UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

et-enhanced turbulent combustion Gete, Zenebe 1991

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

Item Metadata

Download

Media
831-UBC_1991_A7 G47.pdf [ 10.98MB ]
Metadata
JSON: 831-1.0098487.json
JSON-LD: 831-1.0098487-ld.json
RDF/XML (Pretty): 831-1.0098487-rdf.xml
RDF/JSON: 831-1.0098487-rdf.json
Turtle: 831-1.0098487-turtle.txt
N-Triples: 831-1.0098487-rdf-ntriples.txt
Original Record: 831-1.0098487-source.json
Full Text
831-1.0098487-fulltext.txt
Citation
831-1.0098487.ris

Full Text

J E T - E N H A N C E D T U R B U L E N T C O M B U S T I O N By ZENEBE GETE B. Sc.(Mech. Engrg.) Addis Ababa University A THESIS S U B M I T T E D IN PARTIAL F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A P P L I E D S C I E N C E in T H E F A C U L T Y O F G R A D U A T E STUDIES M E C H A N I C A L E N G I N E E R I N G We accept this thesis as conforming to the required standard T H E UNIVERSITY O F BRITISH C O L U M B I A April 1991 © ZENEBE GETE, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British 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 K/JPCrl • Enp~r The University of British Columbia Vancouver, Canada Date 2 9 - 0 4 - 9 1 DE-6 (2/88) Abstract A study of the squish-jet design concept in spark ignition engines, with central ignition, was conducted in a constant volume chamber. The effects of jet size, jet number and jet orientation in generating turbulence and jet enhanced turbulent combustion were investigated. Three sets of configurations with three port sizes were used in this study. The research was carried out in three stages: 1. Qualitative information was obtained from flow visualization experiments via schlieren photography at 1000 frames per second. The flow medium was air. A sequence of frames at specific time intervals were selected to study the results from the respective configu-rations and jet sizes. The swirling nature of the flow is vivid in the offset arrangement. 2. Pre-ignition pressure and combustion pressure traces were measured with a piezo-electric pressure transducer from which characterising parameters such as maximum pres-sure, ignition advance and mass burn rate were analysed. Mass fraction curves were cal-culated using the simple model of fractional pressure rise. A maximum pressure increase of 66% over the reference quiescent combustion case, and combustion duration reduction of 77% were obtained for the offset arrangement with 2 mm diameter port. Comparisons of the times required for 10%, 50% and 90% mass burned are identified and confirmed that it took the 2 mm jet the shortest time to burn 90% of the mixture in the chamber. 3. Two-component velocity measurements were made using an L D V system. Measure-ments were taken in the central vertical plane of the chamber at specified locations. The data collected were window ensemble- averaged for the mean and fluctuating velocities over a number of cycles. Data intermittency and low data rate precluded, however, cycle-by-cycle analysis. Mean tangential velocities were calculated for each case and the data i i were used to construct a movie of the tangential velocity as a function of time, suitable for quantitative flow visualization. The vortical nature of the flow was recorded, the distribution being neither solid body rotation nor free vortex, but some complex fluid motion. The jet scale and orientation influence the in generation of turbulence flow field in the chamber, affecting the rate of combustion and the ensuing maximum pressure rise. The offset jet arrangement gives the best results, whereas radially opposed jets have a reduced effect. Increasing the number of jets in opposed arrangement does not enhance turbulent flow. Turbulent flow in the spark region during the onset of ignition was found to be important. in Table of Contents Abstract ii List of Tables vii List of Figures viii Nomenclature xii Acknowledgement xv 1 I N T R O D U C T I O N 1 1.1 Turbulence and Engine Performance 1 1.2 Squish-jets and Turbulent Combustion in Engines 3 1.3 Concepts and Definitions of Turbulent Flow 4 1.4 Research Motivat ion and Objectives 9 2 L I T E R A T U R E R E V I E W 10 2.1 Introduction 10 2.2 Turbulence in S.I. Engines 11 2.3 Turbulent Combustion 14 2.4 Engine Geometry Effects (Squish, Swir l , Jets) 15 2.5 Jet Flow 17 2.6 Flow visualization 18 2.7 L D V Flow F ie ld Measurement 20 iv 3 E X P E R I M E N T A L A P P A R A T U S 24 3.1 Introduction 24 3.2 Apparatus and Charge Medium 24 3.2.1 Combustion Bomb and Piping 25 3.2.2 Charge Medium and Charging 26 3.3 Ignition 28 3.4 Pressure Measurement 29 3.5 Input/Output Board 30 3.6 Schlieren Photographic Arrangement 32 3.7 L D V Experimental Setup 34 4 D A T A ACQUISITION A N D P R O C E S S I N G 38 4.1 Introduction 38 4.2 Data Acquisition Systems 38 4.3 Data Processing 42 4.4 L D V Data Acquistion and Signal Processing 44 5 E X P E R I M E N T A L R E S U L T S A N D DISCUSSION 47 5.1 Introduction 47 5.2 Flow Visualization 47 5.2.1 Two Offset Jet Arrangement 48 5.2.2 Two Radially Opposed Jet Arrangement 49 5.2.3 Four Radially Opposed Jet Arrangement 50 5.3 Jet Enhanced Turbulent Combustion 51 5.3.1 Air Flow Pressure Results 51 5.3.2 Combustion Results and Discussion 52 5.4 Mass Fraction Burned 55 v 5.5 Flow Field Measurement 56 5.6 Experimental Uncertainties 58 5.6.1 Pressure Measurement 58 5.6.2 Flow Measurement Technique 59 6 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 61 6.1 Conclusions 62 6.2 Recommendations 62 Bibliography 64 Appendices 144 A D E R I V A T I O N O F JET F L O W P A R A M E T E R S 144 B PRINCIPLES OF T H E S C H L I E R E N S Y S T E M 147 C M I X T U R E P R E P A R A T I O N A N D C A L C U L A T I O N 149 D G A S C H R O M A T O G R A P H 150 E E Q U I P M E N T SPECIFICATIONS 152 vi List of Tables Table 5-1: Maximum Pressure Rise and Time for Max. Pressure 71 Table 5-2: Times for 10%, 50% and 90% Mass Burn 72 vii List of Figures Fig. 1-1: The Squish-jet Design, After Evans 73 Fig.2-1: Growth of Jet Profiles 74 Fig.2-2:Jet Profiles, After Trupel and Forthmann,From Abramovich's "The Theory of Turbulent Jets" 75 Fig.2-3: Geometry of a Typical Measurement Volume 76 Fig.2-4: Photograph of the Chamber Assembly 77 Fig.2-5: Optical Arrangement of the L D V System 78 Fig.3-1: Schematic of the Combustion Bomb, Sectional View 79 Fig.3-2: Photographs and Section Views of Ports 80 Fig.3-3: Control and Instrumentation Diagram of the Bomb 81 Fig.3-4: Calibration Curve for Propane Composition Analysis 82 Fig.3-5: Schematic of Ignition System Electrical Wiring 83 Fig.3-6: Photograph of Laser and Accessories 84 Fig.3-7: Block Diagram of the DT2811 I/O Board 85 Fig.3-8: Schematic of the Schlieren Optical Bench 86 Fig.3-9: Doppler Signal From the Photomultiplier Before and After Filtering . 87 Fig.3-10: Section View of the Cyclone Particle Seeder 88 Fig.4-1: Basic Circuit Diagram of the Dual Mode Charge Amplifier 89 Fig.4-2: L D V Measurement Locations in the Central Plane of the Chamber . . 90 Fig.5-1: Schlieren Photographic Sequence; 2-Offset Jets, Dia. = 2 mm 91 Fig.5-2: Schlieren Photographic Sequence; 2-Offset Jets, Dia. = 3 mm 92 viii Fig.5-3: Schlieren Photographic Sequence; 2-0ffset Jets, Dia. = 4 mm 93 Fig.5-4: Schlieren Photographic Sequence; 2-Radial Jets, Dia. = 2 mm 94 Fig.5-5: Schlieren Photographic Sequence; 2-Radial Jets, Dia. = 3 mm 95 Fig.5-6: Schlieren Photographic Sequence; 2-Radial Jets, Dia. = 4 mm 96 Fig.5-7: Schlieren Photographic Sequence; 4-Radial Jets, Dia. = 2 mm 97 Fig.5-8: Schlieren Photographic Sequence; 4-Radial Jets, Dia. = 3 mm 98 Fig.5-9: Schlieren Photographic Sequence; 4-Radial Jets, Dia. = 4 mm 99 Fig.5-10: Comparison of Charge Pressure Development: 2 Offset Jets 100 Fig.5-11: Comparison of Charge Pressure Development: 2 Radial Jets 101 Fig.5-12: Comparison of Charge Pressure Development: 4 Radial Jets 102 Fig.5-13: Cyclic Variation Of Combustion Pressure 2 Offset Jets, Dia. = 2 mm 103 Fig.5-14: Cyclic Variation Of Combustion Pressure 2 Offset Jets, Dia. = 3 mm 104 Fig.5-15: Cyclic Variation Of Combustion Pressure 2 Offset Jets, Dia. = 4 mm 105 Fig.5-16: Cyclic Variation Of Combustion Pressure 2 Radial Jets, Dia.= 2 mm 106 Fig.5-17: Cyclic Variation Of Combustion Pressure 2 Radial Jets, Dia. — 3 mm 107 Fig.5-18: Cyclic Variation Of Combustion Pressure 2 Radial Jets, Dia. = 4 mm 108 Fig.5-19: Cyclic Variation Of Combustion Pressure 4 Radial Jets, Dia. = 2 mm 109 Fig.5-20: Cyclic Variation Of Combustion Pressure 4 Radial Jets, Dia. = 3 mm 110 Fig.5-21: Cyclic Variation Of Combustion Pressure. 4 Radial Jets, Dia. = 4 mm 111 Fig.5-22: Combustion Pressure Rise: 2 Offset Jets 112 Fig.5-23: Combustion Pressure Rise: 2 Radial Jets 113 Fig.5-24: Combustion Pressure Rise: 4 Radial Jets 114 Fig.5-25: Comparison Of Combustion Pressure Rise Dia. = 2 mm 115 Fig.5-26: Comparison Of Combustion Pressure Rise Dia. = 3 mm 116 Fig.5-27: Comparison of Combustion Pressure Rise Dia. = 4 mm 117 Fig.5-28: Mass Fraction Burned : 2 Offset Jets 118 i x Fig.5-29: Mass Fraction Burned : 2 Radial Jets . 119 Fig.5-30: Mass Fraction Burned : 4 Radial Jets . . . 120 Fig.5-31: Comparison of M . F . B . For Different Configurations, Dia. = 2 mm . 121 Fig.5-32: Comparison of M . F . B . For Different Configurations, Dia. = 3 mm 122 Fig.5-33: Comparison of M . F . B . For Different Configurations, Dia. = 4 mm . 123 Fig.5-34: U-Component Velocity on Axis of Jet, Offset Configuration, Dia. = 2 mm 124 Fig.5-35: V-Component Velocity on Axis of Jet, Offset Configuration, Dia. = 2 mm 125 Fig.5-36: U-Component Velocity on Axis of Chamber, Offset Configuration, Dia. = 2 mm 126 Fig.5-37: V-Component Velocity on Axis of Chamber, Offset Configuration, Dia. = 2 mm 127 Fig.5-38: U-Component Velocity on Axis of Jet, Offset Configuration, Dia. = 4 mm 128 Fig.5-39: V-Component Velocity on Axis of Jet, Offset Configuration, Dia. = 4 mm 129 Fig.5-40: U-Component Velocity on Axis of Chamber, Offset Configuration, Dia. = 4 mm 130 Fig.5-41: V-Component Velocity on Axis of Chamber, Offset Configuration, Dia. = 4 mm 131 Fig.5-42: U-Component Velocity on Axis of Jet, 2 Radial Configuration, Dia. = 2 mm 132 Fig.5-43: V-Component Velocity on Axis of Jet, 2 Radial Configuration, Dia. = 2 mm 133 x Fig.5-44: U-Component Velocity on Axis of Chamber, 2 Radial Configuration, Dia. = 2 mm 134 Fig.5-45: V-Component Velocity on Axis of Chamber, 2 Radial Configuration, Dia. = 2 mm 135 Fig.5-46: U-Component Velocity on Axis of Jet, 4 Radial Configuration, Dia. = 2 mm 136 Fig.5-47: V-Component Velocity on Axis of Jet, 4 Radial Configuration, Dia. = 2 mm 137 Fig.5-48: U-Component Velocity on Axis of Chamber, 4 Radial Configuration, Dia. = 2 mm 138 Fig.5-49: V-Component Velocity on Axis of Chamber, 4 Radial Configuration, Dia. = 2 mm 139 Fig.5-50: Distribution of Tangential Velocity, Offset Configuration, Dia. = 2 mm 140 Fig.5-51: Distribution of Tangential Velocity, Offset Configuration, Dia. = 4 mml41 Fig.5-52: Distribution of Tangential Velocity, 2 Radial Configuration, Dia. = 2 mm 142 Fig.5-53: Distribution of Tangential Velocity, 4 Radial Configuration, Dia. = 2 mm 143 xi Nomenclature L Length Scale u Turbulent Intensity (m/s) V Kinematic Viscosity (m2/s) T Time Period (s) U(t) Instantaneous Velocity (m/s) u Time-averaged Mean Velocity (m/s) u Fluctuating Velocity (m/s) R(r) Temporal Autocorrelation Coefficient Rx Spatial Autocorrelation Coefficient Lx Integral Length Scale U Integral time Scale \ Temporal Taylor Microscale Length Taylor Microscale e Dissipation Rate UT Turbulent Flame Speed (m/s) UL Laminar Flame Speed (m/s) AT Turbulent Flame Front Area (m2) AL Laminar Flame Front Area (m2) n Index of Refraction, Counter P Density (kg/m3) k Constant Fringe Spacing (fim) xii A Wave Length K Half Angle lm Measurement Volume Major Axis dm Measurement Volume Minor Axis de-2 Beam Diameter De-2 Diameter of Beam From Laser F Focal Length of Focussing Lens f Doppler Frequency NfT Number of Fringes M . V . Measurement Volume V Volt Poar Pressure (bar) P(t) Pressure Trace (bar) Nc Total Number of Cycles t Time (s) P(t,n) Pressure At Time t and Cycle n MB(t) Mass Fraction Burned Pmax Maximum Pressure (bar) Pi Pre-ignition Pressure UE(t) Ensemble-averaged Mean Velocity (m/s) U(t,i) Instantaneous Velocity at Time t and ith Cycle (m/s) UEA(1) Window Ensemble-averaged Velocity (m/s) UF,EA{i) Window Ensemble-averaged Fluctuating Velocity (m/s) t Time at Centre of Window (ms) At Size of Window (ms) xm Nt Total Number of measurements N{ Number of Velocity Measurements in a Window in the ith cycle j xiv Acknowledgement I would like to express my gratitude to Dr. R.L. Evans for his supervision and encour-agement throughout the course of this study. I wish to thank him for his concern and advice during my stay at U.B.C. Drs. M . Iqbal and S. Green are also thanked for their contribution in the defence. I would like also to thank the academic and technical staff of the Mechanical Engi-neering Department especially: Professor P.G. Hill, for his encouragement and follow up; L.Drakes, for his efficiency in machining and building the experimental apparatus and for his untiring assistance in maintaining, modifying, and manufacturing accessories when time was so demanding; D. Baysuth, for his assistance in installing the electronics of the system. In addition I would like to express my appreciation for my friends whose encourage-ment, discussions and comments at the various stages of the work were very helpful. Finally I wish to thank CIDA for the financial support given to me during my course of study at U.B.C. xv Chapter 1 I N T R O D U C T I O N 1.1 Turbulence and Engine Performance Turbulence is believed to play a key role in enhancing the combustion process in internal combustion (i.c) engines. Large and small scale mixture motions exist affecting large parts of the chamber volume, the mixing process and speed of flame propagation.The mechanism by which such a turbulent flow is brought about deserves due attention in the design and improvement of engine combustion chambers. Much progress has been achieved in improving engine design and performance throughout the history and devel-opment of automotive engineering. Many researchers are still striving to achieve better engine designs and performance along with improved fuel economy. Advancements in spark ignition (s.i.) engines require a thorough understanding of the fundamental processes taking place in combustion chambers. Fluid flow, thermodynamic processes, and chemical reactions, all prevailing simultaneously, are the fundamental pro-cesses that must be dealt with in engines. Any research improvement in engines centres around the knowledge of the aforementioned fields of study . Improvements in the com-bustion chamber that lead to spark ignition engine optimization enable engine designers to design more efficient vehicles. The results would be dividends in fuel economy, and much smoother drivability. Contemporary researchers are giving emphasis to fast-burn mechanisms and lean-burn operations in s.i. engines. Fast burning is achieved by generating turbulent flow 1 Chapter 1. INTRODUCTION 2 just before ignition is activated. Once a fast burning mechanism is maintained, lean-burn operation may then be associated with it by controlling the equivalence ratio. The enhancement of combustion rate leads to improved engine performance, better lean op-eration resulting in a reduction of NOx formation, better fuel economy, and reduced environmental pollution. Power in a combustion chamber is