such as excessive leakage past the piston rings or wearing in of the pis-ton may account for the original low values of I M E P measured in the previous study. 5.3.3 Mass Fraction Burned Results The results of characterizing the combustion process through a mass fraction burn analysis are discussed in this section. The mass fraction burned data calculated on a crank angle basis are displayed in three ways. Figures 5.69 Chapter 5. Experimental Results and Discussion to 5.84 show the cumulative mass fraction burned against crank angle posi-t ion. Tables 5.15 to 5.23 list the durations of the in i t i a l and main combustion periods determined from arbitrary percentage l imits , and measured in terms of crank angle degrees. Lastly Figures 5.89 to5.98 present bar graph pre-sentations of mass fraction burned ratios showing the in i t i a l and main burn durations for each chamber configuration relative to the slowest or standard bathtub chamber. The general observed trends are discussed in terms of the mass fraction burned plots while quantitative comparisons are made from the statistical results of the duration calculations. Mass fraction burned traces Figure 5.69 shows the cumulative mass fraction burned per crank angle for 60 degrees B T D C to 60 degrees A T D C . The simple mass burn analysis, described i n the previous chapter, assumes the start of combustion at spark and the end of combustion where the pressure rise due to combustion is zero. The mass fraction burned is therefore scaled between 0 and 1 from start to end of combustion. For the bathtub slot chamber at full load 33 rps and R A F R of 1.27, the figure shows the start of combustion at spark 38 degrees B T D C , wi th a slow in i t i a l burn period followed by a rapid increase in mass fraction burned and a final leveling off at approximately 45 degrees A T D C . The 1%, 5% and 90% mass fraction burned percentages are also shown. The in i t ia l burn period is taken from spark to 1% or 5% and the corresponding main burn period from l % - 9 0 % or 5%-90%. Chapter 5. Experimental Results and Discussion Full load operation Figures 5.70 and 5.71 compare all six piston chambers for stoichiometric and lean conditions at W O T 33.3 rps. Whi l e gross differences i n the slope are visible, eg. the single slot piston compared to the bowl-in-piston chamber, the different values of ignition advance make direct comparisons unclear. The distinct difference between the bathtub type pistons and the bowl-in-piston type pistons supports the division of the results into these two groups. These figures also indicate a greater variance between the combustion chambers at lean operation. The effect of engine speed on the burn rate is shown in Figure 5.72 for the bathtub chamber at W O T and lean operation. This figure shows the increase i n ignition advance required as the speed was increased. The figure also illustrates that the mass fraction burned curve for a faster burn at higher speed does not necessarily result in a steeper slope when plotted against crank angle (ca). For example the main combustion duration at 50.0 rps for this chamber was 40.3 ca or 2.2 ms, while the similar crank angle period 41.8 ca at 20.0 rps is 5.8 ms. The raw pressure data for the 40.0 and 50.0 rps standard bathtub cases were provided by A . K a p i l [77] of the A F L Group from measurements taken over the same period as this study. These data were collected for 100 cycles. Figures 5.73 to 5.78 for the bathtub type chambers show a slight increase i n the rate of burn for the single and castellated pistons over the bathtub chamber. This improvement was most pronounced for lean operation at 20.0 and 33.3 rps. Figure 5.77 for stoichiometric operation at 50.0 rps shows similar slopes for al l three chambers, the spacing of the plots resulting from Chapter 5. Experimental Results and Discussion 83 the bathtub chamber being advanced a further six degrees. In all cases the single slot chamber shows a marginal improvement over the castellated piston. The corresponding Figures 5.83 to 5.84 for the disc group show the pro-nounced effect of the plain squish motion. In all cases the bowl-in-piston and squish jet chambers produced an increase in the burning rate shown by the reduced ignition advance and steeper slopes. Similar to the bathtub types this affect was emphasised at lean operation. These figures also indicate that the bowl-in-piston without the squish jets produces a faster main combustion. Part load operation Five of the chambers are compared at the fixed 2.5 B M E P 33.3 rps part load condition in Figures 5.85 and 5.86. The bathtub chamber as mentioned was unstable at this condition. Again these figures show that the bowl-in-piston type chambers have a shorter burn duration than the bathtub chambers with the bowl-in-piston having the shortest. This comparison also shows that the disc chamber and the single slot chamber were similar, wi th the castellated piston producing the slowest burn. Similar trends are seen in Figures 5.87 and 5.88 for the 50.0 rps part load case. The remainder of this section covers the selection of the in i t ia l burn period l imits and a discussion of the duration results. Combustion duration The in i t i a l and main burn durations were determined for two ranges, using 1% and 5% mass fraction burned as the cut off for the in i t ia l burn period. A comparison of the two arbitrary l imits show the same general trends in the ef-fect of the combustion chamber configuration on the combustion. Differences Chapter 5. Experimental Results and Discussion in the in i t ia l burn period for the different piston shapes are more pronounced when 1% is used, particularly for the single slot and castellated pistons. For example i n Tables 5.17 and 5.18 comparing the in i t i a l burn periods for the bathtub chambers, at 33.3 rps full load, shows a 32% range between chambers when 1% ini t ia l burn was used and a 22% range when 5% was used. In separating the in i t i a l flame init iat ion and kernel development from the main burn period, the attempt is made to find a l ink between the type and t iming of the turbulent motion generated, and the combustion process. This is discussed further in the evaluation of the measuring technique and uncertainties. Mass fraction burned ratio Calculation of a representative mass fraction burned ratio has been performed on the 5% in i t i a l burn and subsequent main burn period 5%-90% only. Where possible the different configurations are referenced to the standard bathtub chamber. Where this chamber was unavailable the 'slowest' chamber is used. The similarity of the standard bathtub and disc chamber durations allows a comparison on a two group basis as before. Figures 5.89 and 5.90 for full load operation at 20.0 rps show that the modified bathtub pistons marginally decreased the in i t ia l and main combus-tion durations at stoichiometric operation wi th a more pronounced effect on the main burn period at lean operation. These figures also show that while the bowl-in-piston and squish jet chambers have a dramatic effect on both the in i t i a l and main combustion periods, the effect on the in i t ia l period de-creased wi th lean operation. A t lean operation the disc chamber also showed Chapter 5. Experimental Results and Discussion a decreased main burn period. A t the higher speed of 33.3 rps shown in Fig-ure 5.91 and 5.92 the same trend in reduction of the main burn period by squish action is displayed. The reduced ini t ia l burn period observed with the castellated piston was not repeated at lean operation. The 50.0 rps full load case given in Figures 5.93 and 5.94 show some reversal of the effect of the castellated piston on the main burn, that is, a negligible effect at both stoichiometric and lean operation. In all cases the bowl-in-piston produced the shortest in i t ia l and main burn periods, and the single slot chamber generally gave shorter durations than the castellated piston. Final ly , the part load results presented in Figures 5.95 to 5.97 show the greater effect of plain squish or a single jet over the mult i squish jet action in reducing the main combustion period. Whi l e all the squish pistons reduced the main burn period compared to the non-squish disc chamber, the single slot piston also produced a decrease in the in i t ia l burn period at lean operation. In comparing the different combustion chambers tested in this study with each other and wi th other designs, the chamber geometry is only one of the variables. The location of the 6 p a r k plug i n reducing the maximum distance traveled by the flame is also an important variable. The combustion con-figurations used in this study may be roughly equated to those used in a comparative study by Heywood [61], for example, comparing the bathtub chamber to the 'hemi' chamber with side spark plug location, the single slot and castellated pistons to the 'open' chamber with side spark and the disc chamber and bowl-in-piston chamber as given. W i t h i n these l imits these re-sults are consistent wi th Heywood's study, which showed that the geometry had the greatest impact on the main burn period. The central location of Chapter 5. Experimental Results and Discussion the spark was also shown to be a major factor in reducing the combustion duration supported by the more centrally located spark in the bowl-in-piston types showing an improvement of over 50% compared to the single slot side spark maximum improvement of 15%. The following section provides a discussion of the uncertainties contained i n the above results and an evaluation of the experimental technique. 5.4 Experimental Uncertainties and Technique The uncertainty associated with the results of this investigation may be di-vided into two areas. Errors involved in the equipment sensitivity, accuracy and repeatability are first discussed. Uncertainties inherent in the experi-mental analysis techniques are then discussed under the same headings as the experimental work. 5.4.1 Flow Measurement Pressure measurement Both the pressure transducer and the amplifier were repeatedly calibrated and had a sensitivity error less than 1%. Digital ization error of the pressure signal is also less than 1%. The largest source of error lies in the choice of a reference pressure, ie, the inlet manifold pressure assigned at B D C . This however is a scaling error and of l i t t le concern in a comparative analysis. The pressure measurement is assigned an error of 2%. Chapter 5. Experimental Results and Discussion Hot wire measurement The estimate of the error associated wi th the flow field measurements is 30%. This is a combination of the uncertainty in the hotwire measurements and in the assumptions of isotropic and homogeneous turbulence used in the separa-tion of a fluctuating velocity term from a mean in the absence of a definitive mean flow. The complex flow environment of the engine cylinder, the lack of directional sensitivity of the probe, the high temperature gradients between the fluid and the cylinder walls and the scaling of the calibration constants using an analytical model, all contribute to the uncertainty of the flow mea-surements. Sensitivity analysis on the effect of temperature on the hot wire calibration by Dohring [75] estimated the error associated with this analyti-cal model at 23%. The attempt was made to minimize the thermal gradients present, by maintaining the cylinder coolant at a maximum of 80°C and plac-ing the probe at the furthest distance from the surface while maintaining near spark position. The probe was approximately 5 m m from the cylinder head. Recent studies by Heywood [14] show the thickness of the thermal boundary . layer at T D C of the order of 2mm. The effects of the high error in the flow measurements are somewhat re-duced by their systematic nature in this comparative study. A final comment on the hotwire determination of the flow field deals with the restriction of a single point measurement in representing the cylinder flow. For example, the placement of the probe at 5 m m below the cylinder head suggests that the low measurements for the bowl-in-piston chamber are due to the probe being below the main squish action, not a reduction of turbulence wi th this chamber. This explanation is supported by the fired performance Chapter 5. Experimental Results and Discussion results of the bowl-in-piston chamber and previous studies. 