generated by charge burn which gives rise to pressure development and hence torque output to drive the vehicle. The charge medium (fuel air mixture) could be lean, stoichiometric, or rich. Lean mixtures burn more slowly promoting the disadvantages of cycle-to-cycle variations of pressure, leading to vehicle surge (which is undesirable). Due to the associated misfiring around the lean limit, power is reduced, and the amount of unburned hydrocarbons increases. To improve the shortcomings of a lean mixture operation, provisions must be made to facilitate fast burning. This will increase efficiency because the combustion process approximates constant volume combustion, approaching the ideal Otto cycle with greater indicated work. However, fast-burn by itself has the disadvantage of increasing NOx emission due to the high combustion temperature associated with it. To exploit the advantages of fast-lean burn operations, and compensate for the disad-vantages, it is desirable to couple both operations in the combustion processes of engines. This may be achieved by optimizing mixture motion in the combustion chamber, by in-creasing flame speed through increased turbulence, optimizing the ignition system and geometry of the engine, by selecting the number and position of sparks for shorter flame travel, and improving the reaction kinetics taking place in the chamber. The fundamental problems in thermodynamics, fluid mechanics and chemistry make the engine combustion process difficult to understand. Analytical solutions to problems of turbulence and combustion in the engine environment are very complicated due to a lack of knowledge of turbulence and combustion processes. Therefore most studies resort Chapter 1. INTRODUCTION 3 to experimental analysis and empirical equations. Studies are conducted in engines and engine like chambers ("bombs") to get an under-standing of the nature of turbulence and turbulent combustion. Researchers on engines and bombs have shown the influence of higher levels of turbulence on turbulent flame speed and burn rate.Yet there are still questions to be addressed when when turbulence is important in flame development, the level of turbulence required, and the influence of the flow field on combustion . Many engine variables come into play when one considers combustion in s.i. engines. Investigators have sought explanations of the effects of these variables; such as spark number [1], ignition timing, energy and duration of spark [2, 3], air-fuel ratio,fuel com-position and method of delivery into the engine[4, 5],temperature,mass of the charge per cycle, compression ratio, exhaust gas dilution of the charge, turbulence and mixture motion as a function of engine speed, intake charge motion and combustion chamber design. 1.2 Squish-jets and Turbulent Combustion in Engines The idea of a squish-jet piston design that can potentially generate controlled turbulence in the s.i.engine combustion process was proposed and patented by Evans[6] in 1986.The purpose of the squish-jet design is to enhance turbulence in s.i.engines just prior to igni-tion when it is most desired.The squish-jet design shown in Fig.1-1 changes the standard bowl-in-piston design such that channels are used to communicate between the squish area and the bowl of the combustion chamber.Jets are to be generated in the late stages of the compression stroke and change the flow field promoting a fast combustion rate af-ter the subsequent ignition of the charge. The increase in combustion rate is manifested in an increase of pressure, a very important parameter provided it takes place within a Chapter 1. INTRODUCTION 4 very short duration of time after ignition when the piston is at TDC.The shorter the combustion period, the higher the chemical energy released from the mixture within that time period, and the higher the torque-producing pressure when it is most needed. Studies related to the squish-jet design have been conducted in the Department of Mechanical Engineering, the University of British Columbia under normal operating conditions in a test engine. A review of these studies and other related works will be presented in chapter two. The present study was conducted based on the idea of turbulence generation by jets in a combustion bomb assumed to have similar effects to the squish jets in an actual engine without the inclusion of other engine variables. 1.3 Concepts and Definitions of Turbulent Flow-Eddy formation on obstacles, shear flow due to the retarding effect of walls, and the mixing of streams with different velocities result in turbulent motion [26]. Turbulence is a randomly fluctuating motion of small masses of fluid superimposed on any mean flow.The fluctuating velocities of turbulent flow create additional stresses, which are usually much larger than, the viscous stresses created by molecular motions [27]. These stresses are usually known as Reynolds stresses. Reynolds stresses arise principally from larger eddies which extract energy from the mean flow. Large eddies that are most effective in extracting energy from the mean flow are those with vortices whose principal axes are roughly aligned with the principal axis of positive rate of strain [28]. Energy transfer is believed to be associated with the vortex stretching process [29], which gives rise to an increase in the kinetic energy of rotation, as a result of conservation of angular momentum. Energy is communicated to the turbulent motion from the mean flow through the Chapter 1. INTRODUCTION 5 largest eddies and then to the smaller and smaller eddies and finally dissipated by vis-cous friction. Turbulent energy is fed into a flowing medium by virtue of the straining mechanism that takes place against the Reynolds stresses. For a steady state flow, uni-form conditions, the rate of turbulent production is equal to the rate of energy dissipation. Decay times depend on the balance of turbulent generation and the prevailing dissipation mechanisms. For a fluctuating motion, eddy life time may be estimated by the ratio of the respective length scale to the fluctuating velocity scale, L/u . The rate of turbulent energy production is proportional to ^ - [9], and the time scale for molecular dissipation if the length scale were the same would be The rate of viscous dissipation of energy would be proportional to ~j- [9], where L is length scale, u is fluctuating velocity scale, and v is the kinematic viscosity. Large eddy sizes with slow velocities decay slowly than eddies with smaller length scales and higher velocities. Irregularity and randomness of turbulent flows make statistical treatment appropriate for the description and analysis of such flows. It has been established that turbulent flow contains eddies with different sizes and velocities distributed over a wide range. This variation includes both spatial and temporal differences in scale. Therefore most actual turbulent flows are non-homogeneous and nonisotropic. Consequently the energy distribution will also be variable. There exists a spectrum of turbulent energy distribution over the ranges of the fluctuating frequencies where the power spectral density function is used to determine this energy distribution. Turbulent transport of mass and momentum are several order of magnitudes larger than molecular transport for the reason that the mass of eddies and the average distances they travel are much larger than the molecular mass and the molecular mean-free-path. This fact underlies the importance of turbulent flow phenomena for mass, momentum and heat transfer, mixing and flame propagation processes. Quantities used for turbulent description are: Chapter 1. INTRODUCTION 6 Mean velocity: the average velocity measured during a specified time interval, i.e. a period T, with both direction and magnitude expressed by U= \ [ U(t)dt. (1.1) In engines where the instantaneous velocity U(t) is constantly changing, the mean velocity will also be changing, and it depends on the time interval in which averaging is performed. The unsteady nature of the flow hence makes a single mean velocity in the engine environment unrealistic and questionable. A mean velocity makes sense, for analysis purpose, when a quasi-steady flow is assumed within a small crank angle or time window. The instantaneous velocity of a fluid particle U(t) is then customarily written as the sum of the time average velocity U and a fluctuating velocity component u. U(t) = U + u (1.2) Turbulent intensity: the velocity component given by the root-mean-square value of the fluctuating velocity characterizes turbulent intensity. u = v^? (1.3) Turbulent scale: turbulent scales are represntative sizes of turbulent eddies (a certain scale in space) and representative times (a certain scale in time) for the corresponding eddies in the flow field under study [7, 9]. It is essential to specify representative scales in a turbulent flow so that quantitative analysis may be possible. In fact several length and time scales exist in a turbulent flow with specific roles in the description and analysis of the flow. These wide ranges of turbulent flow scales are bounded by the dimensions of the system boundaries and the diffusive action of molecular viscosity of the medium. Chapter 1. INTRODUCTION 7 By making use of the autocorrelation of the velocity fluctuation at a point, a char-acterstic scale of the eddy size distribution within a turbulent flow can be found. The autocorrelation coefficient is defined as: j*-) = i/r f 5^^+11* (i.4) Jo i r where t is time, T is the period of measurement, and r is the autocorrelation time. From the autocorrelation coefficient, the integral time scale of turbulence, Lt is defined as: Lt = I* R{T)<LT (1.5) Jo From the temporal scale Taylor[30] showed the spatial scale of turbulence as: Lx = LtU (1.6) when U is constant and U ^ > u . Lx can be thought of as a characterstic "mixing length" of the turbulent flow. Besides the time correlation, another way of finding a characteristic length scale in a turbulent flow is the spatial correlation. In this method of computation, the turbulent fluctuating components at a correlation distance are utilized to find the autocorrelation coefficient. Spatial autocorrelation is defined as: R. = "(«oM«o + r) = ( L ? ) V(«(x0)2)(«(*o + r)2) where u(x0) is turbulent velocity at x0, and u(x0-\-r) is turbulent velocity at correlation distance r.Then the length scale of turbulence is defined by: (•OO = / Rxdx (1.8) Jo Lx may be regarded as a length scale proportional to the average size of the fluid elements or eddies in the flow. Chapter 1. INTRODUCTION 8 Taylor[30] found a temporal microscale from the temporal autocorrelation in the following form. He used this temporal microscale to calculate the spatial microscale; Xx = U\t According to Tennekes [9] the Taylor microscale can also be given in the following form. <*>'SATV ( 1 1 0 ) where u is the fluctuating component of velocity. The Taylor microscale is related to the macroscale L, by ^- oc ^r. A turbulent Reynolds number R\, based on the microscale A, was introduced by Taylor, where it was shown to be proportional to the ratio of time scales for large to small eddies in addition to being the ratio of inertial to viscous forces; i.e. R\ oc If R\ increases, the time scale of large eddies decreases, but the time scale of small eddies decreases even more, Andrews et al. [59]. In the micro structure regime of turbulence, isotropy is discerned [29], where the dissipative eddies make their detailed structure independent of the large eddies. Kolmogorov analysed the isotropic regime of turbulent flow and derived microscale values of length t] , velocity vn , and time TV , based on the rate of dissipation and kinematic viscosity only. These are given by: 7 7 = ( ^ ) 1 / 4 = A 1 5 - l / 4 ( i 2 A ) - l / 2 ( L n ) e u = M , / 4 ="/-» = IS"" (1.12) r , = ( ^ = , / , , = _ A _ ( M S ) where u is the kinematic viscosity and e is the dissipation rate defined [9] for isotropic turbulence by: Chapter 1. INTRODUCTION 9 In general the state of structures in a turbulent flow is believed to be responsible for facilitating the combustion process either by increasing the flame area through distortion, or by increasing burning rate by means of preferential combustion within the isotropic dissipating eddies. 1.4 Research Motivation and Objectives The presence of several engine variables in an actual engine operation make it difficult to single out the effect of individual parameters. At times it is important to study the effects of each variable on turbulence generation and the ensuing combustion process without the influence of one or another factor. This study concentrated on the influence of jets on the enhancement of turbulence and turbulent combustion in an engine-like chamber; a constant volume bomb. The objective of this study was to investigate the influence of jet scale(size), jet orientation and jet number on turbulence generation in a constant volume chamber with the aim of achieving fast-lean combustion. By assessing the conditions for fast-lean operation, the ultimate goal was to gain information on whether the squish-jet design would be feasible or not in the actual engine chamber design. Flow field measurements were performed so that jet velocities could be investigated for their effect on combustion. Qualitative information on the flow field was acquired from flow visualisation (schlieren photography), and quantitative velocity measurements were made using Laser Doppler Velocimetry (LDV). Chapter 2 L I T E R A T U R E R E V I E W 2.1 Introduction The study of engine design and geometry with the aim of improved performance and fuel economy is not a new field of research. Although research in this discipline is as old as engines and vehicles, increasingly many researchers are investigating engine combustion phenomena, the nature and processes of flow field in engine combustion chambers, and the interactions of turbulence with combustion. The need to meet new environmental regulations, the desire to improve engine performance , and recurring oil crises give the impetus for the continuation and progress of engine studies. Yet, the nature of engine turbulence and combustion are not fully understood due to the complexity of the problems involved. Previous related works are assessed in the following sections. Studies in turbulent flow, means of turbulence generation, combustion in s.i. engines, influence of engine geometry on turbulence, effects of turbulence on combustion and combustion on turbulence are reviewed. Attention is also given to the studies on squish, squish-jet and jet effects on turbulence either in engines or engine-like environments. To characterize a turbulent flow, one needs to acquire information about the flow either from numerical simulations or experimental results. Two experimental techniques are well established to measure the flow field; Hot Wire Anemometry and Laser Doppler Anemometry. Means of flow field measurements used in previous investigations pertaining to turbulence and turbulent 10 Chapter 2. LITERATURE REVIEW 11 combustion are thus discussed. 