5.4.2 Performance Measurements Engine parameters Performance measurements obtained by the slow data acquisition system were dependent on analogue signals from a number of different instruments. As described in the Chapter 3, extensive checks were carried out after each engine rebuild prior to operation of the engine t iming, throttle and torque positions. The conversion of the analogue signals by the data translation system was also periodically checked and found to be within 3% error. This was consistent wi th similar measurements by Dymala-Dolesky three years earlier. The largest error in the performance parameters came from the dynamic instabil ity of the torque system, resulting in a 5% error. The collected samples of 400 data were taken after stable conditions had been maintained for a m i n i m u m of ten minutes. O n misfire or knock occurrence the data collection was aborted. Variations i n the set operating conditions for the different combustion configurations over the extended two year period of this investigation were less than 1% for speed, 3% in air fuel ratio at lean operation, and for part load operation 12% in B M E P . Fired pressure analysis Similar to the motored pressure the error in the pressure data is set to 2%. Volume assignment for I M E P was assessed using the same clearance volume for al l chambers except the disc chamber. The error in the clearance volume Chapter 5. Experimental Results and Discussion 89 and subsequent compression ratio is 3%-5%. The selection of the sample size has been discussed in the method section. Statistical sampling errors were generally lower for the large sample sizes. No strong memory element was observed in the I M E P , in line with Belmont et al[54] who showed that there was l itt le memory effect in the presence of high cyclic variability and lean operation. Figure 5.99 shows the I M E P variation over 200 consecutive cycles for the single slot chamber at W O T , M B T and lean operation. In this case, and in al l fired pressure measurements the cyclic variability discussed is for stable oper-ation, that is without misfire cycles. Similarly Figure 5.100 shows the I M E P variation over 44 non consecutive cycles for the bathtub chamber at the same operating conditions. Referring to the values in Table 5.11, the mean and standard deviation (std dev) are also displayed on the plot. In this case there is a slight increase in the standard deviation for the smaller sample. The error associated wi th the sample size for the single slot is: mean ± 0.9, stan-dard deviation ± 0.64 and coefficient of variance ± 0 . 1 , and for the bathtub: mean ± 2.5, standard deviation ± 1.8 and coefficient of variance ± .14. Mass fraction burned The main approximations made in the mass fraction burned calculation were the calculation of the polytropic oefficient and determination of the end of combustion, Rassweiler and Withrow [49] found that a variation in the poly-tropic coefficient between 1.25 and 1.35 had l itt le effect. A sensitivity study on the coefficient of compression (7C), comparing an ensembled averaged value and individual cycle values, produced a smooth representative trace from Chapter 5. Experimental Results and Discussion individual ly calculated coefficients only. The compression coefficient was cal-culated from an average of values before spark. The use of Rassweiler and Withrow's simple method was supported by Amann's [19] comparison with more complex computer models. In selecting 1% and 5% mass fraction burned for this study, the ini t ia l burn period was also calculated for 2%, 3% and 10%. A recent sensitivity study by K a p i l [77] showed that below 2% mass burn there is a sharp increase in standard deviation of the in i t ia l burn period associated wi th experimental error and measuring techniques. He subsequently used extrapolated values from 5% to get a theoretical standard deviation at zero burn time. However the upturn in the cyclic variability need not necessarily be assigned to ex-perimental error as it appears repeatable and may be a phenomena of cyclic variation in mass fraction burned. Young [4] and Yatsamoto [46] both used 1% as the ini t ia l burn period to separate the 'ignition delay' (a somewhat misleading term), from the main burn period. Other researchers have used 5% and 10% [17, 61]. Statistical sample size error Fina l ly in comparing the data taken over 44, 100 and 200 cycles a statistical check was used to acertain whether there was a significant difference between values. Testing the 'no change'or nul l hypothesis (Ho) at a 5% level of significance showed that a difference in the means of I M E P between two samples less than 5-10 bar caused 'no rejection' of the null hypothesis. Similarly over al l sample sizes and conditions a difference less than 50-100 bar in the peak pressure caused 'no rejection'. A test at the 5% level of significance for a 'no change' Chapter 5. Experimental Results and Discussion 91 hypothesis in the standard deviation showed that a difference less than 30% between two 44 cycle samples, less than 18% for one 44 and one 200 cycle sample, or less than 5% for two 200 cycle samples, caused 'no rejection' of the nul l hypothesis. The effect of the number of cycles on the sample error of all the statistical calculations, ie., I M E P , peak pressure and burn durations is indicated in the number of significant figures given in the appropriate tables. The sample error was based on the 'large sample' approximation for maximum likelyhood esti-mates of mean, standard deviation and coefficient of variance. The equations for these calculations were taken from Bury's statistical text [83]. The statistical analysis was used as a general check on the data obtained and does not constitute a detailed analytical investigation. 5.5 Turbulence, Combustion and Performance The remainder of this section examines the relation between the flow field measurements and the performance measurements obtained in this study. The general findings for the bathtub group of pistons illustrate that the addition of squish jet action to the standard bathtub chamber enhances the mean velocity and turbulent intensity of the in-cylinder flow. The subsequent improvement in performance was confirmed by increased brake thermal effi-ciency and indicated mean effective pressure ( I M E P ) . The results show that the effect of the directed squish motion was stronger when concentrated in one jet,that is, i n the single slot piston. Examinat ion of the M B T ignition advance for the bathtub group confirms that the squish effect is present in the main burn period. Squish action Chapter 5. Experimental Results and Discussion occurred after 20 degrees B T D C , wi th the ignition advance for the bathtub pistons generally i n the region 30 degrees B T D C . The effect of the large squish area of the bowl-in-piston type chambers in reducing in both the in i t ia l and main burn periods was illustrated i n the ignition t iming of 15-20 degrees B T D C for these chambers. Two significant differences in the turbulence and performance results for the bowl-in-piston chambers are apparent. First , the low mean velocity and turbulent intensity measurements for the plain bowl-in-piston were not re-flected in low performance data. Second, the high flow field measurements for the squish jet piston did not result in high performance. However, a comparison wi th the results of previous studies and an examination of the velocity measurement location suggests the first anomaly is a deficit of the measurement recorded not of the physical phenomenon. The turbulence and performance measurements for the squish jet piston, however, were consistent with previous studies. Possible reasons for the ob-served reduction i n performance may be related to the effect of the high level of turbulence and increased surface area. As noted previously, excess turbulence can cause the disruption of the early flame kernel by convection away from the spark point towards the chamber walls causing quenching. The increased surface area of the chamber due to the channel passages also contributes to heat loss and quenching. The potential for improved part load performance through strong squish action was clearly shown for the single slot and bowl-in-piston pistons. These pistons gave a 2%-3% increase in thermal efficiency over the standard bathtub and squish chambers. The coefficient of variance of I M E P for these concen-trated squish action pistons was a th ird to a half lower than the multijet Chapter 5. Experimental Results and Discussion designs, which is an indication of improved drivability. The specific focus of the experimental work in this project has been to ex-pand the information available on squish and squish enhanced turbulence in affecting and promoting fast-lean operation using alternative fuels. Through-out the study, conclusions drawn from this investigation have been reviewed under a range of engine configurations and operating conditions. The conclu-sions drawn from this investigation and the directions identified for further research , are presented in the final chapter. Chapter 6 Conclusions and Recommendations 6.1 Introduction The objective of this research was to investigate the influence of combustion chamber design on turbulence enhancement i n the achievement of fast-lean operation of a spark ignition engine. The specific focus of the experimental work was to examine the effect of squish induced jet action. A modified flat piston was designed for use with a bathtub cylinder head where the jet action was formed by a single slot or slots in a raised ridge on the piston surface. In the further investigation of bowl-in-piston chambers these jets were produced by adding channels from the top surface to the side of the bowl inclined towards the spark position. The evaluation of six combustion chamber configurations was conducted i n four stages: 1. F low field measurements by hotwire at the spark location were taken and subsequently analysed using a window average technique. Mean velocities and turbulent intensities were obtained by this method. 2. Performance measurements for the engine fueled by natural gas over the speed range 20-50 rps for stoichiometric and lean operation at full load and part load were made. 94 Chapter 6. Conclusions and Recommendations 3. Combustion pressure histories were measured over the same range. Char-acterizing parameters of Imep, peak pressure and angle of occurrence of peak pressure were also used to evaluate the engine performance and as a measure of the cyclic variability with each chamber. 4. Mass fraction burned curves were calculated using a simple rate of pres-sure rise model to evaluate the effect of chamber design on the in i t i a l and main burn durations. These durations for two phases were defined as 0-1% and 1-90% and 0-5% and 5-90% mass fraction burned. 6.2 Conclusions The conclusions drawn from this investigation may be summarized under the principal areas of study: 6.2.1 Turbulence Studies • The mean velocity and turbulence intensity at T D C increases with in-creased engine speed. • The addition of a ridge wi th one or more slots to the standard bathtub piston reduces the mean velocity and turbulent intensity before T D C . • The single slot or castellated wall produces or maintains a higher level of turbulence after top dead centre, than the flat piston. • The increased squish area of the bowl-in-piston and squish jet pistons produces a higher mean velocity and turbulent intensity than the smaller squish area of the bathtub type chambers. The bowl-in-piston measure-ments were confirmed by previous studies [17, 18]. Chapter 6. Conclusions and Recommendations 96 • The addition of slots to the single slot ridge reduces the effect of the main squish and jet action. Similar addition of channels to the bowl-in-piston chamber reduces the squish effect. • The squish jet action of forcing the flow into the lower portion of the bowl as observed by Dymala-Dolesky [18] was confirmed by these mea-surements. 6.2.2 Performance and Combustion Studies • Enhanced turbulence by combustion chamber geometry is most effective in improving performance at lean operation. • The larger squish area and more centrally located spark configurations produce the greatest reduction in the in i t i a l and main combustion du-ration. • Squish is most effective i n the main burn period, occurring approxi-mately 20 degrees before T D C , with ignition t iming for the chambers tested generally around 30 degrees before T D C . • The single jet action of the single slot piston, directed towards the spark is most effective in improving the efficiency at high speed and lean mix-tures. • High turbulence intensity is not necessarily beneficial to engine perfor-mance when accompanied by increased heat loss and cyclic variability. • The reduced ignition advance requirements for the modified bathtub pistons indicate their design potential to extend the knock l imi t . • The addition of the squish jet action has the greatest potential for im-proving engine drivability at part load operation by reduction of the Chapter 6. Conclusions and Recommendations cyclic variability. In stating these conclusions, it is important to note that few trends were consistent over all speed and load conditions for the range of specific engine operating conditions tested in this investigation. Conclusions drawn from some previous studies have been l imited to specific operating conditions. The following section contains recommendations on expanding the basis of com-parison and test conditions following the range of engine and operating vari-ables included in this investigation. 6.3 Recommendations The trends identified from this experimental investigation provide direction for further work on the analysis of the flow field developed and the interaction of the jet motion in combustion chamber design. The l imits of single point H W A measurements in an engine have been well documented and confirmed i n this study. It is suggested that further studies of the 'jet effect' should be conducted in a combustion bomb or rapid compression machine, ideally with laser Doppler anemometry at mul t i positions. The promising improvements of the single slot piston over the standard bathtub configuration at high speed and lean operation suggest further that studies should examine leaner mixtures and higher speeds within the l imits of the available apparatus. The bathtub piston modification was selected as the most practical modification for this investigation, because it required changes to the piston only. This however restricted the squish area available. Further experiments using a modified bowl-in-piston chamber and the required step i n the cylinder head should be carried out to achieve the advantages of both Chapter 6. Conclusions and Recommendations the large squish area action and jet slot action. F ina l ly it is recommended that exhaust emission studies be carried out to ascertain whether the advantages of these fast burn squish jet pistons are mitigated by the increased surface areas, and possible quenching effects of the channels leading to unburned deposits. Bibliography [1] K a r i m , G . A . , " A n Examinat ion of The Nature of the Random Cycl ic Pressure Variations in a Spark-Ignition Engine" , J . of the Institute of Petroleum, V o l . 53, No. 519, March 1967. [2] Young, M . B . , "Cyc l i c Dispersion in the Homogeneous Charge Spark Ig-nit ion Engine — A Literature Study" , S A E Paper 810020, 1981. [3] Andrews, G . E . , Bradley, D . , and Lwakabamba, S.B., "Turbulence and Turbulent Flame Propagation — A Cr i t i ca l Appra i sa l " , Combustion and Flame V o l 24, pp 285-304 1975. [4] Young, M . B . , "Cyc l i c Dispersion - Some Quantitative Cause-and-Effect Relationships", S A E Paper 800459, 1980. [5] Damkohler, G . , "The Effects of Turbulence on the Flame Velocities in Gas Mix ture s " , N A C A T M 1112, 1947. [6] Shchelkin, K . L , " O n Combustion in a Turbulent F l o w " , N A C A T M 1110, 1947. [7] Tabaczynski, R . J . , and Ferguson, O R . , a n d Radhakrishnan,K. , " A Tur-bulent Entrainment Model for Spark- Ignition Engine Combust ion" , S A E Paper 770647, 1977. [8] B l i zard ,N.C . , and Keck, J .C . , "Theoret ica l and Experimental Investiga-tion of a Burning Model for Spark- Ignition Engines", S A E Paper 740191, 1974. [9] Tabaczynski, R . J . , 'Turbulence and Turbulent Combustion in Spark Ig-nit ion Engines" ,Prog. Energy Combust. Sci. V o l 2 pp 143-165, 1976. [10] Semenov,E.S., "Studies of Turbulent Gas Flow in Piston Engines," T E C H . T R A N S . F97 N A S A 1963. [11] Fraser, R . A . , and Bracco, F . V . , "Cycle- Resolved L D V Integral Length Scale Measurements in an IC Engine" , S A E Paper 880381, 1988. [12] Hinze, P.O.,Turbulence, 2nd E d M c G r a w - H i l l 1975 [13] Lancaster, D . R . , "Effects of Engine Variables on Turbulence i n a Spark Ignition-Engine", S A E Paper 760159, 1976. [14] Heywood, J .B . , " F l u i d Mot ion within the Cylinder of Internal Combus-tion Engines - The 1986 Freeman Scholar Lecture" , J . of Fluids Engi-neering V o l 109/3, March 1987. [15] Taylor, G.I . , "Statistical Theory of Turbulence," Book Extract 1935. [16] Andrews, G . E . , and Bradley, D . , "The Burning Velocity Of Methane-Air Mix ture s " , Combustion and Flame V o l 19 pp 275-288, 1972. [17] Cameron, C . , " A n Investigation of Squish Generated Turbulence in IC engines", M . A . S c . Thesis U B C , AF1-85-02, 1985 99 Bibliography 100 [18] Dymala-Dolesky, R., "The Effects of Turbulence Enhancement on the Performance of a Spark- Ignition Engine" , M . A . S c . Thesis U B C . 1986. [19] A m a n n , C . A . , "Cyl inder Pressure Measurement and its use i n Engine research", S A E Paper 852067, 1985. [20] Winsor , R . E . , and Patterson, D . J . , " M i x t u r e Turbulence- A Key to Cycl ic Combustion Var ia t ion" , S A E Paper 730086, 1973. [21] W i t z e , P .O . , "Measurements of Spatial Distr ibution and Engine Speed Dependence of Turbulent A i r Mot ion i n an IC Engine" , S A E Paper 770220,1977. [22] Bopp, S., Vafidis, O , and White law, J . H . , "The Effect of Engine Speed O n The T D C Flow field In A Motored Reciprocating Engine" , S A E Paper 860023, 1986. [23] Daneshyar, H . , and Fuller, D . E . , "Definition and Measurement of Tur-bulence Parameters in Reciprocating IC Engines", S A E Paper 861529, 1986. [24] H a l l , M . J . , and Bracco, F . V . , " A Study of Velocities and Turbulence In-tensities Measured in F i r ing and Motored Engines", S A E Paper 870453, 1987. [25] Ikegami, M . , Siaj i M . , , and Nishimoto, K . , "Turbulence Intensity and Spatial Integral Scale During Compression and Expansion Strokes in a Four-Cycle Reciprocating Engine" , S A E Paper 870372, 1987. [26] Dent, J .C . , and Salama, N.S. , "The Measurement O f the Turbulence Characteristics i n an Internal Combustion Engine Cyl inder " , S A E Paper 750886, 1975. [27] Haghgooie, M . , Kent , J .C . , and Tabaczynski, R . J . , "Turbulent Time-Scale Measurements i n a Spark Ignition Engine Using Hot wire Anemom-etry and fast Response Ion Probes", Symposium on Flows in IC Engines A S M E W A M 1982. [28] Rask, R . B . , "Laser Doppler Anemometry Measurement in an Internal Combustion Engine," S A E Paper 790094, 1979. [29] Fraser, R . A . , Felton, P . G . , Bracco, F . V . , and Santavicca, D . A , "Pre l im-inary Turbulence Length Scale Measurements i n a Motored IC Engine" , S A E Paper 860021, 1986. [30] W i t z e , P . O . , " A Cr i t i ca l Comparision of Hot- W i r e Anemometry and Laser Doppler Velocimetry for IC Engine Appl icat ions" , S A E Paper 800132,1980. [31] Matsuoka, S., Yamaguchi, T . , and Umemura, V . , "Factors Influencing the Cyc l ic Variations of Combustion of Spark-Ignition Engine" , S A E Paper 710586, 1971. [32] Gatowski , G . A . , Heywood, J .B . , and Deleplace, O , "Flame Photographs i n a Spark-Ignition Engine" , Combustion and Flame V o l 56 pp 71-81, 1984. [33] Keck, J .C . , Heywood, J .B . , and Noske, G . , "Ear ly Flame Development and Burning Rates in Spark Ignition Engines and Their Cyc l ic Variabil-i t y " , S A E Paper 870164, 1987. Bibliography W i t z e , P . O . , et al. "Measurements and Predictions of the P r e -Combustion F l u i d Mot ion and Combustion Rates in a Spark Ignition Engine" , S A E Paper 831697, 1983. M a r t i n , J . K . , Wi tze , P .O . , and Borgnakke, C , "Combustion Effects on the Preflame Flow Fields in a Research Engine" , S A E Paper 850122, 1985. Barton, R . K . , Kenemuth, et al . "Cycle-by-Cycle Variations of A SI En-gine — A Statistical Analys i s " , S A E Paper 700488, 1970. Gosman, A . D . , "Mult idimensional Model l ing of Co ld Flows and Turbu-lence in Reciprocating Engines", S A E Paper 850344, 1985. Soltau, J .P . , "Cyl inder Pressure Variations in Petrol Engines", Proceed-ings of the Institution of Mechanical Engineers No. 2, 1960-1961. Patterson, D . J . , "Cyl inder Pressure Variations, A Fundamental Com-bustion Prob lem" , S A E Paper 660129, 1966. Hansel, J . G . , " A Turbulent Combustion Model of Cycle-to-Cycle Com-bustion Variations in Spark-Ignition Engines," Combustion Science and Technology, vol 2 pp 223- 225, 1970. Hancock, M . S . , Buckingham, D . J . , and Belmont, M . R . , "The Influence Of A r c Parameters on Combustion in a Spark-Ignition Engine" , S A E Paper 860321, 1986. Anderson, R . W . , "The Effect Of Ignition System Power on Fast Burn Engine Combust ion" , S A E Paper 870549, 1987. Kalghatgi , G . T . , "Spark Ignition, Ear ly Flame Development and Cycl ic Variat ion In IC Engines", S A E Paper 870163, 1987. Lancaster, D . R . , and Kreiger, R . B . , et al . "Effects of Turbulence on Spark Ignition Engine Combust ion" , S A E Paper 760160, 1976. Matekunas, F . A . , "Modes and Measures of Cyc l ic Var iab i l i ty " , S A E Transaction vol 92 pp 1139, S A E Paper 830337, 1983. Yamamoto, H . , and M i s u m i , M . , "Analysis of Cycl ic Combustion Varia-tion in a Lean Operating SI Engine" , S A E Paper 870547, 1987. Nagayana, I., A r a k i , Y . , and Lioka, Y . , "Effects of Swir l and Squish on SI Engine Combustion and Emmis ion" , S A E Paper 770217, 1977. Kuroda , H . , and Nakaj ima, Y . , "The Fast B u r n wi th Heavy E G R , New Approach for Low N O x and Improved fuel Economy", S A E Paper 780006, 1978. Rassweiler, C M . , and Withrow, L . , " M o t i o n pictures of Engine Flames Correlated with Pressure Cards" , S A E V o l 42 no 5 1938. Gatowski , J . A . , Heywood, J .B . , et al. "Heat Release Analysis of Engine Pressure Data " , S A E Paper 841359, 1984. Kreiger, R . B . , and Borman, G . L . , "The Computation of Apparent Heat Release for Internal Combustion Engines", A S C E 6 6 - W A / D G P - 4 , 1966. Bibliography 102 [52] A m a n n , C . A . , "Combustion In The Spark-Ignition Engine" , Keynote address I M e c h E International conference on combustion in engines," May 1988. [53] Cole, J .B . , and Swords, M . D . , " A n Investigation of the Ignition Process in a Lean-Burning Engine using Conditionally Sampled Laser Doppler Anemometry," S A E Paper 800043, 1980. [54] Belmont, M . R . , Hancock, M . S . , and Buckingham, D . J . , "Statistical As-pects of Cycl ic Var iab i l i ty " , S A E Paper 860324, 1986. [55] Daily, J . W . , "Cycle-to-Cycle Variations: A Chaotic Process?", S A E Pa-per 870165, 1987. [56] M a , T . H . , "Effect of Cyl inder Charge Mot ion on Combust ion" , C81/75 I M e c h E , 1975. [57] W i t z e , P . O . , "The Effect of Spark Location on Combustion in a Variable-Swirl Engine" , S A E Paper 820044, 1982. [58] W i t z e , P . O . , and V i l c h i s , F . R . , "Stroboscobic Laser Shadowgraph study of the Effect of Swirl on Homogeneous Combustion i n a Spark-Ignition Engine" , S A E Paper 810226, 1981 [59] Sheppard, C . W . , and Bradley, "Limitat ions to Turbulence Enhanced Burning Rates in Lean B u r n Engines," I M E C H E C46/88, 1988. [60] Saxena, V . , and Rask R . B . , "Influence of Inlet Flows on the Flow Fie ld in an Engine" , S A E Paper 870369, 1987. [61] Heywood, J .B . , "Combustion Chamber Design for O p t i m u m Spark-Ignition Engine Performance", International Journal of Vehicle Design , vol 5 No.3 pp 133-147, 1984. [62] Gruden, D . O . , "Combustion Chamber Layout for Modern Otto E n -g ines" ,SAE Paper 811231, 1981. [63] Overington, M . T . , and Thr ing , R . H . , "Gasoline Engine Combustion-Turbulence and The Combustion Chamber," Ricardo Consulting E n -gineering L t d . S A E Paper 810017, 1981. [64] Shimoda, M . , et al . "Effect of Combustion Chamber Configuration on In-Cylinder A i r Mot ion and Combustion Characteristics of C I Diesel Engine" , S A E Paper 850070, 1985. [65] Jane, P . A . H . , "The Development of a Direct Injection Diesel Combus-tion System for Low Noise Emmisions and Mechanical Loading" , C66/88 I M e c h E 1988. [66] Evans, R . L . , "Internal Combustion Engine Squish Jet Combustion Chamber" , U S A Patent, No 4,572,123 Feb.25, 1986. [67] Evans, R . L . , and Cameron, O , " A New Combustion Chamber for Fast B u r n Applicat ions," S A E Paper 860319, 1986. [68] Nakamura , H . , Ohinouye, T . , et al . Mitsubishi "Development of a New Combustiono System ( M C A - J E T ) in Gasoline Engine" , S A E Paper 790016, 1979. Bibliography 103 [69] Jones, A . " U B C Ricardo Hydra Engine Test Fac i l i ty" , AFL-86-08, 1986. [70] Ricardo Consultant "The Ricardo/Cussons Standard H y d r a Engine and Test B e d " , 1985. [71] Brown, "Methods for Evaluating Requirements A n d Errors In Cyl inder Pressure Measurement", S A E Paper 670008, 1967. [72] Benson, R.S. , and