2.2 Turbulence in S.I. Engines Turbulence may be described as the fluctuating velocity component superimposed on the mean velocity of a viscous fluid flow. Hinze[7] characterized turbulent fluid motion as "an irregular condition of flow in which the various quantities show a random variation with time and space coordinates, so that statistically distinct average values can be discerned". According to Schlichting[8] the most important characteristic of a turbulent flow is that at any point within the flow, the velocity and pressure are not constant in time but exhibit irregularity, and high frequency fluctuations. Tennekes[9] listed some of the characterstics of turbulent flows as: irregularity, diffusivity, large Reynolds numbers, three dimensional vorticity fluctuations and dissipation. Because of irregularity, the motion of turbulent flow in all details cannot be described as a function of time and space. Hence statistical methods have been used to solve problems involving turbulence. Engine combustion processes are greatly influenced by turbulence. The general mo-tion within the cylinder and the associated turbulence specifically during ignition, flame development and propagation affect the burning rate, cycle-to-cycle variations and the lean combustion limit. Turbulence generation in s.i. engines is affected by the shear flow in the inlet mech-anism. Semenov[10] using constant temperature Hot Wire Anemometry (HWA) found that the intake process in a research engine generated the velocity gradient which was responsible for the rise of turbulence. The jet nature of the flow through the intake valve was characterized by the spatial velocity gradients during the intake stroke creating the in-chamber turbulent flow regime in the subsequent process. The influence of the intake process and geometry in generating turbulence was later Chapter 2. LITERATURE REVIEW 12 confirmed in the works of Winsor and Patterson[15], Lancaster[ll], and Fuller and Daneshyar[12] by comparing the effects of shrouded and nonshrouded inlet valves. The shrouded valve geometry increased the average velocities and turbulent intensities more than the nonshrouded valve. Swirling inlet geometry also has significant effect in gen-erating in-chamber flow field.Whitelaw et al[13] concluded that in the absence of swirl and squish, the intake generated mean motion and turbulence decayed considerably by the end of compression. Heywood[14] on the other hand stated that intake generated swirl usually persists through the compression, combustion and even the expansion pro-cess. This phenomenon is due to the effect of the angular momentum imparted on the incoming mixture before admission to the cylinder. Another engine variable involved in creating a turbulent atmosphere prior to the occurrence of ignition is engine speed. Lancaster, Heywood, and Saxenafll, 14, 16] found that turbulence intensity and mean velocity increased with engine speed, and had a greater effect than inlet generated flow in enhancing turbulence during the ignition process. Bracco et al.[17] indicated that the T .D.C. turbulence intensities were relatively insensitive to the intake velocity, and tended to scale more strongly with engine speed. On the other hand, the ensemble averaged velocity was also found to increase linearly with intake velocity for a fixed engine speed. It can therefore be stated that engine geometry, configuration, and speed have important roles in generating turbulence in the combustion chamber. Besides engine inlet geometry and engine speed, other engine variables are also in-volved in the physical and chemical processes taking place in the engine cylinder flows and combustion. Compression ratio, equivalence ratio and mixture composition, and spark location are variables that are salient to combustion chamber processes. However, in the engine environment, it is difficult to study the quantitative effects of individual engine variables. Measurement techniques in engines, the cyclic nature of engines and Chapter 2. LITERATURE REVIEW 13 the cycle-to-cycle variations of quantities make investigations far more difficult. For these reasons studies have been conducted based on simplified models and empirical equations. Engine turbulence measurements have been conducted using HWA [10, 25, 42, 44]. Since HWA is difficult to use in fired operation in engines (due to hostility of the envirn-ment), motored operation measurements and results are widely available in the literature. Valuable studies were conducted, important conclusions drawn from these works despite the disadvantages of using HWA (interference and breakage problems). Recently, Laser Doppler Velocimetry (LDV) has gained popularity in engine flow field research. L D V has been used to investigate flow field phenomena inside an engine cylinder: to study the effects of piston bowl geometry and intake swirl ratio on the structure and evolution of turbulent flow field [18], to measure air velocity in motoring i.e. engines [19], to make cycle-resolved velocity and turbulence measurements both in motoring and firing conditions [17], and to investigate combustion phenomena in s.i. engines [20]. It has also been used to measure the swirl velocity in the combustion bowl of a DI diesel engine [21], to investigate the relationship between fluctuations of the flow field in the region of the spark plug and the efficiency of the subsequent combustion [22], to measure lateral integral scales of the tangential velocity component in a motored i.e. engine [23], and to measure real-time gas flow velocity in the combustion chamber of an engine during firing operation [24]. Direct scale measurement was also made possible by using L D V . Rask[19], Fraser et al.[23], and Witze[25] surveyed and compared HWA and L D V . Comparison of the two showed the choice of L D V measurement technique was more favorable than HWA provided the particle seeding problem was carefully handled. Chapter 2. LITERATURE REVIEW 14 2.3 Turbulent Combustion It is well established that the combustion process in s.i. engines is turbulent. The influ-ence of turbulence on combustion has been qualitatively recognized long ago. However, accurate analytical solutions have not yet evolved from turbulent combustion modelling. A higher level of turbulence increases turbulent flame speed and burn rate. Mallard and Le Chatelier[31] in 1883 showed that turbulence increased both energy transfer and flame surface area. The effect of turbulence on combustion flame speed as investigated by Damkohler[32] for small-scale, high intensity turbulence, for eddies of scale less than the thickness of the laminar flame front was given by; where UT and UL are the turbulent and laminar flame speeds respectively, e is the total turbulent diffusivity, and v is the kinematic viscosity of the fluid. He assumed that the effects of these eddies was to increase the local heat and mass transfer along the flame front, where the local rate of flame propagation is increasing. His expression was based on the transport properties in the unburned gas ahead of the flame. Another mechanism used to explain the increase in flame speed due to turbulence considers the effects of turbulent eddies of a scale larger than the thickness of the flame front. For large turbulent eddies and low intensity turbulence, it was proposed that these eddies have no effect on the local flame velocity; rather simply wrinkle the flame front and hence increase the area [32, 33, 34]. This was expressed by: JTL-TL ( 2 1 5 ) where AT and Ar, are turbulent and laminar flame front areas respectively. Damkohler proposed the area ratio in terms of the turbulent intensity u and laminar Chapter 2. LITERATURE REVIEW 15 flame speed (/^ as: Shelkin[35] considered large-scale, small intensity turbulence and assumed that flame surfaces are distorted into cones with characterstic dimensions which depended on the turbulent intensity and the turbulent scale. He gave the expression | = [1 + ( ^ ) 2 ] 1 / 2 (2-17) showing a linear dependence of turbulent flame speed on intensity for u 3> UL. Similar conclusions were given by [11, 33]. Many experiments verified the approximately linear correlation of flame speed ratio with intensity. Results were also obtained indicating that turbulent flame speed increased linearly with the engine speed, upon which the turbulent intensity may be scaled [11, 15, 17, 36]. In spark ignition, the influence of turbulence during flame kernel formation is low as compared with that of the main combustion period, [37]. But there exists a spectrum of different opinions concerning the structure and propagation process of flame kernels in a turbulent medium [38, 39, 40, 59, 66]. 2.4 Engine Geometry Effects (Squish, Swirl, Jets) Most phenomena taking place in spark ignition engines centre on turbulence. The genera-tion and breakdown of turbulence at T .D.C. in s.i. engines is very essential to improve the performance of engines. Among other variables, engine geometry has attracted consid-erable attention as one of the means of generating and enhancing turbulence; and hence has been used as an important design parameter in the development and improvement of i.e. engines. It has been experimentally proven that a geometry design change leading to the generation of turbulence promotes a fast combustion rate, a desirable effect if this Chapter 2. LITERATURE REVIEW 16 phenomenon happens during ignition onset. However, the turbulence generated at the beginning of ignition should not be so violent as to convect the flame kernel, leading to quenching. Design of the combustion chamber could combine the provisions of swirl, squish and jets independently or in combination. Swirl in engines or engine like chambers was investigated and found to be very influential in generating turbulence, and hence fast combustion [41]. In spark ignition engines, inlet valve geometry greatly influences the flow field in the combustion chamber. It can produce swirl or jet flow into the cylinder. By changing the inlet valve geometry, [14, 17, 59] it was shown that turbulence generated by shrouded valves was greater than nonshrouded valves. Various researchers have studied the effect of squish and squish-jets on combustion. Whitelaw et al [13] concluded that in the absence of swirl and squish, the intake generated mean motion and turbulence decayed considerably by the end of the compression stroke. Based on Evans' [6] invention of the squish-jet arrangement, Cameron[42] investigated turbulence in a single cylinder engine using Hot Wire Anemometry and reported a 50% increase in the turbulence intensity with reference to a standard bowl-in-piston design. An increase in the mass burn rate was also indicated. But in a similar work, Dymala-Dolesky [43] observed that the squish-jet design was not very effective in promoting fast combustion due to reduction of squish motion and flow activity near the spark plug. The strength of the jets was reported to be rather weak, and incapable of reaching the cylinder centre. However, Dymala-Dolesky's work confirmed the importance of squish in increasing the mean and turbulence fluctuation. In this particular study the effect of squish was more pronounced than that of the jets. According to Tippett[44] higher mean velocity and turbulent intensity were obtained with the increased squish area of the Chapter 2. LITERATURE REVIEW 17 bowl-in-piston and squish jet pistons. Tippett's study also agreed with that of Dymala-Dolesky's on the squish-jet action in the lower portion of the bowl piston. In both cases, however, the effective squish-jet size and strength had not been optimized. Cole and Mersky[45], in their study of a combustion bomb chamber investigated how controlled scale (jet size) affected the pressure rise process during combustion. They reported that the rate of pressure rise increased approximately linearly with mixture jet velocity. It was found as well that the width of mixture jet had effect on the rate of pres-sure rise. A research in a two chamber combustion bomb by Klomp and Deboy[46] showed that increasing jet strength reduces mixture burn time. This phenomenon was attributed to increased turbulence, mixing, and jet penetration with increased jet strength. Nakamura et al[47] worked on a gasoline engine with a design of what they called an MCA-jet . A jet valve directed to the spark plug with different nozzle sizes was used in this study to see the effect of jet sizes. It was shown that jet size influences the rate of combustion. In this study the optimum jet diameter was found to be 6 mm; compared with the 4 mm and 8 mm diameter jets. Weak jets and excessively strong jets were not desirable for the mixing and combustion process. From all previous works it is indicative that a change in engine geometry can improve the combustion process, and there is still room to change engine geometrical design to accommodate enhanced turbulent combustion. 2.5 Jet Flow In general a jet may be assumed to issue from a circular orifice into a stream of uniform velocity or into a quiescent fluid, with a uniform outlet velocity. For turbulent flow of high Reynolds number, the flow becomes fully developed shortly after the orifice outlet. A characteristic feature of a turbulent jet is the smallness of the transverse velocity Chapter 2. LITERATURE REVIEW 18 component in any section of the jet compared with the longitudinal velocity component [48]. Hence the usual way of dealing with jets is to disregard the transverse velocity components and work with the longitudinal and fluctuating velocity components. As confirmed by experiments, the velocity profile of jets broadens continuously along the axis. This phenomenon appears as shown in Fig. 2-1. Further from the outlet, the velocity profile becomes lower and wider. Velocity profiles obtained from experiments of Trupel and Forthmann are shown by Abramovich [48] in Fig.2-2. It is also shown that the dimensionless turbulent mixing lengths for different sections of the jet stream remain constant. Abramovich established a law for the growth of the jet as a function of distance along the longitudinal axis. Schlichting [8] has also given the same kind of relationship for jet growth. From this derivation one is able to get the jet width at any section along the axis. Details of the derivation are given in Appendix A. 2.6 Flow visualization Visualization or observation of a fluid flow is essential to the process of understanding the flow phenomenon that may take place in engine combustion chambers under differ-ent circumstances or operating conditions. The operating medium could be a colourless, transparent and non-luminous substance where it is difficult to use direct visual or photo-graphic methods to look at an event. In addition the phenomena to be studied may vary with time, therefore high speed photographic techniques have to be applied to capture sequential flow events. If qualitative information is desired for a given system of colourless fluid flow, where analytical solutions could be very complex, it is of prime importance to devise a method of observing and recording the flow condition through the intended period of time. A technique that is appropriate for such applications or investigations is the schlieren Chapter 2. LITERATURE REVIEW 19 photography method. In this technique, the phenomena of interest involves changes of the refractive index across the field to be visualised. The refractive index may vary across a field because of changes in density of the substance due to changes in temperature and pressure. The schlieren is a device which is used to determine small variations in the refractive index of a transparent material. There is a linear relationship between the refractive index and density of a substance [62]. This relationship is given as: ra = 1 + kp where, n is index of refraction of the substance, and p is the density of the substance k is a constant for a particular gas and particular wavelength of light and can be expressed by; k = (n0 - l)/p0 na and p0 are values at a reference temperature and pressure. The working principle of the schlieren system is described in Appendix B. Various investigators applied the schlieren system in engines and similar machines to study flame formation, growth and development under different conditions [14, 37, 49, 50, 51, 52]. The schlieren system components may vary from investigator to investigator as per application, desire for accuracy of observation, availability and design of setup, but the basic principle remains the same. In general the system may be comprised of at least a light source, a collimator, a focussing mirror or lens, a knife edge or pinhole, or graded filter, and a high speed movie camera for event recording. When considering a light source, appropriate care must be taken in order to account for important properties of the source. These properties are source dimension, brightness, duration and spectral properties. The quality of the optical arrangement that may be Chapter 2. LITERATURE REVIEW 20 used must also be as high as possible so that it may be free from spherical and chromatic abberations. The use of a knife edge or a circular cut-off is crucial in a schlieren system, but is also subject to diffraction which is a natural phenomenon occurring when light contacts an object.The more the interference of the knife edge with the on coming beam, the higher the diffraction on the image plane precluding sharp boundaries of the image. To this effect a compromise must be made between image sharpness and knife edge or pinhole usage. When experimenting with transient flow conditions, it is very important to use a high framing rate camera to obtain a clear resolution of the fluid motion [62] . Short exposure time is required so that events within very short durations may be discerned. Frame rates higher than 100,000 fps or more are quoted in the literature for rapidly changing flows. 