Pick , R., 'Recent Advances In Internal Combustion Engine Instrumentation W i t h Particular Reference to High-Speed Data Acquis i t ion and Automated Test Bed, , ' S A E Paper 740695, 1974. [73] Vines, R . F . , " T h e P la t inum Metals and their A l loy s " , The International Nickel Company, Inc.,New York, New York, 1941. [74] Boisvert, J . , "Turbulent Combustion of Gas-Air Mixtures In a Spark Ignition Engine," M . A . S c Thesis U . B . C . AFL-86-05 1986. [75] Dohring, K . , "The Relative Effects of Intake and Compression Stroke Generated Turbulence In I .C. Engine Durat ion" , M . A . S c . Thesis U B C AFL-86-01 1986. [76] Lancaster D .R . , Krieger, R . B . and Liemesch, "Measurement and Analy-sis of Engine Pressure Data " , S A E Paper 750026, 1975. [77] K a p i l , A . "Cycle-to-Cycle Variations in Spark Ignition Engines", M . A . S c Thesis, U B C , 1988. [78] Coll i s , D . C . , and Wi l l i ams , M . J . "Two- Dimensional Convection from Heated Wires at Low Reynolds Number" , J . of F l u i d Mechanics V o l 6 pp 357-384, 1959. [79] Davies, P . O . A . L . , and Fisher, M . J . , "Heat Transfer From Electrically Heated Cyl inders" , Proc. Roy. Soc. A . , V o l 280 pp 486-526, 1964. [80] Rask, R . B . , "Comparision of Window Smoothed- Ensembled and Cycle-by-Cycle Data Reduction Techniques for Laser Doppler Anemometry Measurements of In-Cylinder Veloci ty" , A S M E 1981. [81] Fransler, T . D . , "Laser Velocimetry Measurements of Swirl and Squish Flows i n an Engine with a Cyl indr ica l Piston B o w l " , S A E Paper 850124, 1985. [82] Catania, A . E . , and M i t t i c a , A . , " A Contribution To The Definition A n d Measurement Of Turbulence In A reciprocating IC Engine" , 85DPG12, 1985. [83] Bury, K . V . , Statistical Models in Appl ied Science John W i l e y and Sons 1975. Tables 104 E N G I N E : Number of cylinders 1 Bore 80.26 m m Stroke 88.9 m m Swept Volume 0.45 litres M a x i m u m Speed 90 rps M a x i m u m Power 15 k W Compression ratio (nominal) 9 : 1 V A L V E A R R A N G E M E N T Overhead cam shaft verticle lift 9 m m I V O : Inlet Valve Opens 12° B T D C I V C : Inlet Valve Closes 56° A B D C E V O : Exhaust Valve opens 56° B B D C E V C : Exhaust Valve Closes 12° A T D C I G N I T I O N S Y S T E M Type Lumenit ion C o i l Lucas SP 12 T i m i n g range 70° B T D C to 20° A T D C Spark P lug Champion N 6 Y or A G Y C Table 3.1: Ricardo H y d r a Gasoline (or Gaseous fuel) Engine Specifications. Tables 105 P I S T O N o: P R E S S U R E • : H O T W I R E 44 cycles S P E E D 20.0 (rps) 33.3 (rps) 50.0 (rps) 66.7 (rps) std bathtub 0 • 0 • 0 • single slot 0 • 0 • 0 castellated 0 • 0 • o • bowl/piston 0 • 0 • 0 0 squish jet 0 • 0 • 0 • std disc 0 • 0 • 0 • Table 3.2: Motored operating conditions for Pressure and Hotwire measurements at W O T . P I S T O N P R E S S U R E o: 44 cycles • : 200 cycles S P E E D 20.0 (rps) 33.3 (rps) 50.0 (rps) 66.7 (rps) R A F R 1.00 1.27 1.00 1.27 1.00 1.27 1.00 1.27 std bathtub 0 0 0 0 0 0 single slot • • • • • • castellated • • • • • • bowl/piston • • • • • • squish jet 0 0 o o 0 0 std disc 0 0 0 0 o 0 Table 3.3: F i red operating conditions for Pressure measurements at M B T and Ful l Load ( W O T ) . P I S T O N P R E S S U R E o: 44 cycles • : 200 cycles S P E E D 33.3 (rps) 50.0 (rps) B M E P 2.5 (bar) 3.5 (bar) R A F R 1.00 1.27 1.00 1.27 std bathtub single slot • • • • castellated • • • • bowl/piston • • • • squish jet o 0 std disc 0 0 Table 3.4: F i red operating conditions for Pressure measurements at M B T and Part Load. Tables 106 MOTORED PRESSURE HOTWIRE ANALYSIS ANALYSIS HOTWIRE HOTWIRE HW-Cal ISAAC2VAX PRESS- ANAL EXTRACTION 1 SMTH--AVE TEMP--1800 ISAAC2VAX EXTRA CTION HW-ANAL TURBULENCE Table 4.1: Motored data Analysis program flow chart. Tables 107 FIRED PRESSURE ANALYSIS: 200 cycles HOTPRES2 ISAC22VAX FIRERUN MASSBURN-ALL FIRED PRESSURE ANALYSIS: 44 cycles HOTPRES ISAAC2VAX FIRERUN-44 MASSBURN-ALL PERFORMANCE ANALYSIS DATAQ CRUNCH Table 4.2: Fired data Analysis program flow charts. Tables 108 P I S T O N Compression coefRceint jc Expansion coefficient 7 e std bathtub 1.34 1.3.5 single slot 1.33 1.35 castellated 1.32 1.35 bowl/piston 1.30 1.35 squish jet 1.27 1.35 std disc 1.37 1.35 Table 5.1: Compression and Expansion coefficients for the motored condition. Tables 109 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 3.20 7.10 256 28.9 25.4 24 1.02 single slot 3.09 6.87 276 26.9 24.6 24 1.03 castellated 2.70 5.99 298 24.9 21.5 22 1.03 bowl/piston 3.40 7.60 240 30.8 27.1 11 1.02 squish jet 2.75 6.10 294 25.2 22.0 16 1.03 std disc 3.11 6.92 262 28.3 24.8 19 1.04 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 2.60 5.94 248 29.9 21.3 29 1.30 single slot 2.80 6.16 250 29.6 22.1 28 1.30 castellated 2.36 5.27 285 26.0 18.9 31 1.30 bowl/piston 2.96 6.57 226 32.8 23.5 21 1.30 squish jet 2.33 5.20 284 26.1 18.6 24 1.31 std disc 2.77 6.17 243 30.5 22.0 29 1.30 Table 5.2: Engine performance as per SAEJ1349 for different piston geometries for stoi-chiometric and lean R A F R at M B T , W O T , and 20.0 rps. Tables 110 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 5.30 7.12 254 29.2 25.5 35 1.02 single slot 5.18 6.90 272 27.2 24.7 30 1.01 castellated 5.09 6.77 272 27.3 24.2 26 1.01 bowl/piston 5.98 7.94 241 30.7 28.4 12 1.01 squish jet 4.81 6.40 288 25.8 22.9 20 1.01 std disc 5.64 7.50 248 29.9 26.9 32 1.02 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 4.55 6.12 245 30.3 21.9 38 1.28 single slot 4.80 6.50 234 31.6 23.1 31 1.28 castellated 4.47 5.96 251 29.5 21.3 33 1.29 bowl/piston 5.27 7.01 219 33.7 25.1 20 1.28 squish jet 4.17 5.54 272 27.3 19.8 27 1.31 std disc 4.78 6.39 238 31.2 22.9 33 1.28 Table 5.3: Engine performance as per SAEJ1349 for different piston geometries for stoi-chiometric and lean R A F R at M B T , W O T , and 33.3 rps. bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. single slot 7.98 7.08 262 28.4 25.4 33 1.01 castellated 7.49 6.67 279 26.5 23.9 31 1.01 bowl/piston 8.77 7.75 245 30.2 27.8 16 1.02 squish jet 7.05 6.29 299 24.8 22.5 21 1.01 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. single slot 7.40 6.50 230 32.3 23.4 37 1.27 castellated 6.94 6.17 249 29.8 22.1 37 1.27 bowl/piston 7.82 7.00 222 33.4 24.9 22 1.27 squish jet 6.14 5.47 279 26.6 19.6 28 1.27 Table 5.4: Engine performance as per SAEJ1349 for different piston geometries for stoi-chiometric and lean R A F R at M B T , W O T , and 50.0 rps. Tables 111 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 2.02 2.70 349 21.3 9.7 35 1.05 single slot 1.93 2.56 348 21.3 9.2 31 1.05 castellated 1.79 2.38 371 20.0 8.5 31 1.03 bowl/piston 1.90 2.51 308 24.0 9.0 23 1.04 squish jet 1.90 2.54 339 21.9 9.1 26 1.05 std disc 1.84 2.48 373 19.9 8.9 32 1.06 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. std bathtub 2.20 3.00 355 20.9 10.5 35 1.27 single slot 1.94 2.56 331 22.4 9.2 33 1.31 castellated 1.91 2.55 364 20.3 9.1 34 1.30 bowl/piston 1.92 2.54 291 25.4 9.1 24 1.31 squish jet 1.92 2.54 325 22.8 9.1 29 1.27 std disc 1.75 2.35 350 21.1 8.4 32 1.25 Table 5.5: Engine performance as per SAEJ1349 for different piston geometries for stoi-chiometric and lean R A F R at M B T , 2.5 bmep, and 33.3 rps. bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. single slot 3.88 3.45 314 24.4 12.4 37 1.02 castellated 3.97 3.53 321 23.1 12.6 36 1.02 bowl/piston 3.78 3.36 305 24.3 12.0 24 1.03 bp bmep bsfc Vth Torq. Ign.Adv R A F R P I S T O N k W bar g / k W h r % N m deg. single slot 3.88 3.46 281 26.3 12.4 41 1.28 castellated 4.01 3.56 300 24.7 12.7 40 1.28 bowl/piston 3.92 3.48 277 26.7 12.4 25 1.28 Table 5.6: Engine performance as per SAEJ1349 for different piston geometries for stoi-chiometric and lean R A F R at M B T , 3.5 bmep, and 50.0 rps. Tables 112 R A F R 1.02 1.06 1.12 1.17 1.22 1.27 1 1.32 P I S T O N I G N I T I O N A D V A N C E (degrees btdc) std bathtub 24 26 27 28 29 29 31 single slot 20 24 27 29 29 28 28 castellated 22 25 28 29 29 31 32 bowl/piston 11 14 15 18 21 21 20 squish jet 16 19 21 22 23 24 25 std disc 19 21 26 28 28 29 30 R A F R 1.02 1.06 1.12 1.17 1.22 1.27 1.32 P I S T O N B R A K E T H E R M A L E F F I C I E N C Y ( % ) std bathtub 28.9 29.7 29.6 30.1 29.7 29.9 29.3 single slot 26.9 28.0 28.6 29.2 29.4 29.6 29.6 castellated 24.9 25.4 25.5 26.0 26.0 26.0 26.0 bowl/piston 30.8 31.7 32.1 32.2 32.9 32.8 32.9 squish jet 25.2 25.5 26.2 26.1 25.8 26.1 26.0 std disc 28.3 29.4 30.0 30.4 30.6 30.5 30.5 Table 5.7: Ignition Advance and Brake Thermal for R A R F ^ l . 0 0 - 1 . 3 5 at M B T , W O T , and 20.0 Efficiency for different piston geometries rps. Tables 113 R A F R 1.02 1.06 1.12 1.17 1.22 1.27 1.32 P I S T O N I G N I T I O N A D V A N C E (degrees btdc) std bathtub 35 36 37 36 38 38 40 single slot 30 30 30 31 31 31 31 castellated 26 30 31 31 32 33 34 bowl/piston 12 15 18 19 19 20 21 squish jet 20 24 26 26 27 27 29 std disc 32 36 36 37 33 33 34 R A F R 1.02 1.06 1.12 1.17 1 1.22 1.27 1.32 P I S T O N B R A K E T H E R M A L E F F I C I E N C Y ( % ) std bathtub 29.2 30.0 30.0 30.5 30.4 30.3 30.1 single slot 27.2 28.8 29.6 30.5 31.1 31.6 31.7 castellated 27.3 28.5 29.2 29.6 29.8 29.5 29.3 bowl/piston 30.7 32.3 32.9 33.3 33.4 33.7 33.9 squish jet 25.8 26.7 26.9 27.1 27.2 27.3 27.1 std disc 30.0 30.8 31.5 31.5 31.8 31.2 30.9 Table 5.8: Ignition Advance and Brake Thermal Efficiency for different piston geometries for R A R F « 1 . 0 0 - 1 . 3 5 at M B T , W O T , and 33.3 rps. R A F R 1.02 1.06 1.12 1.17 1.22 1.27 1.32 P I S T O N I G N I T I O N A D V A N C E (degrees btdc) single slot 33 32 33 35 35 37 38 castellated 31 32 34 36 35 37 38 bowl/piston 16 19 21 21 21 22 23 squish jet 21 24 26 26 28 28 31 R A F R 1.02 1.06 1.12 1.17 1 1.22 1 1.27 1.32 P I S T O N B R A K E T H E R M A L E F F I C I E N C Y ( % ) single slot 28.4 29.6 30.7 31.4 32.2 32.3 32.7 castellated 26.5 28.2 29.0 29.7 29.8 29.8 29.7 bowl/piston 30.2 31.8 32.5 33.2 33.2 33.4 33.7 squish jet 24.8 25.5 26.1 26.2 26.5 26.6 26.5 Table 5.9: Ignition Advance and Brake Thermal Efficiency for different piston geometries for R A R F wl.00-1.35 at M B T , W O T , and 50.0 rps. Tables 114 I M E P t 1 tr COV CT COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A FR=1.27 std bathtub 818.7 12.8 1.58 701.5 20.4 2.94 single slot 821.7 14.1 1.72 764.2 13.6 1.79 castellated 808.5 13.0 1.61 726.2 14.1 1.95 bowl/piston 788.1 14.6 1.86 697.1 13.8 1.98 squish jet 768.9 10.8 1.41 673.9 13.3 1.99 std disc 826.9 8.0 0.98 743.8 19.0 2.58 P E A K tT COV tT COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 3957 318 8.1 3237 402 12.6 single slot 4032 316 7.8 3819 344 9.0 castellated 4030 263 6.6 3802 337 8.9 bowl/piston 4418 738 16.7 4361 201 4.6 squish jet 3527 124 3.6 3169 168 5.3 std disc 3918 388 10.0 3761 441 11.9 I P E A K tT COV tT COV c.a. c.a. % c.a. c.a. % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 16.9 2.1 12.7 16.2 2.1 13.3 single slot 20.4 3.0 14.8 19.0 3.0 15.9 castellated 18.6 2.3 12.4 16.8 2.9 17.3 bowl/piston 12.6 5.8 46 12.0 1.4 11.7 squish jet 15.5 1.4 9.0 14.3 1.8 12.5 std disc 19.0 2.6 13.7 16.6 2.1 12.6 Table 5.10: Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , W O T , and 20.0 rps. Tables 115 I M E P f1 COV COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A FR=1.27 std bathtub 859.9 16.6 1.95 763.9 28.8 3.81 single slot 878.7 12.7 1.45 813.3 26.0 3.20 castellated 841.8 14.4 1.72 757.1 31.8 4.21 bowl/piston 867.9 13.6 1.57 745.5 24.8 3.34 squish jet 803.7 19.6 2.46 726.4 23.5 3.27 std disc 875.6 12.4 1.44 777.8 25.6 3.33 P E A K cr COV /* cr COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 4331 341 8.0 3729 378 10.3 single slot 4396 316 7.2 3965 436 11.0 castellated 3930 292 7.4 3631 450 12.4 bowl/piston 3878 943 24.4 4371 259 5.9 squish jet 3383 371 11.1 3204 245 7.7 std disc 4224 459 11.0 3990 597 15.1 I P E A K cr COV cr COV c a . c a . % c a . c a . % P I S T O N R A F R = 1 . 0 0 R A FR=1.27 std bathtub 14.1 2.1 14.9 13.5 1.8 13.3 single slot 16.9 2.7 16.0 17.7 3.2 18.1 castellated 19.5 2.8 14.3 17.8 3.1 17.4 bowl/piston 14.0 9.8 70 13.1 1.8 13.7 squish jet 15.5 4.4 28 14.9 2.3 15.4 std disc 16.0 2.0 12.5 14.4 2.4 16.6 Table 5.11: Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , W O T , and 33.3 rps. Tables 116 I M E P V- cr COV f1 tr COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A FR=1.27 std bathtub 884.1 17.0 1.92 800.3 36.0 4.53 single slot 906.7 18.8 2.08 841.0 36.7 4.37 castellated 864.0 17.5 2.03 790.7 36.6 4.64 bowl/piston 888.6 15.7 1.77 775.4 28.8 3.72 squish jet 772.5 18.2 2.38 722.8 42.5 5.94 P E A K f1 COV cr COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A FR=1.27 std bathtub 4910 276 5.7 4316 553 12.9 single slot 4448 335 7.6 4209 433 10.3 castellated 4009 392 9.8 3811 460 12.1 bowl/piston 4333 458 10.6 4346 300 6.9 squish jet 3345 176 5.3 3298 247 7.6 I P E A K cr COV COV c a . c a . % c a . c a . % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 12.1 1.9 15.7 12.8 2.4 18.8 single slot 16.4 2.7 16.5 15.6 2.6 16.7 castellated 16.0 3.5 21.9 16.9 3.0 17.8 bowl/piston 13.9 4.6 33.1 10.4 3.6 35 squish jet 15.5 2.5 16.1 12.7 3.5 27 Table 5.12: Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean operation at, M B T , W O T , and 50.0 rps. Tables 117 I M E P COV COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 422.6 11.3 2.7 414.3 24.1 5.8 castellated 400.0 27.6 6.9 394.3 31.7 8.0 bowl/piston 381.0 6.2 1.6 384.9 7.78 2.0 squish jet 397 11.7 3.0 414 12.7 3.1 std disc 418 127 30 400 31 7.9 P E A K CT COV COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 1586 207 13.0 1517 180 11.9 castellated 1395 212 15.2 1372 179 13.0 bowl/piston 2009 133 6.6 1999 112 5.6 squish jet 1539 329 21.6 1707 170 10 std disc 1698 326 19 1524 230 15 I P E A K o~ COV V- COV c a . c a . % c a . c a . % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 21.7 4.7 21.7 15.5 5.1 33 castellated 13.5 7.4 55 12.8 7.2 56 bowl/piston 15.4 2.7 17.5 16.9 1.4 8 squish jet 14.0 8.1 58 17.5 3.1 17.8 std disc 17.5 6.4 36 13.9 4.3 31 Table 5.13: Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , 2.5 bmep, and 33.3 rps. Tables 118 I M E P f 1 cr COV cT COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 539 10.2 1.9 526 20.1 3.8 castellated 547 22.2 4.06 544 33.7 6.2 bowl/piston 506 8.5 1.7 501 13.3 2.7 P E A K cr COV V- or COV k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 2349 251 10.7 2325 312 13.4 castellated 2271 281 12.4 2126 325 15.3 bowl/piston 2523 127 5.0 2517 171 6.8 I P E A K V- tr COV tT COV c.a. c.a. % c.a. c.a. % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 18.5 2.8 15.1 17.1 2.5 14.6 castellated 19.9 3.3 16.6 16.6 4.1 24.7 bowl/piston 12.2 3.3 27 15.7 2.5 15.9 Table 5.14: Imep, peak pressure and angle of occurance of peak pressure for different piston geometries for stoichiometric and lean R A F R at, M B T , 3.5 bmep, and 50.0 rps. Tables 119 I N I T I A L a COV cr COV 0-01% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 10.2 1.2 11.9 14.1 1.6 11.3 single slot 9.5 0.9 9.7 13.1 1.4 10.4 castellated 9.4 1.0 10.6 13.5 1.8 13.6 bowl/piston 5.4 1.4 27 10.2 1.4 13.5 squish jet 7.8 0.9 12.2 12.0 1.5 12.9 std disc 11.3 1.4 12.6 15.0 1.6 10.7 M A I N t 1 cr COV f 1 cr COV 01-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 35.1 3.2 9.1 44.1 6.4 14.7 single slot 33.8 2.3 6.9 38.4 2.7 7.0 castellated 33.0 2.1 6.5 37.4 2.7 7.2 bowl/piston 19.5 2.5 13 22.4 3.0 13 squish jet 28.5 5.7 20 33.1 7.2 22 std disc 34.0 3.3 9.8 37.6 4.5 12.2 Table 5.15: Init ia l (0-01% massburned) and M a i n (01-90% massburned) combustion durations for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 20.0 rps. Tables 120 I N I T I A L COV CT COV 0-05% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 12.2 1.4 11 16.3 1.7 10.6 single slot 11.5 1.0 8.8 15.7 1.5 9.3 castellated 11.3 1.1 9.4 16.1 1.8 11.1 bowl/piston 6.3 1.2 19 11.4 1.3 11.1 squish jet 9.4 1.0 10.4 14.4 1.5 10.5 std disc 13.1 1.7 13.3 17.2 1.9 11.2 M A I N COV cr COV 05-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 33.1 3.0 9.2 41.8 6.3 15.2 single slot 31.8 2.3 7.1 35.8 2.6 7.3 castellated 31.1 2.1 6.6 34.8 2.6 7.4 bowl/piston 18.6 2.4 13 21.2 3.1 14.6 squish jet 26.9 5.7 21 30.8 7.1 23 std disc 32.2 3.0 9.5 35.3 4.4 12.6 Table 5.16: Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 20.0 rps. Tables 121 I N I T I A L f1 tT COV tT COV 0-01% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 15.8 2.1 13.2 18.4 2.5 14 single slot 13.9 1.4 9.8 15.1 2.0 13.1 castellated 11.4 2.0 18 18.0 2.8 15.5 bowl/piston 8.8 1.3 14.5 10.7 3.0 28 squish jet 12.0 1.6 14 15.2 1.9 12.9 std disc 16.2 2.7 17 17.5 3.4 20 M A I N cr COV f1 cr COV 01-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 40.4 4.4 11.0 47.0 5.7 12.2 single slot 36.0 2.2 6.1 39.9 3.2 7.9 castellated 36.8 2.8 7.6 39.9 3.5 8.7 bowl/piston 24.4 5.7 24 26.2 7.8 30 squish jet 26.9 4.2 16 31.9 6.2 20 std disc 38.8 5.0 13.1 43.0 6.7 15.9 Table 5.17: Init ia l (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 33.3 rps. Tables 122 I N I T I A L tT COV tT COV 0-05% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 18.3 1.7 9.2 21.2 2.1 10.1 single slot 16.4 1.5 8.9 18.1 1.9 10.3 castellated 14.6 1.5 10.4 20.8 2.4 11.5 bowl/piston 10.2 1.2 11.4 12.4 2.4 19 squish jet 13.7 1.3 9.5 17.8 1.8 9.9 std disc 18.6 2.3 12.5 20.8 3.5 17 M A I N tT COV tT COV 05-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 37.9 4.5 11.9 44.2 8.2 12.4 single slot 33.5 2.1 6.3 36.9 2.9 7.8 castellated 33.6 2.2 6.7 37.0 3.1 8.5 bowl/piston 23.1 5.8 25 24.4 7.9 32 squish jet 25.2 4.2 17 29.2 6.4 22 std disc 36.4 4.8 13.2 39.7 6.9 18 Table 5.18: Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 33.3 rps. Tables 123 I N I T I A L COV 0~ COV 0-01% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 19.2 3.2 17 21.3 3.3 16 single slot 15.9 2.2 14 13.9 8.5 61 castellated 15.3 3.8 25 19.9 10.7 54 bowl/piston 11.5 1.9 17.0 11.3 3.2 28 squish jet 12.2 1.3 10.7 15.1 3.1 21 M A I N cr COV f 1 cr COV 01-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 39.8 4.6 11.5 43.0 5.4 12.6 single slot 38.4 2.7 7.2 44.5 8.6 19.4 castellated 38.3 3.9 10.0 44.4 9.0 20.2 bowl/piston 22.1 5.1 23 23.9 8.4 35 squish jet 26.8 3.6 14 30.3 7.3 24 Table 5.19: Init ial (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 50.0 rps. Tables 124 I N I T I A L cr COV tT COV 0-05% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 22.1 2.0 9.1 24.0 2.6 10.9 single slot 19.2 1.7 8.9 20.3 4.0 19.8 castellated 18.2 4.1 23 23.1 9.1 39 bowl/piston 13.0 1.6 12.2 13.1 2.8 22 squish jet 14.7 1.2 8.4 17.9 2.3 13 M A I N tT COV cr COV 05-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub 36.9 4.1 11.2 40.3 5.2 12.8 single slot 35.1 2.6 7.5 38.1 4.3 11.3 castellated 35.5 4.1 11.5 41.2 7.8 18.9 bowl/piston 20.6 5.2 25 22.1 8.4 38 squish jet 24.2 3.4 14 27.5 7.0 26 Table 5.20: Init ial (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , W O T , and 50.0 rps. Tables 125 I N I T I A L V- COV cr COV 0-01% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 18.8 2.9 15.4 20.5 6.0 29.3 castellated 21.0 4.7 22.6 23.0 5.9 25.9 bowl/piston 13.9 1.4 9.8 16.0 1.5 9.2 squish jet 16.9 2.8 16.5 18.3 3.7 20.7 std disc 20.6 4.2 20.5 24.6 38.8 16.0 M A I N cr COV a COV 01-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 45.5 5.0 11.1 53.7 8.4 15.6 castellated 52.0 9.2 17.6 53.3 10.1 19.0 bowl/piston 25.9 1.6 6.4 28.5 2.1 7.5 squish jet 30.6 4.0 13.1 37.2 6.0 16.3 std disc 48.1 7.2 15.2 51.8 6.7 13.1 Table 5.21: Init ia l (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 2.5 bmep, and 33.3 rps. Tables 126 I N I T I A L t 1 cr COV cr COV 0-05% k P a k P a % k P a k P a % P I S T O N R A FR=1.00 R A F R = 1 . 2 7 std bathtub unstable unstable single slot 22.2 2.2 10.0 24.1 5.0 21.0 castellated 24.2 4.4 18.1 26.8 4.9 18.5 bowl/piston 15.6 1.1 7.2 17.8 1.2 7.1 squish jet 18.9 2.2 11.7 20.7 2.2 10.8 std disc 23.71 4.3 18.3 27.3 2.9 10.7 M A I N • COV tT COV 05-90% k P a k P a % k P a k P a % P I S T O N R A FR=1.00 R A F R - 1 . 2 7 std bathtub unstable unstable single slot 42.1 4.6 11.0 50.1 7.8 15.7 castellated 48.8 8.9 18.3 49.5 9.8 19.8 bowl/piston 24.2 1.6 6.6 26.7 2.1 7.7 squish jet 28.6 4.0 14.3 34.8 6.3 18.3 std disc 45.0 6.3 14.2 49.0 5.8 12.1 Table 5.22: Init ia l (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean at R A F R M B T , 2.5 bmep, and 33.3 rps. Tables 127 I N I T I A L cr COV cr COV 0-01% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 18.3 3.6 19 17.2 6.6 38 castellated 21.7 8.5 39 27.5 7.8 28 bowl/piston 13.9 3.3 24 18.3 4.5 25 M A I N cr COV cr COV 01-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 44.3 4.0 8.9 51.8 6.7 13.0 castellated 44.1 6.3 14 49.7 7.8 15.8 bowl/piston 26.8 5.1 19 29.8 5.5 18.4 Table 5.23: Init ial (0-01% massburned) and M a i n (01-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 3.5 bmep, 50.0 rps. I N I T I A L cr COV /* cr COV 0-05% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 22.7 2.3 10.1 24.1 5.4 23 castellated 24.2 8.6 36 30.7 6.2 20 bowl/piston 16.7 3.0 18 20.1 4.6 23 M A I N V- cr COV A* cr COV 05-90% k P a k P a % k P a k P a % P I S T O N R A F R = 1 . 0 0 R A F R = 1 . 2 7 single slot 40.0 3.3 8.3 44.9 6.1 13.5 castellated 41.7 6.8 16 46.6 7.3 15.8 bowl/piston 24.0 5.0 21 28.0 5.7 20 Table 5.24: Init ial (0-05% massburned) and M a i n (05-90% massburned) combustion duration for different piston geometries for stoichiometric and lean R A F R at M B T , 3.5 bmep, and 50.0 rps. NAT. GAS = £ c GASOLINE = £ = AVL OPTICAL Instilment signalj r» out COOLING WATER $ "ATI — H Y D R A E N G I N E O I L C O O L I N G M O D U L E . sh.n W A T E R D Y N A M O M E T E R TT TACHO IETER TORQUE LOAD CELL TEST CELL CONTROL ROOM C O N T R O L  C O N S O L E . control *~ signals out ^ tnstruawnt * signals in Instrument " signals out G I B M "PT A / D S Y S T E M - I N C O M P U T E R analog slgnalsV ^ d i g i t a l "out ) C I R C U I T B O A R D H O U S I N G •—condit ioned signals out 1 control ^ signals C O N V E R T E R C A B I N E T control signals , n power at In or f out T R A N S F O R M E R POWER SUPPLY 5 B n cn Figure 3.1: Ricardo Hydra engine, dynamometer and control systems layout. to oo Figures 129 Figure 3.2: Ricardo Hydra MKIII Gasoline (or gaseous fuel) Engine cross-sectional and longitudinal views. DETAIL O F C O M B U S T I O N C H A M B E R Figures 131 S E C T I O N T - T Figure 3.4: Single slot and castellated piston geometri Figures 132 Piston no.7 CR « 9.0:1 Figure 3.5: Bowl-in-piston and squish jet piston geometries. Figures 133 Figure 3.6: Instrumentation layout for the Ricardo engine test cell. Figure 3.7: Hot wire probe location through the spark plug entry for the bathtub and flat cylinder heads. Figures 135 RS232 Un* to PC Figure 3.8: Acquisition hardware arrangement with fast pressure data acquisition hook up. Figures 136 Figure 4.1: Comparison of the effect of window size on the turbulent intensity profile for the bathtub chamber at WOT, 33.3 rps. Figures 137 -90 -60 -30 0 30 60 CRANK ANGLE DEGREES FROM TDC Figure 4.2: Polytropic coefficent calculated from the ensembled pressure for single slot chamber at MBT, WOT and 33.3 rps for RAFR=1.27. Figures 138 2 0 0 0 1500 1000 500 0 . std bathtub single slot caste 11ated bowl/piston squish jet ' '• std disc ,iff \ \ if X 1 V 'ii V ''/ '// \ \ \ jJL-— 1 1 1 1 1 1 1 1 1 1 1 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.1: Motored pressure profiles for different chamber geometries at W O T , 20.0 rps. Figure 5.2: Motored pressure profiles for different chamber geometries at W O T , 33.3 rps. Figures 139 Figure 5.3: Motored pressure profiles for different chamber geometries at W O T , 50.0 rps. C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.4: Motored pressure profiles for different chamber geometries at W O T , 66.7 rps. Figures 140 Figure 5.5: Motored pressure profiles for the single slot piston at WOT, for three speeds; 20.0, 33.3, and 50.0 rps. 800 700 600 500 400 300 std bathtub single slot castellated bowl/piston \ squish Jet i \ std disc if% $' % if \ i \ \ //'/ •/ / ' \ -180 -150 -120 -90 -60 -30 30 60 90 C R A N K A N G L E D E G R E E S F R O M T D C 120 150 180 Figure 5.6: Motored temperature profiles for different chamber geometries at WOT, 33.3 rps. Figures 141 20 C R A N K A N G L E D E G R E E S F R O M T D C 10 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.7: Window ensembled and cycle ensembled mean velocity and turbulent inten-sity profiles for the single slot piston at WOT, 33.3 rps. Figures 142 Figure 5.8: Mean velocity profiles for different chamber geometries at WOT, 33.3 rps. sta; bathtub s i n g l e s l o t -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.9: 33 3 rps. Turbulent intensity profiles for different chamber geometries at WOT, Figures 143 20 Figure 5.10: Mean velocity profiles for the 'bathtub' group of chambers at WOT, 20.0 rps. 5 _, '. , C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.11: Turbulent intensity profiles for the 'bathtub' group of chambers at WOT, 20.0 rps. Figures 144 20 18 16 s td bathtub s i n g l e s l o t c a s t e l l a t e d 14 . 1 1 1 1 1 1 1 1 1 1 r r~ -180 -150 -120 -90 -60 -90 0 30 60 90 120 150 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.12: Mean velocity profiles for the 'bathtub' group of chambers at W O T , 33.3 rps. 5 Figure 5.13: 33.3 rps. Turbulent intensity profiles for the 'bathtub' group of chambers at W O T , Figures 145 H o o •J a > 20 18 16 14 12 10 8 6 4 2 0 — std disc — bowl/piston — squish Jet i 1 1 1 1 1 1 — r r -180 -150 -120 -90 -60 -30 0 30 60 .90 120 150 CRANK ANGLE DEGREES FROM TDC Figure 5.14: Mean velocity profiles for the 'disc' group of chambers at WOT, 20.0 rps. std d1sc bowl/piston squish Jet I _| , " I - 1 - " j !