2.7 L D V Flow Field Measurement Two techniques of velocity measurement have been employed to obtain information on the flow regimes prevailing in i.e. engines. A number of researchers have made use of HWA to measure velocity fields inside combustion chambers [10, 25, 42, 44, 60]. L D V is the other technique that is being used for flow field measurement. Witze [25] made a critical comparison of both methods for i.e. engine applications with conclusive remarks in favor of L D V . L D V is an optical technique applicable to measure velocity fields of liquids and gases. The physical principle of L D V is based on the shift of frequency of waves due to the motion of the source or reciever. The presence of light scatterers or reflectors in the path of a light beam change the frequency of the given light, where the change in frequency Chapter 2. LITERATURE REVIEW 21 depends on the velocity of the interfering particles and the direction of reflection. Since the frequency shift thus obtained is very small [53], it cannot be observed by simple optical means. Hence the principle of hetrodyning of two beams of equal wavelength and intensity is used to measure the change in frequency which contains the particle velocity; the required information about the flow. Two coherent laser beams are crossed at the focal point of a common lens, where the wave structure of the incident light creates interference fringes throughout the beam intersection region known as the measurement volume [25]. Viewed in a plane perpendic-ular to the common axis of the beams, the interference fringes appear as a pattern of dark and light straight lines. Effort must be made to satisfy the condition of crossing-over the beams at the waist so that a series of uniformly spaced bright and dark fringes, which are mandatory to L D V setup, are obtained. The spacing of these fringes df is related to the wave length A of the laser light and the half-angle of the two intersecting beams, K , such that d f ~ 2sinK i2-18) The measurement volume is defined by the crossing of two beams, dectated by the optical arrangement. An ellipsoidal surface with a major and minor axes lm and dm respectively define the geometry. A typical ellipsoidal measuring volume is shown in Fig.2-3 with characteristic dimensions: lm = ^ (2.19) sinn dP-7 dm = 2.20 COSK d.-> = (2-21) where de-2 is the beam diameter at the e - 2 point intensity, F is the focal length of the focussing lens, and De-2 is the diameter of the beam as it comes out of the laser. The Chapter 2. LITERATURE REVIEW 22 total number of fringes Nfr is calculated from Nfr = ^ = 1.27-^-. (2.22) df De-2 When a small particle ~ df/A [25] passes through the fringe pattern, it will scatter light. A typical Doppler burst from a photodetector is indicated in Fig.3-9. The shape of the signal is distributed according to the change in scattered light intensity as a particle traverses the fringes. The amplitude of the output signal from the photodetector is directly proportional to the intensity of the scattered light. Due to the Gaussian distribution of light intensity across the measuring volume, the signal has a Gaussian envelope. The physics behind the L D V technique lies in fringe generation and the time required for a particle to move perpendicular to the fringes. The frequency, f, of a Doppler burst is related to the velocity component of the particle Up normal to the fringes by the equation: UP = fdf = f-4— (2.23) Zsinn where f in this case is the Doppler frequency measured by any frequency measuring device. Two-component velocity measurement requires hetrodyning four beams, two colours at the measurement volume. The TSI 9100-7 L D V system measures two components of a given flow by crossing four laser beams (two green, two blue) at a single point in the flow and then measures the amount of light scattered by the movement of the particles. The change in frequency, or Doppler shift, of this scattered light, as observed by a stationary detector, is used to calculate the flow velocity. Three techniques of hetrodyne detection are commonly in use where the operation is based on the methods by which the two mixing beams are used to generate a Doppler frequency shift. Chapter 2. LITERATURE REVIEW 23 One of the techniques is the dual beam system. In this system, the scattered lights from two incident beams are hetrodyned. A single beam of a certain wave length is split into two beams of equal intensity, the same wave length and polarity before crossing them over. At the beam axis e~2 of maximum intensity is used to define the diameter De-2 of the beam. The intensity of the beam depends on the radial distance from the beam axis and has a Gaussian distribution. The optical arrangement of an L D V system is shown in Fig.2-5. Beam splitters divide the initial laser beams into two sets of equal and parallel beams, i.e. two green and two blue beams. When the light scattered from the measurement volume is collected by the same lens used to focus the incident laser beams, the arrangement is known as a backscatter mode. In a forward scatter mode a second receiving lens is placed on the other side of the test section, behind which the photodetector is located. Even though the back scatter arrangement is beneficial in cases where access to the measurement volume from the opposite direction is difficult, Adrian and Earley [54] demonstrated that a much higher portion of light is scattered in the forward direction. Thus higher laser power is required for back scatter arrangements. Chapter 3 E X P E R I M E N T A L A P P A R A T U S 3.1 Introduction Study of fluid flow and combustion phenomena in an engine-like environment (out of the actual engine combustion chamber) allows minimization of the number of engine variables and isolation of the behavior and effects of the desired variables. Constant volume combustion bombs are suitable for such phenomenological studies. To exploit the advantage of obtaining first order information on the effects of jet size, jet number and jet orientation in producing turbulence, and to study the effects of this jet generated turbulence on combustion, a cylindrical combustion bomb was used in this research. 3.2 Apparatus and Charge Medium The design of the body of the combustion bomb was initially done by Milane[55] and later used by Bauwens[41] for swirl combustion experiments. In this experiment, the same design was made use of for the body of the combustion chamber. However, alterations were made so that the bomb could accommodate the necessary designs pertaining to the present work. A complete new set of inlet ports for jets were incorporated in such a way that interchangeability of different sizes was possible. The piping network and control system are described in the following sections. 24 Chapter 3. EXPERIMENTAL APPARATUS 25 3.2.1 Combustion Bomb and Piping The body of the cylindrical combustion chamber was machined from stainless steel. The diameter of the bore is 80.8 mm with a length of 43 mm. A section of this chamber is shown in Fig.3-1. Two transparent quartz windows of thickness 25.4 mm each were used on both flat sides of the cylindrical chamber allowing optical access for flow visualization and flow field measurements. Three set of inlet ports were available in this combustion chamber as follows; • two off-set jet ports • two radially opposed jet ports • four radially opposed jet ports. Admission inlets were located in the central plane of symmetry perpendicular to the cylinder longitudinal axis. The two offset ports were set off from the central horizontal plane by 9.5 mm (one port above and the other below the plane) giving room to use different nozzles with different inclinations of jets with reference to the central plane. This arrangement enabled the onset of initial swirl in the chamber during injection. In contrast the radial jet ports were designed so that jets converge at the center of the bomb where the spark was located. In these arrangements counterflows prevailed during injection. Various ports were also available for the spark plugs, exhaust to the vacuum pump, and the piezoelectric pressure transducer. Inlet nozzle inserts were made of brass with a set of diameters 4, 3, and 2 mm. Photographs of sample nozzles (jets) are shown in Fig.3-2. The total volume communicating with the combustion chamber (cylinder volume plus cavities in the inlet ports) was measured by filling the bomb with a liquid and measuring the volume it occupied. The volume was found to be 230 ml with a reading error of ±10 Chapter 3. EXPERIMENTAL APPARATUS 26 ml. The volume of the cylinder is 222 ml and hence the cavities would not be more than 18 ml. The length to diameter ratio of the cylinder (aspect ratio) was about one half, which is a compromise between the two extremes of i.e. engine cylinder piston positions, i.e. top dead center and bottom dead center. The piping is illustrated in the control and instrumentation diagram, Fig.3-3. 3.2.2 Charge Medium and Charging In the study of turbulence generation by jets, the medium of operation was air. In the study of turbulent combustion, a premixed charge of propane and air was used throughout the experiment. The mixture was prepared by controlling the partial pressures and making use of the equivalence ratio desired and using the ideal gas law assumption. Mixture calculation is given in Appendix C. The conditions of charging process, however, were the same for all experiments. Commercially available extra dry air with 10 ppm moisture and instrument grade propane with minimum purity of 99.5% were used to form a stoichiometric mixture. The mixture was stored in a high pressure cylinder to a maximum storage pressure of 2 M P a such that an accidental combustion within the cylinder would not exceed the 17 MPa maximum working pressure of the cylinder. Making use of the partial pressures calculated from the ideal gas assumption, a pre-viously evacuated cylinder was slowly filled with fuel until the desired partial pressure at room temperature was achieved. Then air was slowly added to the cylinder until the final 2.0 M P a pressure was obtained. The new mixture was allowd to homogenize for at least 36 hours before a sample was analysed by gas chromatograph. The gas chromatograph analysis was repeated several times until satisfactory mixture composition was obtained. Chapter 3. EXPERIMENTAL APPARATUS 27 The resulting fuel air ratio is deemed to have 2% uncertainty in composition. The re-quired volume percentage of the fuel was then found from the calibration curve depicted in Fig.3-4. The charging system consisted of a charging tank, piping connecting the charging tank to the bomb, a charging valve, check valves for each inlet port, and a solenoid valve that controls air flow used to operate the charging valve. Since the charging valve was not designed to be exposed to the high pressure and temperature of combustion, a spring-operated check valve rated for high pressure was installed between the combustion chamber and charging valve. The check valve also reduced the inlet cavity minimizing the amount of mixture pushed back into the inlet which could burn after the flame has reached the chamber wall. The charging tank was isolated from the combustion chamber and was in thermal equilibrium with the atmosphere during charging. Pressure in the charging tank and replenishing line was monitored by a strain gauge type absolute pressure transducer, Sensym Model LX1830 A Z . This pressure transducer was used to determine the upstream pressure in each cycle, and was controlled by the I/O card driven by a previously loaded program. To control the charging process, an air-operated charging valve, ASCO Model P210C94 controlled by a direct action, three-way D.C. solenoid valve triggered by the I /O board was used. The air valve has a strong spring, and a much higher energy was required (900 kPa air in this case) to trigger this valve. A strong spring-operated valve (as compared with electric valves) was necessary to close the valve while the flow is still choked[41]. In a previous similar arrangement, this air valve was found to be more effective than electric valves. In addition, it was also found that D.C. solenoid valves were not satisfactory in controling the charging process [41]. The available DC solenoid valves for this purpose Chapter 3. EXPERIMENTAL APPARATUS 28 are pilot-operated, and are designed for minimum power requirement of the coil imply-ing that the actuating force on the diaphram is quite low. Low actuating forces result in unstable and inconsistent valve closure. On this account DC solenoid valve was not used on the charging line. The actuating air pressure (compressed air at 900 kPa) was controlled by an A S C O Model 8321A1 solenoid valve. The piping branched downstream of the charging valve so that the flow could be directed to the desired ports of injection via the different jet nozzles. 3.3 Ignition Ignition energy to the spark elctrodes was supplied with a capacitive discharge ignition system,(Heathkit Model CP1060) through a standard automotive ignition coil. Twelve volt DC power was supplied to the ignition system. A schematic of the electrical wiring of the ignition system is depicted in Fig.3-5. Steel electrodes of diameter about 3mm were installed in the chamber for central ignition. One of the electrodes was grounded to the body of the chamber, while the other electrode, insulated with a ceramic tubing, was connected to the live line of the power supply. The spark gap was located at the centre of the chamber and lies along the longitudinal axis of the cylindrical combustion chamber. The ignition system was controlled by the DT2811 I/O board with software driven by the computer. An Opto-isolator was installed between the ignition system and board to protect the board from any overload and consequent damage. In the case of combustion experiments, after the chamber was filled with mixture, ignition could be controlled by the operator by changing the delay parameter in the program. Chapter 3. EXPERIMENTAL APPARATUS 29 3.4 Pressure Measurement To study the effects of jet size, jet number and jet orientation in producing turbulence, and enhancement of turbulent combustion, it was necessary to keep track of the pressure trace and its development through time in both the charging and combustion processes. The pressure in the chamber was meassured by a piezoelectric pressure transducer (Kistler type 6123 A2). The charge signal sensed by the transducer was amplified by a dual mode charge amplifier (Kistler type 5004) which converted the signal to a voltage output. This voltage was scaled to an equivalent Mechanical Unit (bar in this case) to be used as a conversion factor in the data processing stage. The voltage output from the amplifier was supplied to the I/O board via the screw terminal board and data was then acquired. The fact that piezoelectic crystal elements do not depend on displacement of me-chanical elements makes these transducers suitable for measurement of rapidly changing pressure. However, a known initial pressure value which corresponds to the signal at initial time must be used because the pressure signal provided by piezoelectric elements is differential only. Unless this initial pressure is recorded results may not be reliable. In this experiment, since the chamber was evacuated to about 3 kPa each cycle, the voltage recorded at the beginning of the charging process was taken to correspond to the precharging pressure. The pressure sensor was mounted slightly recessed in the cylindrical wall to reduce the susceptibility of the piezoelectric pressure transducer to thermal loading. However, the recess was not deep enough to produce acoustic resonance. The exhaust line from the chamber was opened and closed by a solenoid valve (VERSA E S M 2301-40 ) triggered by the I/O board. The vacuum level was monitored by a vacuum pressure transducer Sensym Model SCX30 A N . A simultaneous connection to the vacuum pump was also done so that opening the exhaust line and triggering the pump motor could Chapter 3. EXPERIMENTAL APPARATUS 30 take place at the same time. The vacuum pump (CENCO Model Hyvac7) ran until the desired vacuum level (3kPa) was obtained from the transducer. The pressure transducer was calibrated before it was used for measurement. The vacuum level was recorded by the I/O board for later reference of subsequent pressure data analysis. Output from the vacuum transducer was amplified at a gain of 100 and fed to the desired channel of the board. 3.5 Input/Output Board A Data Translation card Model DT2811-PGH Input/Output board designed for use with an I B M P C / X T / A T or compatible microcomputer was used as an interface between the system and terminal to control the apparatus and processes. Fig.3-7 depicts a block di-agram of this I /O board. Analog to Digital (A /D) , Digital to Analog (D /A) conversion and digital Input/Output (I/O) subsystems are the main features of this model. The DT2811 is typically used for industrial process control and low noise laboratory environ-ments (see the operating manual for further details). This board can be plugged into any one of the fully-bussed expansion slots in the microcomputer backplane. The analog to digital converter, which is the main component of the A / D subsys-tem acquires analog voltage inputs from the test section through external sensors such as piezoelectric transducers, thermocouples, etc. and converts them into 12-bit digital codes. The codes are transmitted by the board to the host computer which processes them according to the software preloaded in the computer's memory and the commands issued to the processor by the operator during the execution of the program. The A / D subsystem is equipped with a gain circuit which permits software programmable ampli-fication of the input signal. The DT2811-PGH board provides gains of 1, 2, 4, and 8 for high level inputs. Chapter 3. EXPERIMENTAL APPARATUS 31 In this subsystem 16 single-ended or 8 differential input channels are available for A / D conversions. There are also provisions for unipolar or bipolar configurations i.e. 0-5V, ±5V and ±2.5V. Straight binary output codes are obtained for unipolar inputs and offset binary codes for bipolar inputs. Provisions for operation in single or continuous modes are available. In the single conversion mode, the card performs conversion on a selected channel and stops at the completion of conversion. In the continuous conversion mode, repetitive hardware triggered conversions can be performed on a selected channel. Conversion stops when the trigger is disabled. Both single and continuous conversions were used in this experiment. A l l A / D conversions were initiated by triggers, either software or hardware. Software triggers were issued by the program running the board and hardware triggers by an electrical pulse generated externally by the user or internally by the onboard clock. The D / A subsystem consists of two 12-bit D / A converters, DACO and DAC1. The DACs convert a digital input recieved from the host processor into a voltage output required by the test rig control application. The DACs provide two independent channels , each capable of operating at a throughput of up to 50kHz. Output ranges can be independently configured to be unipolar ( 0 to +5v) or bipolar ±5V or ±2.5V). Transfer of digital data between the P C bus and one or more peripheral devices connected to the DT2811 is provided by the digital I /O subsystem. It consists of two digital I/O ports, port 0, and port 1, where port 0 is dedicated to inputs and port 1 is dedicated to outputs. In a data input operation, the processor reads the status of the digital input/output lines and data is transferred to the computer. In a data output operation, the processor sends data to the digital I /O lines and data is transferred from the computer to an I/O device. The DT2811 is accessed via a 10-bit address received from the host processor. The seven most significant bits (MSB) select the board. The three least significant bits (LSB) Chapter 3. EXPERIMENTAL APPARATUS 32 are decoded onboard and select various registers in the three subsystems. Address bits beyond the tenth bit are ignored. The DT2811 occupies 8 byte locations in the P C bus I/O space. A / D subsystem oc-cupies five locations, D / A subsystem occupies two locations and one location is occupied by the digital I /O subsystem. The base address can be configured to be located between 200(hex) and 3F8(hex) in increments of 8(hex). In this experiment the factory set base address 218H was maintained throughout. The DT707 screw terminal panel was the only accessory installed in this experimental rig. This terminal panel provided direct connection points to all three subsystems and has connection provisions to external trigger and external oscillator inputs. The connection to the host computer was via an integral 50-conductor flat ribbon cable which plugs directly into the DT2811 board's J-conductor. This screw terminal along with a series of optical isolating relays (opto-isolators) were located next to the combustion bomb. The opto-isolators relay control to the various solenoid valves, ignition coil, vacuum pump motor, and camera system with optical, but no electrical connection to the DT707. Any possible destructive electrical surge back to the DT2811 and to the computer could be prevented by installing these relays. 3.6 Schlieren Photographic Arrangement What is the nature of the flow field inside the chamber during and after charging? How far are the jets penetrating to the center of the chamber where ignition is to be initiated? Are the jets interacting and influencing each other? How long does the turbulence thus generated persist and contribute to the enhancement of combustion? Without going to the details and complications of quantitative analysis, substantial qualitative information could be obtained from photographs taken during a specific event to address the questions Chapter 3. EXPERIMENTAL APPARATUS 33 posed above. The schlieren photographic method was chosen in this work for the purpose of flow visualization. The working principles of the schlieren system are well established and the method has been used widely in flow visualization experiments [57, 62]. Making use of the optical properties of light, this method ultimately exploits changes in refractive index of fluid in the test section which is directly proportional to the density gradient. From the change of the optical path of the light beam, fluid motion or combustion phenomenon can be observed. A schematic of the optical bench is depicted in Fig.3-8. A n Argon-Ion laser (Model Spectra Physics 165 ) with a maximum rating of 5 watts was used as a light source throughout this study. The transient nature of the flow in the charging process required a high frequency or a continuous light source, and thus a laser was used as a light source. A collimated green beam of wave length 514.57/m was generated from the system and was passed through the optical arrangement in the laser bench and then reflected back to the work section by two 45° inclined mirrors.The beam from the mirrors passes through a negative lens and expands to a bi-convex lens of focal length 140mm. At the focal point of this condensing lens, a pinhole of diameter 0.5mm was inserted to obtain an approximate point source. This point source was also located at the focal point of the collimating lens L I , which was located in front of the test section. The collimated beam passed through the chamber and was collected by lens L2. Lens 2 in turn focussed the image at its focal point. L I and L2 were of equal diameter, 300mm, and equal focal point of 1000 mm. It was at this focal point of the second lens that the knife edge was inserted. By fine adjustments, it intercepted the light beam from reaching the film plane. The knife edge was found to give better resolution than graded filters or slots. Of particular importance was the sensitivity of the knife edge, where the greater the interception, the higher the sensitivity but this was also associated with higher levels of diffraction. Chapter 3. EXPERIMENTAL APPARATUS 34 A focussing lens of diameter 100 m m was located beyond the knife edge because no camera lens was to be used in recording pictures. The image was directly focussed on the film plane i n such a way that the 16 m m film was almost filled. A high speed camera, H Y C A M Mode l K2001 with a nominal frame rate of up to 8500 fps for 16mm film was used to take movie films at 1000 fps ( this was the maximum rate the camera could run without film breakage). In cold air flow photography, camera triggering was done by the computer where the 3-way solenoid valve and camera line were wired togther so that the valve opening corresponded to the camera beginning to take pictures. 3.7 L D V Experimental Setup In this section the L D V experimental arrangement is explained. A T S I Model 9100-7 L D V system was used to measure the velocity field in the chamber. The Argon-ion laser was also used as a light source for the L D V system. Light beams of 514.5 nm and 488.0 nm wavelengths emitted from the Argon-ion were utilised at a power of 400 m W under operating conditions. The beam had 1.4mm diameter when it exited the laser. The beam exiting the light source was passed through a collimator used to position the beam waist at one focal point ahead of the focussing lens. This was a necessary condition to ensure that the beams were focussed at their waist to obtain a uniform fringe spacing at the crossing point. The collimated beam was then passed through a dispersion prism to separate the required two colours of beams. The green beam at 514.5 nm and the blue beam at 488 nm wavelengths to be used for final crossing of the four beams were then passed through a series of optics comprised of polarization rotators, beam spacers, beam displacers, steering modules and a beam expander. The beams from the beam expander were reflected by two 45° inclined mirrors guiding the four beams to the focussing lens positioned at 250 m m from the test section. The crossing point Chapter 3. EXPERIMENTAL APPARATUS 35 of the beams inside the chamber was carefully checked for the desired fringe spacing. This crossing point was located in the vertical, central plane of the test section. With a standard 50 mm beam spacing and lens focal length of 248.9 mm, the half-angle between the beams at the crossing point was 5.541°. The measuring volume dimensions were: colour green blue Wavelength: A 514.5T7m 488.077m Half angle: K 5.541° 5.541° Fringe spacing:^ 2.664/im 2.527/im No. of fringes:A^/r 20 20 M.V.diameter: dm 117.01/*m 110.99/im M.V.length: lm 1.206mm 1.144mm Scattered light from particles passing through the measurement volume was collected by the same focussing lens used to focus the on-coming beams from the laser source. The signals travelled along the same path and were collected by the recieving optics. Two lenses of focal length 250 mm focussed the collected light onto the respective light detectors' aperture of 0.2 mm each. By focussing the recieving lens exactly on the measurement volume, the level of noise could be minimised. A n R C A vacuum tube of 13% quantum efficiency was used as a photomultiplier. The output of the photo-multiplier is proportional to the intensity of the incident light which itself is proportional to the square of the amplitude of the light wave. A typical velocity signal output from a photomultipler is shown in Fig.3-9a where the pedestal signal is of low frequency. The pedestal signal is a measure of the time a particle takes to travel the whole measuring volume. By high-pass filtering, the pedestal is removed resulting in a clear Doppler burst as indicated in Fig.3-9b. Two TSI Model 1980B counter type signal processors were installed to measure the Chapter 3. EXPERIMENTAL APPARATUS 36 frequencies of the photomultipliers' output signals with a 2 ns resolution clock. Counter signal processors are ideal for processing signals produced with laser velocimeters be-cause they provide accurate measurements on short bursts of signals. They measure the duration of a given number of cycles in a Doppler burst using a high resolution clock al-lowing the calculation of frequencies for discontinuous, unrelated bursts giving accurate, independent signal measurements. In addition, the system has a feature that allows the measurement of the time for a preselected number of cycles. The counter units consisted of three major modules each. Both counters have input conditioners, timers and data transfer modules. The counter module in one counter unit was interconnected to the MI-990 data transfer module in the second counter unit by a 36 pin ribbon cable so that internal identification could be effected on the data acquisition and subsequent data reduction processes. One counter was assigned to measure the U-velocity component while the second counter was to measure the V-velocity component. The input conditioner was used to reduce the photomultiplier signal into a form that could be handled by the timer module. It is comprised of a series of low and high pass filter settings, selections for mode of operation, selectors for gain and number of cycles per measurement. A signal should reach a threshold level of 50 mV in order to be validated as a datum. The accepted signal would trigger the timer module whose function was to measure the period of the signal. Flow INformation Display (FIND) software developed by TSI was loaded in the com-puter and used to acquire, process and store data. One of the most important require-ments in L D V systems is the presence of adequate particle seeding in the flow medium. Measurement of fluid velocity is made possible by measuring the velocity of particles assuming that the particles closely follow the flow. Since commercially available seeders are very expensive, a simple particle seeder was designed and manufactured in the work-shop. Fig.3-10 shows a section of the seeder. Charge enters the seeding chamber at an Chapter 3. EXPERIMENTAL APPARATUS 37 angle of 60° from the horizontal and picks up particles by creating a cyclone during the bomb charging process. Magnesium oxide was used for seeding and the size distribution was measured to be in the range of 0.5pm to 10pm. A desiccator was used to avoid particle dampening from moisture in the air which otherwise could reduce the data rate by fouling the windows. Chapter 4 D A T A ACQUISITION A N D P R O C E S S I N G 4.1 Introduction The operation of the experimental system, data collection , data processing procedures and techniques are discussed in the present chapter. Preliminary procedures and methods of operations are first discussed, followed by data processing methods. System control procedures, the charging mechanism, data collection, flow field photography and combus-tion phenomena are presented. In addition, data acquisition and processing for studying quantitative flow field in the chamber by means of L D V is discussed. 