——-(  j. .fl j— -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 CRANK ANGLE DEGREES FROM TDC Figure 5.15: Turbulent intensity profiles for the 'disc' group of chambers at WOT, 20.0 rps. Figures 146 Figure 5.16: Mean velocity profiles for the 'disc' group of chambers at W O T , 33.3 rps. Figure 5.17: Turbulent intensity profiles for the 'disc' group of chambers at W O T , 33.3 rps. Figures 147 r-H o o > 20 18 20.,0-srps 33,: 3 rps 66.7 rps 30 60 90 120 150 CRANK ANGLE DEGREES FROM TDC Figure 5.18: Mean velocity profiles for the bathtub chamber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. '. 20.0 rps \ 33.3 rps \ 66.7 rps I i i i i i i i 1 1 r 1 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 CRANK ANGLE DEGREES FROM TDC Figure 5.19: Turbulent intensity profiles for the bathtub chamber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. Figures 148 20 18 16 . 14 12 20.0 rps 33.3 rps 50.0 rps 30 60 90 120 150 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.20: Mean velocity profiles for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. Figure 5.21: Turbulent intensity profiles for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.23: Turbulent intensity profiles for the squish jet chamber at WOT, for three speeds; 20.0, 33.3 and 50.0 rps. Figures 150 H o o > 20 18 16 14 12 10 8 6 4 2 0 — 20.0 rps -•' 33.3 rps — 66.7 rps V IVC EVO • -180 -150 -120 -90 -60 -30 0 30 60 90 CRANK ANGLE DEGREES FROM TDC Figure 5.24: Mean velocity profiles for the disc chamber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. 20.0 rps 33.3 rps — 6 6 . 7 rps 1 I i I I I I I l I r l -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 CRANK ANGLE DEGREES FROM TDC Figure 5.25: Turbulent intensity profiles for the disc chamber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. Figures 151 2.5 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.26: Mean velocity profiles scaled with mean piston speed for the bathtub cham-ber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. 0.8 0.6 _ 0 . 4 0.2 20.0 rps 33.3 rps 66.7 rps C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.27: Turbulent intensity profiles scaled with mean piston speed for the bathtub chamber at WOT, for three speeds; 20.0, 33.3 and 66.7 rps. Figures 152 1 1 h 1 1 1 1 1 1 T r i I -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.28: Mean velocity profiles scaled wi th mean piston speed for the single slot chamber at W O T , for two speeds; 20.0 and 33.3 rps. l 5 o i ! f-1 i 1 1 1 1 1 1 1 1 i r 1 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.29: Turbulent intensity profiles scaled wi th mean piston speed for the single slot chamber at W O T , for two speeds; 20.0 and 33.3 rps. -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.30: Mean velocity profiles scaled with mean piston speed for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. l C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.31: Turbulent intensity profiles scaled with mean piston speed for the castellated chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. Figures a a a a. co Z o H co Z < s H O O H > 2.5 1.5 0.5 154 20.0 rps 33.3 rps IVC E V O -180 I I -120 -90 -60 C R A N K A N G L E D E G R E E S F R O M T D C 120 150 180 Figure 5.32: Mean velocity profiles scaled wi th mean piston speed for the bowl-in-piston chamber at W O T , for two speeds; 20.0 and 33.3 rps. D a B OH cn z o H z < a S ;» H t—4 CO Z Ed H Z t-i z a m OH D H 0.8 . 0.6 0.4 0.2 . 20.0 rps 33.3 rps 1 1 1 --180 -150 -120 -90 60 -30 0 30 60 C R A N K A N G L E D E G R E E S F R O M T D C 180 Figure 5.33: Turbulent intensity profiles scaled wi th mean piston speed for the bowl-in-piston chamber at W O T , for two speeds; 20.0 and 33.3 rps. Figures 155 1 1 1 1 1 1 1 1 1 1 r i I -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.34: Mean velocity profiles scaled wi th mean piston speed for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. g 0.2 . a ^> D CQ -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.35: Turbulent intensity profiles scaled wi th mean piston speed for the squish jet chamber at W O T , for three speeds; 20.0, 33.3 and 50.0 rps. 1 1 1 1 1 1 1 1 1 1 r 1 I -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.36: Mean velocity profiles scaled wi th mean piston speed for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps. -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.37: Turbulent intensity profiles scaled wi th mean piston speed for the disc chamber at W O T , for three speeds; 20.0, 33.3 and 66.7 rps. Figures 157 o Z a h. a <! s a a H a < tf ca 34 32 30 28 26 24 22 20 o — ° s t d b a t h t u b s i n g l e s l o t B---e c a s t e l l a t e d a a B • _ - B B"" 1 1.05 T 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1.35 1.4 Figure 5.38: Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub ' group of chambers at M B T , W O T and 20.0 rps. >« o z a o a a a •J < S a s H a n 34 32 30 28 26 24 22 20 s t d b a t h t u b s 1 n g 1 e s l o t B - - - Q c a s t e l l a t e d . A ' B--..-•J..B-. A - " -e -B — B- c I 1.05 I T " 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1.35 1.4 Figure 5.39: Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub ' group of chambers at M B T , W O T and 33.3 rps. Figures 158 o z a o E a < S a: « B H a < oS 03 34 . 32 30 28 . 26 24 22 20 std bathtub single slot B---e castel lated . . . A * 1 I I I 1 1 1 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 R E L A T I V E AIR F U E L RATIO Figure 5.40: Brake thermal efficiencies for stoichiometric to lean operation for the 'bath-tub ' group of chambers at M B T , W O T and 50.0 rps. >> o z a i—i o E a a < S cS a x H a <! OS m 34 . 32 30 28 26 24 . 22 20 std disc A - - - O bowl /piston B - - - B squish Jet . — •-Br' __.„.-G B B o 1 1.05 I I 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO I I 1.3 1.35 1.4 Figure 5.41: Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 20.0 rps. Figures 159 o z a CD 34 32 30 . a a 28 <: S as a s 26 24 . 22 20 B -B B - B 0 — 0 s td d i s c bowl / p i s t o n s q u i s h Jet —I r~ 1.25 1.3 1.35 1.4 1 1 1 r— 1 1.05 1.1 1.15 1.2 R E L A T I V E AIR F U E L RATIO Figure 5.42: Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at MBT, WOT and 33.3 rps. >• v z a a a a < 2 oS a x H a as CD 34 32 . 30 28 26 . 24 22 . 20 . - - - * * * * * A ' — o—e s td d i s c bowl / p i s t o n B - - - B s q u i s h Jet 1 i i l 1.05 1.1 1.15 1 1 1 1 1.2 1.25 1.3 1.35 1 R E L A T I V E AIR F U E L RATIO Figure 5.43: Brake thermal efficiencies for stoichiometric to lean operation for the 'disc' group of chambers at MBT, WOT and 50.0 rps. Figures 160 o o H CD cn W a o; o a a a o z D < Z o t—t H Z o 45 40 35 30 . 25 g 20 . 15 s td bathtub s i n g l e s l o t c a s t e l l a ted 1 1.05 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1 1.3 1 1.35 1.4 Figure 5.44: Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 20.0 rps. 45 O Q t-i CO. CO a a A o a Q a o z $ Q < Z o H .2 40 . 35 . 30 25 20 15 o — 0 s td bathtub s i n g l e s l o t c a s t e l l a t e d .-0 B : ; rGr& ' - B - - " - " o ' - • " - A * - B - " A - - - - - A 1.05 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1.35 1.4 Figure 5.45: Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 33.3 rps. Figures 161 o Q H CQ co cc o a Q o z $ D <i Z o H t—1 Z o 45 40 . 35 30 . 25 20 15 o — ° s td bathtub a---*, s i n g 1 e s l o t c a s t e l l a t e d T 1.05 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1.35 1.4 Figure 5.46: Ignition advance for stoichiometric to lean operation for the 'bathtub' group of chambers at M B T , W O T and 50.0 rps. o a H CQ cn a a OS o a D a o z Q < Z o r—( Z 40 35 30 25 20 2 15 10 Q — ° s t d d i s c A---A bowl / p i s t o n Q--- 0 s q u i s h Jet / / B--~~" / _ - - B - " " " ' JB .A - - -A - . . . " " • A >• . -A' A'" 1 1 1 1 1 1 1 1.05 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1.35 1.4 Figure 5.47: Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 20.0 rps. Figures 162 40 35 30 25 20 . 15 10 B--_.-e er' a — 0 s td d i s c bowl / p i s t o n Q - ' - B s q u i s h j e t —i 1 n 1 1 1.05 1.1 1.15 1.2 1.25 R E L A T I V E AIR F U E L RATIO 1.3 1 1.35 1.4 Figure 5.48: Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 33.3 rps. 40 35 30 25 20 a 15 10 Br' tT 0 — ° s t d d i s c b o w l / p i s t o n s - - " 6 1 s q u i s h j e t 1 1 1 1 1 1 1 1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 R E L A T I V E AIR F U E L RATIO Figure 5.49: Ignition advance for stoichiometric to lean operation for the 'disc' group of chambers at M B T , W O T and 50.0 rps. Figures 163 5000 f i r e d p ressure motored p r e s s u r e 4000 3000 2000 1000 T -540 -450 -360 -270 -180 -90 C R A N K A N G L E D E G R E E S F R O M T D C 180 Figure 5.50: Ensembled fired and motored pressure profiles over four strokes for the single slot chamber at W O T and 33.3 rps. F i red trace for M B T and R A F R = 1 . 0 0 . Figures 164 « CL. a tf a cu Q a j CQ S a co Z a 5000 4000 3000 2000 1000 s td bathtub s i n g l e s l o t c a s t e l l a t e d 1 1 1 1 1 I I I -180 -150 -120 -90 -60 -30 0 30 60 90 C R A N K A N G L E D E G R E E S F R O M T D C 120 150 180 Figure 5.51: F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 . o a. Jt a tf D cn cn a tf e. Q a •a CQ S a CO Z a 5000 4000 . 3000 2000 1000 . s t d bathtub s i n g l e s l o t c a s t e l l a t e d -180 ~~1 1 -150 -120 -90 -60 60 90 C R A N K A N G L E D E G R E E S F R O M T D C 150 180 Figure 5.52: F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 20.0 rps for R A F R = 1 . 2 7 . Fig uies 5000 165 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.53: F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for RAFR=1.00. a ai D a ai OH D a pa S a in Z a 5000 4000 . 3000 . 2000 1000 s t d bathtub s i n g l e s l o t c a s t e l l a t e d 1 1— 180 -150 -120 -90 -60 -30 0 30 60 90 C R A N K A N G L E D E G R E E S F R O M T D C 180 Figure 5.54: F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T and 33.3 rps for R A F R = 1 . 2 7 . Figures 166 5000 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.55: F i red pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 . 5000 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.56: F ired pressure profiles for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 . Figures 5000 s t d d i s c b o w l / p i s t o n -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.57: F ired pressure profiles for the 'disc' group of chambers at M B T , W O T , an 20.0 rps for R A F R = 1 . 0 0 . 5000 . s t d d i s c b o w l / p i s t o n -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.58: F ired pressure profiles for the 'disc' group of chambers at M B T , W O T , an 20.0 rps for R A F R = 1 . 2 7 . Figures 168 5000 s td d i s c b o w l / p i s t o n 1 I I I I I I 1 1 1 1 1 -180 -150 -120 -90 -60 -30 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.59: Fired pressure profiles for the 'disc' group of chambers at MBT, WOT, and 33.3 rps for RAFR=1.00. 5000 : , a 0. M a 0. D cn cn § O. O H J CQ s a cn Z a 4000 3000 2000 . 1000 "i r -180 -150 -120 -90 -60 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.60: Fired pressure profiles for the 'disc' group of chambers at MBT, WOT, and 33.3 rps for RAFR=1.27. Figures 169 a On M a a D co CO a cc a, Q H a S 63 co 2 H 5000 ' 4 0 0 0 . 3000 . 2000 1000 . bowl/piston squish Jet I I - 1 8 0 - 1 5 0 - 1 2 0 90 - 6 0 - 3 0 C R A N K A N G L E D E G R E E S F R O M T D C 120 150 180 Figure 5.61: F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 . 5000 a OH M a ai D co co a ai a, Q a oa S a CO 2 a 4000 . 3000 . 2000 1000 . bowl/piston squish Jet - 1 8 0 - 1 5 0 - 1 2 0 - 9 0 - 6 0 - 3 0 0 30 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.62: F i red pressure profiles for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 . Fig ures 17.0 « Cu Jt a C6 cn cn cc ft. Q B CQ S B cn Z a 2500 2000 1500 1000 . 500 s i n g l e s l o t c a s t e l l a t e d 30 60 -180 -150 -120 -90 -60 -30 0 90 C R A N K A N G L E D E G R E E S F R O M T D C 180 Figure 5.63: Fired pressure profiles for the 'bathtub' group of chambers at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.00. a a. Jt a c£ cn cn a ns O H Q a CQ S a CO Z a 2500 2000 1500 1000 500 s i n g l e s l o t c a s t e l l a t e d —1 1 1 1 1--180 -150 -120 -90 -60 -30 30 -1 1 1 1 60 90 120 150 180 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.64: Fired pressure profiles for the 'bathtub' group of chambers at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.27. Figures 171 0 0. « OH P to a OH OH Q w CQ H cn Z w 2500 2000 1500 . 1000 500 s t d d i s c b o w l / p i s t o n s q u i s h j e t I I 1 -180 -150 -120 -90 -60 -30 0 30 60 90 CRANK ANGLE DEGREES FROM TDC 120 150 180 Figure 5.65: Fired pressure profiles for the 'disc' group of chambers at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.00. a Cu J* a ai D co CO 63 ai OH Q a •J oa 63 CO Z a 2500 2000 1500 . 1000 500 std d i s c b o w l / p i s t o n s q u i s h Jet CRANK ANGLE DEGREES FROM TDC Figure 5.66: Fired pressure profiles for the 'disc' group of chambers at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.27. Figures 172 3000 2500 •* 2000 & Jo cn cn § 1500 Q a CD 1000 500 . s . l n g l e s l o t c a s t e l l a t e d b o w l / p i s t o n 1 1 1 1 1 1 1 r -180 -150 -120 -90 -60 -30 0 30 60 90 CRANK ANGLE DEGREES FROM TDC 120 150 180 Figure 5.67: Fired pressure profiles for three different chamber geometries MBT, Bmep=3.5, and 50.0 rps for RAFR=1.00. 3000 2500 . 2000 . 1500 1000 500 s i n g l e s l o t c a s t e l l a t e d b o w l / p i s t o n U \ fi Ii \ \ \ \ Ii \ \-\.-Vy, r - ^ -180 -150 -120 -90 -60 -30 0 30 60 90 CRANK ANGLE DEGREES FROM TDC 120 150 180 Figure 5.68: Fired pressure profiles for three different chamber geometries at MBT, Bmep=3.5, and 50.0 rps for RAFR=1.27. Figures 173 I I I i I I I I l -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.69: Mass fraction burned curve for the bathtub chamber at MBT, WOT and 33.3 rps for RAFR=1.27 r I l I I 1 1 1 " 6 ° -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.70: Mass fraction burned curves for different chamber geometries at MBT, WOT and 33.3 rps for RAFR=1.00 Figure 5.71: Mass fraction burned curves for different chamber geometries at MBT, WOT and 33.3 rps for RAFR=1.27 Figures 175 Figure 5.72: Mass fraction burned curves for the bathtub chamber at M B T , W O T and R A F R = 1 . 2 7 for five speeds; 20.0, 33.3, 40.0, 50.0 and 66.7 rps. C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.73: Mass fraction burned curves for the 'bathtub' group of chambers at M B T W O T , and 20.0 rps for R A F R = 1 . 0 0 . -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.74: Mass fraction burned curves for the 'bathtub' group of chambers at M B T W O T , and 20.0 rps for R A F R = 1 . 2 7 . I I I 1 1 1 1 1 1 -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.75: Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for RAFR=1 .00. I I I I i i i -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.76: Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 33.3 rps for RAFR=1.27. Figures 178 Figure 5.77: Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 . I I i I i i i 1 -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.78: Mass fraction burned curves for the 'bathtub' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 . Figures 179 -60 -45 -30 -15 0 15 30 45 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.79: Mass fraction burned curves for the 'disc' group of chambers at MBT, WOT, and 20.0 rps for RAFR=1.00. -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.80: Mass fraction burned curves for the 'disc' group of chambers at MBT, WOT, and 20.0 rps for RAFR=1.27. Figures 180 -60 -45 -30 -15 0 15 30 45 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.81: Mass fraction burned curves for the 'disc' group of chambers at MBT, WOT, and 33.3 rps for RAFR=1.00. -60 -45 -30 -15 0 15 . 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.82: Mass fraction burned curves for the 'disc' group of chambers at MBT, WOT and 33.3 rps for RAFR=1.27. Figures 181 Q H Z oi D m z o H O •< CC Eh in cn <: 2 0.8 0 .6 . 0 .4 . 0.2 b o w l / p i s t o n squ ish Jet 1 1 1 1 1 1 1 -60 -45 -30 -15 0 15 30 45 C R A N K A N G L E D E G R E E S F R O M T D C 60 Figure 5.83: Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 0 0 . Q a z OH D m z o H O < a a cn cn < 0.8 . 0 .6 . 0.4 . 0.2 . b o w l / p i s t o n s q u i s h j e t T " T -60 -45 30 -30 -15 0 15 C R A N K A N G L E D E G R E E S F R O M T D C 45 60 Figure 5.84: Mass fraction burned curves for the 'disc' group of chambers at M B T , W O T , and 50.0 rps for R A F R = 1 . 2 7 . Figures 182 I 1 !—i 1 1 1 1 1 1 -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.85: Mass fraction burned curves for five different chamber geometries at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.00. Figure 5.86: Mass fraction burned curves for five different chamber geometries at MBT, Bmep=2.5, and 33.3 rps for RAFR=1.27. Figures 183 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.87: Mass fraction burned curves for three different chamber geometries MBT, Bmep=3.5, and 50.0 rps for RAFR=1.00. -60 -45 -30 -15 0 15 30 45 60 C R A N K A N G L E D E G R E E S F R O M T D C Figure 5.88: Mass fraction burned curves for three different chamber geometries at MBT, Bmep=3.5, and 50.0 rps for RAFR=1.27. Figures 184 1.2 O i—i a o « z rt p « z o o < ai fa cn cn •< s 0.6 0.6 . y o.4 0.2 . I n i t i a l burn:0 - 5% main burn :5 - 90% n i i i i r bathtub s i n g l e c a s t l e bowl s q . J e t d i s c Figure 5.89: Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T , and 20.0 rps for R A F R = 1 . 0 0 . 1.2 0.8 . 0.6 5 CC Q a z ai P co z o H O < ai b. cn 0 .2 cn < 2 r> 0.4 1 I n i t i a l burn:0 - 5% main burn :5 - 90% i i i i i r bathtub s i n g l e c a s t l e bowl s q . J e t d i s c Figure 5.90: Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for M B T , W O T and 20.0 rps for R A F R = 1 . 2 7 . (2 i-l n cn CD to o H p a CO M» C O ^ • p co n •1 cn i-i > to o a cr n a CD a. cr P n 00 •i P a-co H CD CD a -CD cr p a cr <•> a* p MASS F R A C T I O N B U R N E D RATIO o o co cr <D O H p a co C O C O I-I t3 CD O I-I Tl P o o" a cr a a CD ex. n p 0Q i-« P T3 a* cn i-l CD CD a-CD cr p a cr MASS F R A C T I O N B U R N E D RATIO 5 Oct c I-l CD cn o lO o o CO cr CD I-l O i i O O C T T Figures 186 Figure 5.93: Mass fraction burned ratio bar graphs, relative to the bathtub chamber, MBT, WOT, and 50.0 rps for RAFR=1.00. < S 1.2 0.6 § : 0 - 8 ai a a z. ai D m z O 0 .4 H O < CH CL. 0.2 1 I n i t i a l burn:0 - 5% main burn :5 - 90X i i r bathtub s i n g l e c a s t l e bowl 1 1 s q . J e t d i s c Figure 5.94: Mass fraction burned ratio bar graphs, relative to the bathtub chamber, for MBT, WOT, and 50.0 rps for RAFR=1.27. Figures 187 1.2 5 tf Q a z tf D CQ z o H o <: cn cn «s! s 1 . o.a 0.6 a o.4 0.2 bathtub s i n g l e c a s t l e bowl s q . J e t d i s c Figure 5.95: Mass fraction burned ratio bar graphs, relative to the disc chamber, for M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 0 0 . cn cn < S 1.2 1 . 0.6 P 0.8 <I tf Q H Z tf D CQ Z 2 0.4 H O <! tf 0.2 I n i t i a l burn:0 - 5% main burn :5 - 90J> { |--- i i i r bathtub s i n g l e c a s t l e bowl s q . J e t d i s c Figure 5.96: Mass fraction burned ratio bar graphs, relative to the disc chamber, for M B T , Bmep=2.5, and 33.3 rps for R A F R = 1 . 2 7 . Figures 188 < cd a co co •< s 1.2 H 0.8 •<! tt P a z cd D CQ z 2 0.4 J 0.6 . 0.2 1 i n i t i a l burn-.O - 5% main burn :5 - 90* bathtub s i n g l e c a s t l e I 1 bowl s q . J e t d i s c Figure 5.97: Mass fraction burned ratio bar graphs, relative to the castellated chamber, M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 0 0 . 1.2 0.6 H 0.8 cd D a z cd CQ Z 2 0.4 j u <: cd a c o 0.2 co •< 2 m I n i t i a l burn:0 - 5% main burn :5 - 90* i 1 r 1 1 r bathtub s i n g l e c a s t l e bowl s q . J e t d i s c Figure 5.98: Mass fraction burned ratio bar graphs, relative to the castellated chamber, for M B T , Bmep=3.5, and 50.0 rps for R A F R = 1 . 2 7 . Figures 189 «3 a OS D co co a OS o. a > H O a a fa a z < a IS Q a p—i o z 900 875 850 825 800 775 750 . 725 700 IMEP per c y c l e mean mean+std dev mean-std dev " I 1 1 I I I 20 40 60 80 100 120 C Y C L E N U M B E R 140 160 180 200 Figure 5.99: Indicated mean effective pressure per cycle for the single slot chamber at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 : 200 cycles Figure 5.100: Indicated mean effective pressure per cycle for the bathtub chamber at M B T , W O T , and 33.3 rps for R A F R = 1 . 2 7 : 44 cycles Appendix A Instrument Specification and Calibration This appendix contains the manufacturers calibration curves for: • Kis t ler piezo-electric pressure transducers; 6121 • Mer i am Laminar flow element 50MW20-1.5 (Fuel flow rate) • M e r i a m Laminar flow element 5 0 M C 2 - 4 F ( A i r flow rate) Specifications for the Kist ler pressure transducer and sample calibration data are also presented. Pressure transducer Quasi static calibration of the pressure transducer and amplifier was car-ried out in accord wi th Ricardo consultants recommendations [70]. Calibra-tion was used to confirm the manufacturers constants and to confirm faulty operation suspected from the operating pressure trace. The calibration curves for the three pressure transducers used during this investigation are given i n Figures A . l to A . 3 . The high input resistance of the Kist ler 5004 amplifier unit , set on 'long' response makes it suitable for a quasi static calibration procedure using a dead weight tester. The amplifier unit was set to the sensitivity and range used during the fired operation, v iz . , sensitivity « 14 p C / b a r and gain 10. After resetting the amplifier a weight was suddenly applied and the response 190 endix A. Instrument Specification and Calibration M O D E L 6121 R A N G E U N I T S Range 0 . . . 250 bar Cal ibrated partial ranges 0 . . . 25 bar 0 . . . 2.5 bar Overload 350 bar Sensitivity w -14 p C / b a r Natura l frequency > 55 k H z Frequency response ± 1 % 6 k H z Linearity, al l ranges < ± 1.0 % F S O Hysteresis < 1.0 % F S O Acceleration sensitivity: axial < 0.003 bar/g transverse < 0.0002 bar/g Shock resistance 2000 g Thermal sensitivity shift: 20 . . . 350 ° C < ± 3 % 200 ± 50 ° C « ± 1 % Calibrated i n range 20 . . . 350 ° c Operating temperature range -196 . . . 350 " C Transient temperature error < 0.02 bar Table A . l : Pressure transducer specifications for Kist ler model 6121. Appendix A. Instrument Specification and Calibration 192 Pressure (psi) A m p output (mvolt) Pressure (bar) Charge output (pC) 15 128 1.03 17.9 35 253 2.41 35.4 55 415 3.79 58.1 75 533 5.15 74.6 105 775 7.24 108.5 125 910 8.58 127.4 155 1140 10.69 159.6 205 1450 14.14 203.0 Table A . 2 : Pressure transducer calibration data for Kist ler model 6121. No. 317205 voltage measured on a digital oscilloscope. Cal ibrat ion data for the pressure transducer No. 317205 is given i n Table A . l and indicated on Figure A . l . The charge produced by the transducer is calculated from: Q = V * Cg* k where V is the amplifier output in volts, Cg is the value of the range capacitor in p F and A; is the sensitivity setting of the amplifier expressed as a part of unity, eg.; sensitivity of 1.4 mechanical units per volt is equivalent to 0.14. Laminar flow elements The laminar flow elements calibration curves are given in Figures A . 4 and A . 5 . Corrections for standard air pressure and temperature (21.1 °C,101.3 kPa) were calculated using the manufacturers supplied tabulated values. These values were approximated to linear equations for use i n the D A T A Q and C R U N C H programs. The viscosity of natural gas, used i n these equations was determined from the approximate gas composition given in Appendix C . Appendix A. Instrument Specification and Calibration 193 Druckaufnehmer Capteur de pression Pressure transducer Type 6121 SN 317205 Kalibrierter Bereicti Gamme elalonnee iDar] Calibrated range 0...25O 0-.25 0-S.5 Betriebstemperaturbereich Gamme de temp, d'utiitsalion f*C) -60.-350 Operating temperature range Emplindlicfttieil Sensibilite [pC/bar] Sensitivity -1t,0 -n.o -14,0 Kalibriert bet Etalonne a 20 *C Calibrated at by Sh Date 3.12.8? Unearitat _ . Linearite <±%FSO 0,3 o,3 0,3 1 bar s 10' N • 1^"' = 1,019...at = 14,50...psi 1al = 1kpcm_,B 1kgfcrrr' = 0,960665 bar 1psi=0,06894...bar Figure A . l : Ki s t ler and laboratory calibration curves for pressure transducer Mode l 6121 No. 317205, (used for single slot castellated, bowl-in-piston and fired squish jet tests). Appendix A. Instrument Specification and Calibration 194 Daickaufnehmer Capteur de pression Pressure transducer Type 6121 S N 282737 Kafibrierter Bereich Gamine etalonnee [bar] Calibrated range 0 - 2 5 0 0 _ . 2 5 0 - 2 . 