4.2 Data Acquisition Systems Fluid flow, the charging system, the exhaust system, photography and data acquisition were fully controlled by the DT2811 interface card configured and inserted in one of the back slots of the C O M P A Q computer dedicated for this study. Communication between the operator and the system was via the computer by running a preloaded program and by interactively entering the necessary parameters at the prompt on the screen. The following procedure was followed in entering the required parameters for proper execution of the control systems, data acquisition, data storage plus taking photographs during the charging period. First, the number of cycles desired to run the system for a specific condition were entered. The required vacuum pressure level and pressure of the charging tank upstream 38 Chapter 4. DATA ACQUISITION AND PROCESSING 39 of the chamber were predefined from previous calibrations. These pressure levels were monitored by the strain guage type vacuum pressure transducer and the absolute pres-sure transducer respectively. To replenish the charging tank, a calibrated charging time number (computer number) or a predetermined pressure level compatible with load ca-pacities of the components was used. For all experiments a time gap was incorporated in the program between charging the tank and charging the chamber, this time enabled the fluid to settle before it was charged to the chamber. The time required for charging was then prompted and a computer number corre-sponding to a real time was entered. This time depended on the sampling frequency selected on the I /O board. The sampling frequency used throughout the whole experi-ment was 3 kHz. A number corresponding to the time before ignition and an ignition time number were then entered. The number of data points stored corresponded to the charg-ing time number, the time specified between charging and the onset of spark initiation and the time specified after triggering the ignition system. By setting the between-cycle time number (time which allowed an interval for chamber cooling especially for combus-tion experiments), the input procedure would be complete. After the required parameters were entered into the program, the system ran auto-matically until the specified number of cycles were finished. At the beginning of each cycle, the chamber was evacuated to a vacuum of about 3kPa by opening an on-line solenoid valve in the exhaust system and starting the vacuum pump. The pressure was recorded for later use in the data processing procedure to find the actual pressure rise in the chamber. Once the vacuum line was closed, the line from the main storage tank was opened by triggering a solenoid valve. The charging tank was replenished with air (for flow visualization and flow field measurement), and unburned mixture (for combustion tests) to about 350 kPa absolute pressure. This pressure was monitored by the absolute pressure transducer installed upstream of the charging tank. As mentioned above, this Chapter 4. DATA ACQUISITION AND PROCESSING 40 pressure level was controlled by the interface card. In addition the equivalent output volt-age was observed on a voltmeter for fast checking of the pressure and comparison during the experiment. The sensitivity of the absolute pressure transducer was 33.3 mV/psi with a 2.5±0.3 volt zero offset. A voltage output of 4.4±0.10 volt had been recorded for most cases. After the zero offset was deducted from the reading value, the actual pressure was calculated. The charging line was then opened by triggering the 3-way solenoid valve admitting compressed air at 900 kPa used to operate the spring loaded charging valve. A fixed time of 300 ms was used in all procedures of charging, i.e. this was the time for which the charging line remained opened. But the inertia of the plunger in the valve required about 30 ms additional time before it actually closed. The evacuated chamber was charged by making use of the different configurations of jet sets. As explained in the objectives of this work, a set of three ports of varying diameter for each of the three types of configurations was used. As a result of variations in diameter of the jet ports, the entry velocity was obviously different in each case. Estimation of the pressure drop in the piping showed 110, 90 and 70 m/s velocities for 2, 3 and 4 mm diameter ports respectively. Once the charging began, the piezoelectric pressure transducer installed in the cham-ber started responding and the data register was enabled. 1200 data points were collected in a 400 ms time span in each cold flow cycle. In cold flow the ignition triggering line was disabled, whereas in the case of combustion experiments this line was enabled and data were read after the initiation of ignition. There were 600 data points read for combustion, equivalent to a 200 ms duration of time. In cold flow where the primary concern of the experiment was flow visualization, a 16mm high speed movie camera, H Y C A M Model 2001 run at 1000 fps was used through-out. The picture taking procedure was synchronized with the timing of the 3-way solenoid Chapter 4. DATA ACQUISITION AND PROCESSING 41 valve so that when the charging started, movie pictures were taken simultaneously. Ko-dak 7222 black and white film was found to be more effective than Kodak 7250 and 7231 films. Due to the poor resolution of the pictures generated with the latter film types, all schlieren photographs shown were from the 7222 film. During combustion experiments, recording the rate of pressure rise in each case was important and pressure trace recording took place simultaneously with ignition onset. For all cycles and configurations, the ignition system was triggered at a time delay of 66.6 ms after charging. This time was found to be the minimum time that the system could be delayed after charging. A standard automotive ignition system was used to trigger ignition of the premixed charge in the combustion bomb. In all cases, the data from the pressure transducer, was electrical charge. This charge was fed to the dual mode charge amplifier. The charge amplifier whose schematic is shown in Fig.4-1 internally converted the charge to a voltage output. The charging and combustion process voltage outputs from the charge amplifier were then supplied to the Analog/Digital (A /D) conversion subsystem on the board which was enabled by the computer program and triggered by an on-board clock. The computer I/O system acquired, digitized and stored in memory a total of 1200/600 data points for charging and combustion processes respectively, per cycle. The A / D board increments the -5 to +5 voltage range (set on the board through DIP switches) into 4096 intervals i.e. into 2 1 2 steps converting the pressure signal into two byte data. The -5 to +5 range was selected in order to measure all possible voltages from the amplifier. This range was hardware selected, and as a result the other transducers and analog inputs which send data to the same board also operated within the same voltage range. The data recorded was then converted into digital data by using the internally built-in conversion relation in the program. The relation is: V = 10/4096 * Data - 5 (4.24) Chapter 4. DATA ACQUISITION AND PROCESSING 42 The digital data was stored in binary format for each datum point. After the data sensing was complete, the binary data was transferred from the computer to diskettes after which a new cycle was ready to commence. The average absolute pressure in the chamber by the end of charging ranged from 220 to 250 kPa. Hence, the flow to the chamber remained choked at the valve throat. 4.3 Data Processing Three different kinds of data processing procedures were used in this work; processing of films to obtain schlieren photographs, processing of pressure time histories, and processing of velocity field data. Flow visualization via the means of schlieren photography was previously discussed. Special film processing techniques were required in developing and printing the movie films which were used in this work. The quality of the pictures depended on how much care was taken in the entire picture taking and printing process. Light filtering, time re-quired to expose the negative on a printing paper, and intensity of the light contributed to the clarity of the printed outputs in addition to other systematic errors that existed in the picture taking process. For each jet size in a configuration, the films were devel-oped and made ready as movie pictures. Jet flow developments, circulating flows in the chamber, degree of jet penetration and persistenece of jets are to be observed in these pictures. Pressure time traces were stored in binary format on diskettes as mentioned earlier. A data processing program that could read a number of cycles of raw data converting them to ASCII, ensemble average over a given number of cycles or take a single cycle was written. The procedure of converting raw data was the same for cold flow and combustion processes. The digital (voltage) data was converted to pressure by a scaling Chapter 4. DATA ACQUISITION AND PROCESSING 43 factor, a function of the gain and charge amplifier settings in mechanical units/volts. The pressure in bars was determined by: Pbar = {Data * - 5) * Scale. (4.25) 4096 This equation takes into acount the resolution of the A / D board, the output range of the board which was jumper configured (±5)Volt, and the scale adjustment used in the charge amplifier where the corresponding time depended on the selected sampling frequency. The scale was set to 5 for combustion tests and set to 1 for tests without combustion. Final pressure trace data were obtained by ensemble averaging over a number of cycles. The following formula was used to process pressure data. PW=Jf-i:P(t>n) (4-26) C fl— 1 where Nc is the number of cycles and t is time. Fifty cycles were used for averaging combustion data while twenty cycles were used for charging pressure traces. The final phase of analysis was evaluation of the mass fraction burned process from the ensemble average pressure traces. The mass fraction burned MB(t) in an adiabatic constant volume chamber is nearly equal to the fractional pressure rise, even if dissociation exists [26]; MB(t) = £ ^ _ Z ^ (4.27) 'marc ' i where MB(t) is the mass fraction burned at a given time, P(t) is the chamber pressure at a given time, Pi is the pre-ignition pressure, and Pmax is the maximum pressure following combustion. This simplified procedure scales the mass burn fraction between 0 and 1. The start of combustion was taken at the triggering of the ignition system while end of combustion was determined where the rate of pressure rise was zero or negative. Chapter 4. DATA ACQUISITION AND PROCESSING 44 4.4 L D V Data Acquistion and Signal Processing The desire to measure two-component velocities of the flow resulted in the implementation of the L D V system where measurement took place on the central vertical plane of the chamber. Signals were measured on the axis of the jet, and on the horizontal axis of the chamber at four locations as illustrated in Fig.4-2. At each station five cycles were measured for ensemble averaging. The signal to be measured was the Doppler frequency shift of the intensity of light scattered by the particles as they traversed the measurement volume. The signals de-tected by the photomultipliers were fed to the two counter-type signal processors where the pedestal and noise were removed by the low-pass and high-pass filters set at 30 kHz and 30 MHz respectively. The FIND software, (developed and provided by TSI ) was used to acquire and store data. The time between data was also recorded. The data from the two counters were collected intermittently and were identified as to which counter they came from. FIND was further used to convert the raw data to time history outputs. However, a problem existed in further data processing using FIND due to the transient nature of the flow; since FIND was developed to process steady flow phenomena. Therefore a need arose to look for a means of processing the velocity time histories to extract the necessary information about the flow. Analyzing data from bomb measurements was not as straightforward as might be ex-pected. In a steady flow situation it is the usual practice to decompose the instantaneous fluid velocity U(t) into a mean velocity U and a turbulent fluctuation u{t) as expressed by equation 1-2, where in that case the mean is a simple time average. But in the combus-tion bomb, the charging process and ensuing fluid motions were time dependent. Thus mean velocity was a function of time rather than a constant. Chapter 4. DATA ACQUISITION AND PROCESSING 45 It is common to use ensemble averaging to define the instantaneous mean velocity as a function of crankangle degree or time [11, 25, 19] over a number of cycles. UE(t) = ^ f:U(t,i) (4.28) where U(t,i) is the instantaneous velocity at time t during the ith cycle of the measure-ment. But there is a problem of denning the mean velocity where the L D V data rate is rather low and it is impossible to generate a continuous or nearly continuous velocity record for each cycle. Data rate was limited by the particle generation capacity, and therefore the magnitude of cycle-to-cycle variation could not be identified. Consquently, an ensemble-averaging scheme, outlined below has been used for the present data analy-sis. First, the data was split into 10 ms windows. A mean velocity was then calculated by simply arithmetically averaging the data points which fell in that window. The average at each window midpoint was defined by: M*) = ^ E E ^ ( « ± Y ) (4-29) where N is the number of velocity measurements recorded in a window during the ith cycle, Nc is the number of cycles, and Nt is the total number of measurements. The corresponding equation for the ensemble-average root-mean-square velocity fluctuation is { i Nc Ni 1 uFtEA(i) = \ T 7 E E I 7V< i=i j=i (4.30) where ui,i ~ -UEA-This definition of fluctuation intensity includes cyclic variations in the mean flow as well as the turbulent fluctuations about each cycle's mean [61]. However, cyclic variations in Chapter 4. DATA ACQUISITION AND PROCESSING 46 a bomb are likely to be much smaller than in an operating engine due to the combined effects of engine variables on the latter. Chapter 5 E X P E R I M E N T A L R E S U L T S A N D DISCUSSION 5.1 Introduction Results obtained from experiments conducted on the combustion bomb for the generation of turbulence by means of different port sizes and the ensuing turbulent combustion are presented and discussed in this chapter. In the first part, results from flow visualization experiments are discussed. This is followed by a discussion and analysis of jet-enhanced turbulence and pressure traces, combustion pressure developments and mass fraction-burned results. The effect of jet size, number, and orientation in generating turbulence and its enhancement of combustion are presented. In addition flow field measurements acquired from L D V measurements are discussed. 5.2 Flow Visualization Flow visualization photographs taken by means of schlieren photography system for cold flow condition are presented in Figs.5-1 to 5-9. During the experiment three systems of configurations were used; the offset arrangement, the 2-radial jet and the 4-radial jet arrengements. A sequence of photographs of the turbulence generated by the 2,3, and 4 mm jet sizes for the various arrangements are depicted. In all the photographs displayed, the frames selected began at 73 ms and ended at 346 ms after the charging valve was opened. This range covered the time in which fluid was being charged into the chamber and deemed to cover the event. 47 Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 48 5.2.1 Two Offset Jet Arrangement For the two offset ports configuration shown in Fig.4-2 top sketch, the photographs show jet flows commencing, developing, breaking down into swirling eddies and dissipating. In all diameter cases the jets are seen to emerge at about 80 ms or later, develop for a short while, and then reach the opposite walls of the chamber, but the flow is then dominated by the swirl initiated due to the offset configuration. The two jets can be clearly seen in frames between 81 and 91 ms for the 2 mm port, between 85 and 140 ms for the 3 mm port, and between 81 and 190 ms for the 4 mm port, after which the swirling flow is seen to give the jets a tangential component of velocity which then results in bulk swirling motion in the chamber. Qualitative comparison within this arrangement shows that the intensity of a jet is inversely proportional to the respective jet diameters studied. This means the smaller the diameter, the higher the strength of the jet. The strength of the jets in the respective ports is manifested in the development and dominance of the ensuing bulk swirling flow in the chamber. Tangential velocity calculations from L D V measurements given in figs.5-50 to 5-53 also show the magnitude and duration of the bulk swirling flow for the respective ports. The tangential velocities for the offset arrangement are larger in magnitude and persist for a longer duration of time. For the 2 mm diameter jet, swirling motion along with impinging jets dominates the initial charging period, after which a decaying swirl was prominent . It can be seen that the swirling motion persisted for the 2 mm jet longer than the other two diameters. Even though the behavior of the flow for the 3 mm jet appeared to be similar to the 2 mm jet, the strength difference was significant. Swirling motion started later by about 20 milliseconds and decayed earlier as well. The 4 mm jets show the jets dominating for a longer time, but they were not strong enough to penetrate to the centre of the chamber. The swirling motion appeared to be Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 49 suppressed and the turbulence decayed earlier than for the other jet sizes. This can be explained by the fact that due to the size of the jet, the entrance velocity was lower which in turn meant that the momentum and energy were lower than the other jet sizes leading to earlier dissipation given the same conditions. Viscous friction, the chamber walls, and spark plug rods were factors that could be responsible for dissipation in all cases. A n approximate calculation of the rate of momentum of the initial charging process was performed based on the L D V measurement taken on the axis of the jet 4.2mm away from the entrance of the jet. A parabolic profile of the flow was assumed with the maximum velocity being at the centre. This velocity is used to calculate the flow rate and the mean, bulk outlet velocity. The density of air at 20°C (1.21kg/s) is then used to evaluate the mass flow rate. Together with the mean, bulk velocity, the mass flow rate resulted in the following rate of momentum for each considered case. The rate of momentum for the 2 offset, 2 mm jet was 2.Ql*10~1kg — m/s/s, 6.06* 10-1fc<7 — m/s/s for the 4 mm jet, 2.08* 10-1fc<7 — rn/s/s for the 2 mm jet in the 2 radial jet arrangement, and 1.96 * 10-1fc<7 — m/s/s for the 2 mm jet in the 4 radial jet arrangement. However, these values are subject to change with time due to the unsteadiness of the charging process, and hence may not be used to evaluate the overall rate of momentum of the charging process in the chamber. 5.2.2 Two Radially Opposed Jet Arrangement Figs.5-4 to 5-6 show schlieren photographs of flows for the 2 radial jet arrangement within which 2, 3, and 4 mm diameter jets were used. The frames presented are for the same time range as in the earlier case. The commencement of jet flow is seen to be between 81 and 90 milliseconds. In this case also, the two jets are clearly seen between 85 and 211 ms, between 85 and 190 ms, and between 85 and 140 ms for the 2, 3, and 4 mm ports respectively. Strong jets capable of reaching the center momentarily were generated. The Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 50 persistence of the jets in this case also varied inversely with the jet diameter. Jets are visible for a longer time in the 2 mm case than the other sizes. The in-chamber flow was dominated mostly by the jets in the 2 mm case. The jet interaction did not generate persistent and long lived turbulence structures in the 3 mm case either. Even though the level of swirl was not strong, it started and decayed earlier in the 4 mm jet case. The turbulent nature of the flow was vivid. In this arrangement, weak swirling turbulent motion could be attributed to the opposing nature of the jet flows. 5.2.3 Four Radially Opposed Jet Arrangement Figs.5-7 to 5-9 show photographs of air charge flow for the 4 radial jet arrangement where four diametrically opposing jets converge towards the center of the chamber. The same three jet sizes were used in this case also. As can be seen in the photographs and from observations during the execution of the experiments, the jets were broken down the moment they were generated and the turbulent flow was short lived compared with the 2 offset and 2 radial configurations. It is barely possible to see the commencement and developments of jets in these pictures. The reason for the phenomenon shown in the photographs lies, as in the case of the 2 radial jet case, in the fact that counter flows have the effect of cancelling each other with limited chance of long-lived and strong swirling turbulent motion. Thus in this arrangement the flow settled much quicker than the rest of the arrangements. The large turbulent scale in the offset arrangement developed to the size of the cham-ber. It takes a longer time to break down swirl of the chamber size to the Taylor mi-croscale at which dissipation may take place. In the two radially opposed arrangement, a large scale of half the chamber size developed and it takes half the time of the previous case for the swirl break down to the Taylor micro scale. By the same analogy the largest Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 51 eddy size that can develop in the 4 radially opposed jets case is one quarter the size of the chamber taking a quarter of the time that may take the full chamber size eddy to dissipate. The decay time was a function of the large scale motions generated in the chamber. Comparing the different arrangements, the 2 offset jet arrangement was found to be more effective in generating turbulence that was strong, and with longer life. With this arrangement, the turbulence generated by the 2 mm jets was more favorable than by the 3 and 4 mm jets. The offset gap has not been optimized to determine whether it has any significant effect in changing the turbulence structure. 5.3 Jet Enhanced Turbulent Combustion 5.3.1 Air Flow Pressure Results Prior to combustion, air was used to charge the chamber to get information on the flow and the pre-ignition pressure rise that took place for every cycle. A study of the cyclic variations of pressure rise without combustion showed that the experiments were quite repeatable for all arrangements. A maximum pressure of 245 ± 10 kPa was obtained in most cases except in the case of the 4 radial jets arrangement, where the maximum pres-sure rise was about 220 ± 10 kPa. This reduction in charge pressure might be associated with the pressure loss upstream of the chamber due to the additional piping required. For all cases, the time for the maximum pressure was about 330 ms after the charging valve was triggered to open, and 30 ms after the valve was triggered to close. Comparisons of average pressure rise are shown in the graphs of Figs.5-10 to 5-12 within each group, i.e. different diameters are compared within each configuration. The 2 offset and 2 radial jet arrangements show that the rate of pressure rise was directly proportional to jet diameter for the first 100 ms after the charging valve was opened. Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 52 The charge pressure with 4 mm diameter developed quickly but slowed after about 200 ms. Similar phenomenon was observed for the 3 mm jet as well. The 2 mm jet charge pressure developed slowly with a near linear manner and reached approximately the same maximum pressure as in the other jet sizes at the same time. The 4 radial jet set pressure rise was different from that of the other 2 arrangements. Charging with 2 mm jets gave a faster rate and maximum pressure compared to the 3 and 4 mm diameter jets. 5.3.2 Combustion Results and Discussion A quiescent combustion experiment was executed to be used as a reference to combustion pressure rise obtained with different jet sizes in the various arrangements. It is important in an engine or a bomb to obtain appropriate turbulent motion at the desired time, just prior to ignition. Violent and insignificant turbulent flow fields are not useful in facilitat-ing the conditions for releasing the energy from the combustible fuel. The combustion rate effect due to jet sizes and configurations will be evaluated for a specific ignition time delay, (66.6 ms). Comparisons of combustion pressure traces and the maximum pressures obtained associated to the variables considered will be utilized to evaluate results. Figs.5-13 to 5-21 show combustion pressure developments on a pressure-time axis (where time is referenced to time of triggering the ignition system) to consider the cyclic repeatabilities of the experiments. The 2 offset jet arrangement shows good repeatability with 1.2% deviation from the average maximum. Repeatability deteriorates with the changing configuration and increasing jet diameter. Average combustion pressure rises along with the quiescent combustion pressure rise are plotted in Figs.5-22 to 5-27 for all cases. As expected the pressure rose slowly for quiescent combustion. Table 5-1 shows the maximum pressures obtained and time taken for the respective maximum pressures for each jet size within a configuration. The time taken to reach a maximum pressure of 837 kPa was about 56 ms for quiescent combustion. Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 53 Considering the 2 offset arrangement, it took about 13 ms, 14 ms, and 12.8 ms to reach 1200 kPa, 1286 kPa, and 1386 kPa maximum pressures for 4 mm, 3 mm and 2 mm jet diameters respectively. A l l times were referenced to the time of triggering the ignition system. Reduction in the time to maximum pressure of 76.7%, 75% and 77% were obtained for 4 mm, 3 mm, and 2 mm jets compared with quiescent combustion. Corresponding maximum pressure increases of 48%, 54% and 66% were obtained. In this arrangement there was no noticable difference in the rate of pressure rise. A large increase in the maximum pressure over the quiescent combustion case is noted. The rate of combustion increase is as well seen the short duration of combustion as a result of increased fluid motion. The fluid motion at and during ignition is responsible for the fast combustion associated with the different jet flows. This condition allowed fast release of the chemical energy of the fuel well before wall quenching had the effect of reducing the maximum pressure. In the quiescent relatively longer time was required to completely burn the charge, whereby by then, heat transfer to the cold chamber wall significantly reduced the maximum pressure that could be attained had the combustion been conducted in an adiabatic process. A study of the results for the 2 radial jet arrangement indicated that 19 ms, 22ms and 21.5 ms were required to reach maximum pressures of 1162 kPa, 1088 kPa and 1250 kPa for the 4 mm, 3 mm and 2 mm diameter jets respectively.Comparing the times taken to reach the maximum pressures with that of quiescent combustion, reductions of 66%, 61% and 62% were achieved. Percentage increases of maximum pressure compared to the maximum pressure of quiescent combustion are 39%,30% and 49%. No significant combustion rate differences were seen in the first 15 milliseconds of combustion, however, different slow burning rates then took place until the respective maximum pressures were reached. The phenomena in the 4 radial jet sets were different in magnitude and behavior Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 54 from those of the previous sets. Maximum pressures were on the order of the quiescent combustion case. Maximum pressures of 1000 kPa, 824 kPa and 888 kPa and time lapses of 56 ms, 67 ms and 42 ms were recorded for the 4 mm, 3 mm and 2 mm jet diameters. The time taken to reach the maximum pressure for the 4 mm was equivalent to the time taken to reach the maximum pressure in the quiescent combustion case. A 25% reduction of time to maximum pressure was obtained for the 2 mm port. Maximum pressure increases of 20% and 6% were achieved for the 4 and 2 mm ports respectively over the maximum pressure in the quiescent combustion. It took more time for the 3 mm port to reach the maximum pressure than the quiescent case, and the maximum pressure was less by 2%. As noted in the flow visualization section, the level of turbulence was very low and short-lived for this particular configuration. This condition gave rise to a very slow combustion phenomenon in the order of the quiescent combustion, having a similar effect on the maximum pressure and rate of combustion as discussed above. The rate of combustion decreased because of the slow fluid motion in the chamber, while the effect of heat transfer to the cold walls of the chamber had a pronounced effect in reducing the maximum pressure that could be achieved otherwise. In addition to comparisons within a group, configuration comparisons were conducted by considering the same diameter of port one at a time. Results are plotted in Figs.5-25 to 5-27. The 2 mm jets used in the offset arrangement were found to have the highest maximum pressure and the shortest time to maximum pressure. The combustion rate difference is clearly marked on the graph. The trend is for the maximum pressure to decrease from the offset to the 2 radial and then to the 4 radial jet arrangements. In all cases the offset arrangement was effective in giving rise to maximum pressure with incrased rates as compared with the other sets. According to Witze and Vilchis [49], cyclic variation strongly depended on the de-tachment of the early flame from the spark plug. Low cyclic variation was observed with Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 55 high swirl. Mayo [56] observed that a reduction in cyclic variation was brought about by reducing the burning duration. In light of the above findings, the results of this study are in agreement. Observations from flow visualization and combustion results agree that the cyclic variations were lower and the swirl lasted longer with the offset jet configuration than the radial jet configuration. 