5 Betriebsiemperaturbereich Gamme de temp, d'utilisation PC] -80-350 Operating temperature range EmpfincJIicfikeit Sensib le [pC/Oar] Sensriiviry -H,7 -H,6 Kalibriert b*i Etafonne a 20 *C Calibrated a) by Sh .Date 15.12.86 Lineantat Linearite ' < i % F S U 0,3 0,3 0,3 1 bar = 10* N • rrr* = 1,019...al = 14.50...psi 1 at = 1 kp • crrr' = 1 kgl t i r r 1 • 0,980655 bar Figure A.2: Kistler calibration curve for pressure transducer Model 6121 No. 282737, (used for standard bathtub and disc chamber tests). Appendix A. Instrument Specification and Calibration 195 Druckaufnehmer Capteur de pression Pressure transducer Type 6121 SN 3 1 7 1 2 5 Katibrierter Bereicfi Gamme etalonnee [bar] Calibrated range 0...2S0 0...25 0...2.5 Betriebatemperaturtwrettfi Gamme de temp, d'utilisatlon [*C] -S0...350 Operating temperature range Emptindltcnkett Sensibilite [pC/bar] Sensitivity -14.8 -14.5 -14.5 Kalibriert bei Etalonne d 20 9 C Calibrated at by Sh Date 11.8.88 Linearital _ „ Linearite <+%FSO 0.3 0.3 0.3 1 bar * 10' N • itr' = 1.019...at = u,50...psi 1at= 1kpcm-'= 1kgtcrrr' = 0,980665 bar 1psi-0.06894...bar 0 25 50 75 100 125 150 175 200 225 250 0 2.S 5 7.5 10 12,5 15 17.5 20 22.5 25 0 0.2S 0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 Abhftngigkett der Empfindlfc*ikect [ . , [ . : , . I ] , ; , | j j • — ' I i . j . — 4 - 3 i , von der Temperatur —f 1 '• : . l ~ ^ ± ^ s | = ' , "[ ^ — . ^ 1 • """"] Sensibilite en fonction r * " " " l ^ • • | ; [ \ '• | • --- I* ' - 1 ^ oC/bar de ta temperature J "*"~ ~t' , I I ! I ' , I J_ • ... j — - — j - — - j . . . • Sensitivity versus temperature | • • . . . . . j . •••_—] I ...... I . .1 1 Figure A . 3 : Kis t ler calibration curve for pressure transducer Model 6121 No. 317125, (used for the motored squish jet test only). Appendix A. Instrument Specification and Calibration 195 Druckaufnehmer Capteur de pression Pressure transducer Type 6121 S N 3 1 7 1 2 5 Kalibrierter Bereicti Gamme etalonnee [bar] Calibrated range 0...250 0...25 0...2.5 Setriebstemperaturbereich Gamme de temp, d'utilisation f*C) -80...350 Operating temperature range Empfindlidikeit Sensibilile [pC/bar] Sensitivity -H.B -1"i.5 - T i . 5 Kaltbrtert bei Etalonne a 20 a C Calibrated at by Sh Date 11.8.88 Linearitat Linearite < + T » r S O 0.3 0.3 0.3 1bar = 10' N mr' = 1.019...at = 14.50...psi 1at=1kpcm-'=1kgtcm- , = 0.980665 bar 1psi=0.06894...bar 0 25 50 75 100 125 150 175 200 225 250 0 2.5 5 7.5 10 12,5 15 17.5 20 22.5 25 0 0.25 0.5 0.75 1 1.25 1.5 1.75 2 2.25 Abhangigkert der Empfindlichkeit U . . . "j ; i.•..••TT . 1 . | ; ; • ' | ' • I ' I ' ; j ' . ' ' ] 3'-* von der Temperatur I ; : t ; ; : • ! ,,' • ' T *—-— j ; ; ' . ' -.^~J Senaibilrte en tonction L^rrrr.jr. I 1.11 : ; ' 1 ' .. ITT. , 7^|TTZL^T4 - ^ . 8 pC/bar de la temperature ~ + — - 4 -!• • •• 4- — | •—-|~~' •-•{• — •• Sensitivity versus temperature —f- -3'.. 350 C C Figure A . 3 : Kis t ler calibration curve for pressure transducer Mode l 6121 No. 317125, (used for the motored squish jet test only). Appendix A. Instrument Specification and Calibration 196 Figure A . 4 : Mer i am calibration curve for Mode l 50MW20-1.5 No. S-4875-l,( used for Natura l gas flow rate). Appendix A. Instrument Specification and Calibration 197 Figure A . 5 : Mer i am calibration curve for Mode l 50MC2-4F No. S-4875-2,( used for A i r flow rate). Appendix B Hot Wire Anemometry Specification and Calibration This appendix presents the specifications for the hotwire probe and bridge unit and typical calibration data. Cal ibrat ion of the hotwire system was carried out i n a small wind tunnel for velocities between 0.5 and 16 m/s i n accord with D I S A recommendations. The internal probe and cable resistance were compensated for, through the rezeroing of the bridge unit for the short-circuited cable, and by subtracting the internal probe resistance from the measured resistance value. Typica l values of sensor resistance after annealing for 6-8 hours at the operating tem-perature were 10.8 fi ambient and 12.1 fi operating. The resistance drop during the annealing process, before a stable value was reached, was approx-imately 1 fi. Balance of the bridge was obtained by applying a square wave to the probe exposed to the maximum velocity i n the wind tunnel and adjusting the cable compensation for a clear response signal. A pitot tube and inclined alcohol manometer provided flow calibration data against a mean voltage unit . These data were then used i n the calibration program H W - C a l to obtain the analytical model constants. The exponential constant n varied from 0.36 to 0.66 over the wires used , showing a good agreement with the in i t i a l theory of K i n g [78]. Sample data for wire No. 1 used for the bowl-in-piston chamber flow field 198 Appendix B. Hot Wire Anemometry Specification and Calibration 199 P R O B E : M o d e l TSI 1226 No. 65202 Internal Resistance o.57 fi W i r e material TSI P la t inum Ir id ium, PI2.5 Diameter 6.3 pm Length 1.5 m m Thermal coeff. of resistance .00089/ °C Operating temperature 600 ° C B R I D G E : Mode l D I S A Type 55M10 C T A Standard F I L T E R : Mode l D I S A Type 55D25 Aux i l i a ry unit Low pass filter 20 k H z High pass filter off Table B . l : Hot wire anemometry equipment and specifications. measurements is given i n Table B.2 and illustrated in Figure B . l . The am-bient and operating resistance of this wire were 11.34 fi and was 16.96 fi respectively. The calibration constants were obtained from a curve fitting routine. Appendix B. Hot Wire Anemometry Specification and Calibration 200 A Pressure (Pa) Voltage (V) Reynolds No. Nusselt No. 1.4 3.44 0.63 .571 4.2 3.69 1.09 .664 8.2 3.86 1.52 .732 14.4 4.00 2.02 .789 20.4 4.10 2.40 .832 31.4 4.22 2.98 .884 44.6 4.33 3.55 .932 59.5 4.41 4.10 .971 76.0 4.49 4.64 1.01 95.5 4.57 5.20 1.05 112 4.62 5.63 1.07 145 4.71 6.41 1.12 172 4.77 6.98 1.15 189 4.81 7.32 1.17 Table B.2: Hotwire calibration data for wire No. 1, yielding the calibration constants: 4=0.1575; B=0.4908; and n=0.360. Appendix B. Hot Wire Anemometry Specification and Calibration 201 HOT WIRE CALIBRATION : #1 24.6.88 A-.157516, B».490792, N..360255 1.4 R E Y N O L D S N U M B E R Figure B.l: Hotwire calibration curve for Wire No. 1 Ramb = H-34^/2^ = 11.9511, (used for the bowl-in-piston chamber tests). Appendix C B C Natural Gas Properties This appendix presents calculations of the properties of B C Natural gas, used i n the firing tests. The molecular weight, heating values, viscosity and sto-ichiometric air fuel ratio are calculated from the typical composition given i n Table C . l . The figures given are those used i n the original calculation of constants for the D A T A Q and C R U N C H programs and used i n previous A F L Group studies. This gas composition was compared to monthly averages from B C Hydro over the test period and found to be wi th in 1 %. This led to a 1 % variation i n stoichiometric air fuel value. C O M P O S I T I O N V O L U M E % Methane 94.00 Ethane 3.30 Propane 1.00 Iso-Butane 0.15 N-Butane 0.20 Iso-Pentane 0.02 N-Pentane 0.02 Hexane 0.01 Nitrogen 1.00 Carbon Dioxide 0.01 Water content: 3-4 lbs /mft 3 Table C . l : Composition of B C Natura l Gas. The molecular weight and heating values, shown in Tables C.2 and C.3, 202 Appendix C. BC Natural Gas Properties 203 C O M P O N E N T V o l Fraction M o l . W t . Mass (kg/kmol) Methane CH4 0.940 16.040 15.078 Ethane C2H$ 0.033 30.070 0.992 Propane C$H& 0.010 44.097 0.441 Butane C 4 H 1 0 0.004 58.124 0.232 Nitrogen N2 0.010 28.013 0.280 Carbon Dioxide co2 0.003 44.010 0.132 T O T A L S 1.000 17.156 Table C.2: Molecular weight of B C Natural gas. C O M P O N E N T Mass (%) H H V (kJ /kg) L H V (kJ /kg) Methane CHA 0.879 * 55496 = 48781 * 50010 = 43959 Ethane C2H$ 0.058 * 51875 = 3008 * 47484 = 2754 Propane CzHg 0.026 * 50343 = 1309 * 46353 = 1205 Butane C^Hxo 0.013 * 49500 = 644 * 45714 = 640 Nitrogen N2 0.016 0 0 0 0 Carbon Dioxide C02 0.008 0 0 0 0 T O T A L S 1.000 53742 48558 Table C.3: Higher and Lower Heating values of B C Natura l gas. were determined using a convenient approximation, that is, hydrocarbons of higher order tha C^H^o were included with the butane figures. From the average molecular weight of 17.156 the gas constant and density are determined by: R 8.3143 R MW 17.156 = 0.4846. kJ/kgK P = z*R*T at 21 .1 °C, 101.3 k P a , z « 1 this gives: p = 209 /7 / ( °K) k g / m 3 . The viscosity of the natural gas at 0 °C may be obtained from a similar Appendix C. BC Natural Gas Properties 204 C O M P O N E N T Vol .Frac. Vi M o l . W t . MWi Viscosity Mi ViMW? Methane CHA 0.940 16.040 102.6 386.29 3.765 Ethane C2H6 0.033 30.070 84.8 15.35 0.181 Propane C3H.S 0.014 48.105 75.0 7.28 0.020 Carbon Dioxide C02 0.003 44.01 139.0 2.78 0.020 Nitrogen N2 0.010 28.013 166.0 8.80 0.053 T O T A L S 1.000 420.50 4.116 Table C.4: Viscosity calculations for B C Natura l gas. simplified composition using: 1 _ E Vt * Pi* MWj7 pgas — 1 where t/j is the volume fraction, MW{ is the molecular weight and pi is the viscosity, of component i. In this approximation butane and higher com-ponents are included wi th propane. From the above the viscosity at 0 °C is determined by: 420.5 Vga, = = 102.16/xpoise at 0 °C 4.11b The viscosity at other temperatures is obtained from: pga. = 9.879 * Tf (T + 163.17) The viscosity of Natural gas at 21.1 ° C is 108.96 /zpoise. The stoichiometric air fuel ratio may be determined from the equation for the complete combustion of one mole of B C Natural gas approximated by: .94(7 H4 + . 033C 2 # 6 + MC3H6 + .004C4Hlo+ .OIJV2 + . 003CO 2 + 2.072O 2 + 2.072(3.76)JV2 1.055CO 2 + 2.039ff 2 O + (7.791 + .01)AT2 Appendix C. BC Natural Gas Properties 205 The stoichiometric air fuel ratio is then determined by: A F R = 2.072(1 + 3.76). 28.97  .94(16) + .033(30) + .01(44) + .004(58.1) + 0.01(28) + .003(44) Appendix D Pressure Filtering Methods This appendix presents examples of the filtering method used on the motored and fired pressure profiles. F i l ter ing of the motored pressure signal was performed i n the region of intake valve closing to remove the transient effect caused by the valve motion, detrimental to the temperature calculations. Init ial ly low pass filtering using a 's ixth order butterworth' method was considered. To effectively remove the pressure spikes at low speeds a cut off frequency of 6 K h z was required. The pressure data at this speed was digitalized at 36 k H z . This method however causes a phase shift over the whole range or a step i n the data i f applied to a specific data. Figure D . l illustrates the effect of filtering at 6 k H z on motored pressure obtained at 50.0 rps. Figure D.2 shows the best approximation obtained using a replacement method varying amounts of end tension. This was also deemed unsatisfactory. The preferred method used i n the preceding study involved an averaging and smoothing routine over the 600 points (180 to 60 degrees B T D C ) . Fig-ure D.3 compares the effect of averaging over a 30 point (6 degree) and 60 point (12 degree) windows. The 6 degree window was used over al l speeds for the motored pressure. Similarly an averaging and smoothing method was applied to the fired pressure i n the region 150 to 55 degrees B T D C , using a 12 degree window. 206 Appendix D. Pressure Filtering Methods 207 Figure D.4 shows an expanded view of the pressure profile for the single slot piston at 33.3 rps before and after smoothing. Appendix D. Pressure Filtering Methods 208 2000 1500 1000 500 . NO FILTERING FC 6KHZ ( -150, -60) T -180 ~ l — -90 "T -150 -120 - -60 C R A N K A N G L E D E G R E E S F R O M T D C "1— •30 Figure D . l : Motored pressure trace at 50.0 rps low pass filtered at 6 k H z over the entire range and over an in i t i a l region. Appendix D. Pressure Filtering Methods 209 a CM - * a a D CO CO a a Q a j m S a CO 2000 J 1500 J 1000 J 500 NO SMOOTHING P(END POINTS)*.03 P(END P0INTS)=.02 / / & w -180 -150 -120 -90 1 1--60 -30 C R A N K A N G L E D E G R E E S F R O M T D C Figure D.2: Motored pressure trace at 50.0 rps wi th the region between 150 and 60 degrees B T D C replaced with a section under end tension Appendix D. Pressure Filtering Methods 210 500 AVE 30 POINTS NO END TENSION AVE 60 POINTS NO ENO TENSION NO SMOOTHING OR FILTERING I 1 1 1 -120 -90 . -60 C R A N K A N G L E D E G R E E S F R O M T D C •150 -30 Figure D.3: Motored pressure trace at 50.0 rps with the region between 180 and 60 degrees BTDC averaged and smoothed over 6 and 12 degree windows. Appendix D. Pressure Filtering Methods 211 1 I i i i i 1 -180 -150 -120 -90 -60 -30 0 C R A N K A N G L E D E G R E E S F R O M T D C 400 50 . -180 -150 -120 -90 -60 -30 0 C R A N K A N G L E D E G R E E S F R O M T D C Figure D.4: Expanded fired pressure traces at 33.3 rps, with and without averaging and smoothing applied over 12 degree windows in the region between 150 and 55 degrees BTDC. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0080823/manifest

Comment

Related Items