5.4 Mass Fraction Burned In the previous chapter it was explained that the mass fraction burned for an adiabatic, constant volume chamber would be calculated from the fractional pressure rise and the maximum pressure during combustion. Graphs plotted from these data are given in Figs.5-28 to 5-33. For the purpose of comparison among different jet sizes and configu-rations, times taken for 10%, 50% and 90% mass burned are identified in the respective cases and tabulated in table 5-2. Comparison within each configuration showed the 2 mm diameter jet had the shortest times to burn 90% of the mixture mass in the chamber. As stated earlier this can be explained by the jet generated turbulent mixing process favouring faster combustion than the 3 mm and 4 mm cases. Overall comparisons among the configurations studied revealed that the offset ar-rangement was more effective in promoting fast combustion than the other arrangements. Figs.5-31 to 5-33 illustrates this phenomenon. The advantage of the offset geometry over the radial arrangements was because of initial swirl formation and the subsquent slow decay of turbulent motion. Given the same charging pressure for all cases, an approximately equal mass has been charged within a very short time(300 ms). It can be argued that the momentum inherent to the turbulent motion was responsible for the increased rate of combustion. The increased velocity Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 56 associated with the 2 mm port gave rise to the increased momentum effectively delaying dissipation, therefore resulting in a faviourable condition for fast burning. 5.5 Flow Field Measurement Two-component velocity measurements as a function of time were conducted in the cen-tral vertical plane of the chamber at the locations indicated in Fig.4-2, along the axis of the jet and along the horizontal axis of the chamber. The processed data for the time dependent flow field at the respective stations are illustrated graphically in figs.5-34 to 5-49. In each graph, the window ensemble-averaged mean component velocity and the corresponding ensemble-averaged fluctuation velocity are displayed. The locations of measurements for each case are also pointed in the insets by a dot mark. The dimen-sionless ratio r/R is the ratio of a radius at a measurement location from the centre to the radius of the chamber. In all velocity graphs, time is referenced to the beginning of Doppler signal recieving, not to the time of triggering the charging valve. Care must be taken if comparison of flow visualization and L D V results is desired. The two-component results show non-simultaneous measurements of two orthogonal components of velocity at the given locations. When one considers averaged quantities, the disadvantage of non-simultaneous measurements are not significant [19]. In all cases the U-component velocity shows a decreasing trend in time while the V-component velocity has an increasing trend during the process of charging. The fluc-tuation velocities decrease all the way through the charging period and after charging. The fluctuating forces which give rise to turbulent flow appear to be stablizing with time. Comparison of the graphs indicate that the mean velocities are of approximately equal magnitude for the same port diameter irrespective of configuration. The processed data represent the 2 mm diameter port in all sets of configuration and the 4 mm port in the Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 57 offset arrangement. A peak mean U-velocity of 70 m/s and a peak mean V-velocity of 50 m/s were noted for the 2 mm port. Similar observation shows a 55 m/s and 35 m/s of U and V component velocities respectively for the 4 mm port. Flow fields away from the center seem to be similar. There is considerable difference of flow at or near the centre of the chamber. Flow activity around the center deteriorated much faster in the 4 radial arrangment than in the other two sets. Tangential velocities were calculated for each considered case at specified time inter-vals. It is extremely hard to interpret what is hapening in the flow from measurements made at only one location. Different flow structures can pass near or through a par-ticular measurement point making the explanation of what is happening difficult other than at the measured point. However, the data collected was used to construct a movie of the tangential velocity as a function of time. The net effect is a quantitative flow visualization. Figs.5-50 to 5-53 show the mean tangential velocity at the 4 locations in the central plane. The lines indicate the magnitude and direction of each local tangential velocity. A reference velocity scale is shown in each frame. The figures reveal a type of vortical flow in the chamber. But the velocity distribution is neither solid body nor free vortex. A similar conclusion was drawn by Rask [19] from the study of flow field in an engine cylinder. It is beyond the scope of the present work to conduct 3D flow field study which would have revealed the whole nature of the flow. The approach in this study was to get preliminary information at the indicated locations. The information thus gathered shows the swirling nature of the flow and confirms the existence of changes in angular momentum. The figures show the differences in flow fields of the different configurations. The result obtained from L D V measurement associated with flow visualization results can be used to evaluate the effects of port sizes and orientataions on enhancing turbulence. Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 58 It was found that the offset arrangement has the largest effect of generating turbulence which is believed to be the reason for the combustion results obtained. The effectiveness of the configurations is of the same order ( 2 offset, 2 radially opposed and 4 radially opposed jets) as in the previous discussions. 5.6 Experimental Uncertainties An assessment of the limitations of the measuring techniques used is discussed in this section. Uncertainties involved in this investigation might be caused by errors in appa-ratus or instrument construction, systematic errors or random errors in the apparatus or instruments. 5.6.1 Pressure Measurement The pressure transducer upstream of the charging tank was repeatedly calibrated with a sensitivity error of 2%. The error in the maximum pressure of the charging tank after replenishing was found to be less than ± 3 % . Both the piezoelectric pressure transducer and the charge amplifier are specified with an accuracy of ± 1 % . The error associated with the pressure data is therefore within ±2%. The pre-ignition pressure may have varied by about 10 kPa. A conservatve estimate on the effect of peak pressure is that a variation in pre-ignition pressure has a ten fold effect on peak pressure [65]. Thus the variation in peak combustion pressure due to initial pressure variation is about 100 kPa. This represents an error of less than 12% on the peak pressure of combustion. Conversion systems and data acquisition errors are included in this error. The effect of variable initial pressure on the rate of mass fraction burned is considered negligible. A systematic error may exist in the calculation of the mass fraction burned as a function of the fractional pressure rise. The relation given in eqn.4-27 is derived for Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 59 adiabatic combustion with no dissociation, but the combustion within the bomb exhibits heat transfer and possibly dissociation effects. The effect on the calculated mass fraction is thus assigned up to 2%. Another source of error related to the combustion process may be the fuel compo-sition. The Hewlett Packard research gas chromatograph was used to determine the stoichiometry of the mixture. From the calibration of the machine with different known mixture volumes, a mixture strength uncertainty of 2% was obtained. This error is con-sidered to result in a negligible effect on the burning duration. The effects of impurities in the fuel and air was shown to be less than 1% [64]. 5.6.2 Flow Measurement Technique The L D V system measures the time it takes to traverse a fringe spacing for a specified number of cycles. There were 20 fringes in the present study. The velocity is then calculated by dividing the fringe spacing by the measured time. The timer has 125 MHz clock with a 2T/S resolution. If it is assumed that the system is measuring on a real Doppler burst, the error in the counter measurement of time is 0.25% at 20 MHz, and 1.0% at 80 MHz [63]. 80 MHz corresponds to putting the photomultiplier output directly into the counter. The data obtained was random rather than continuous velocity variation. The intermittency in data is due to the time resolution error . The digital data system uses 12 MSB (Most Significant Bits) binary bits, for approximately 0.25% accuracy. TSI measured and specified the beam angle with an error of 0.1%. The total error of the Doppler measurement is, therefore, approximated to be less than 2%. Another source of error associated with the L D V measurements is velocity biasing. Velocity biasing is due to the higher probability of faster particles entering the measuring volume than low-velocity particles thereby signalling more data points. Thus the mean velocity obtained by a simple arithemetical averaging would be biased towards the higher Chapter 5. EXPERIMENTAL RESULTS AND DISCUSSION 60 velocity data points. The data in the present work have not been corrected for velocity bias. A source of error with the highest potential for causing large uncertainty is the mea-surement of noise instead of Doppler signals. The system was checked by confirming that no-particle-seeding corresponded to zero data rate. In addition, it was checked by blocking one beam and making sure that the data rate went to zero. It is difficult to quantify the error that might result from such noise. A limited number of velocity measurements were used to calculate the ensemble-averaged mean and rms velocity fluctuation. It was thus subject to statistical sampling uncertainty. The method used by Rask [19] to calculate the errors was applied to estimate the uncertainties. Generally, the error in the mean is proportional to the relative rms fluctuation (fluctuation normalized by the mean) divided by the square root of the number of samples. Up to 30% error is approximated for the mean. Similarily, the error in the rms fluctuation is inversely proportional to the square root of the number of samples. The uncertainty associated with the rms fluctuation is assigned up to 50%. Chapter 6 C O N C L U S I O N S A N D R E C O M M E N D A T I O N S The objective of this study was to investigate the influence of jet size, jet orientation and jet number in enhancing turbulence in a constant volume chamber and on the combustion rate. Three groups of configurations were examined. 1. Offset arrangement, 2 jets 2. 2 Radially opposed arrangement 3. 4 Radially opposed arrangement. Within each group three different port (jet) sizes were studied; 1. 2 mm diameter 2. 3 mm diameter 3. 4 mm diameter. Evaluation of the three configurations was conducted in three phases: 1. Flow visualization by schlieren photographic technique. 2. A study of jet enhanced combustion. Combustion pressure measurements were conducted. The maximum pressure and time to maximum pressure were the char-acterizing parameters. From the pressure traces mass fraction burned curves were calculated using the simple model of fractional pressure rise proportional to mass-burned fraction. The rate of combustion was then evaluated from this data. 61 Chapter 6. CONCLUSIONS AND RECOMMENDATIONS 62 3. Flow field measurements using L D V were conducted and analyzed for selected groups of jet ports. The respective mean tangential velocities were calculated. 6.1 Conclusions A summary of the conclusions drawn from this study are the following. • Jets are an effective means of enhancing turbulence and combustion rate provided they are not arranged opposedly. • The flow field generated in the chamber was influenced by jet size. The 2 mm port has the most influence. • The rate of combustion and maximum pressure rise were influenced best by the offset configuration. • The radially opposed jet arrangements were not effective because of the counter flow resulting in quick turbulence breakdown and dissipation. • An increased number of jets in opposed arrangments does not enhance turbulence and turbulent combustion. • Jet orientation has an influence on the flow field. The offset configuration generated more slowly decaying flow than the rest of the configurations. • Turbulent flow arround the spark has important influence on the initial flame for-mation and subsquent pressure rise. 6.2 Recommendations The results obtained from this experimental study suggest that further work be done on the combustion bomb, both experimental and numerical. The gap between the offset Chapter 6. CONCLUSIONS AND RECOMMENDATIONS 63 jets needs to be optimized. Increased numbers of jets must be investigated in an offset arrangement. A variable ignition delay could be another parameter of study. In addition a 3D flow field study by numerical simulation seems important in revealing the full scale chamber flow field. The trends identified in this work were not in an engine environment where a number of variables come into play. A combustion chamber design including the offset arrange-ment is recommended. This design may take into account previous findings in squish and squish-jet applications. The orientation of the jets is recommended to be towards the spark. Bibliography [1] Young, M . B . , "Cyclic Dispersion in the Homogeneous Charge Spark Ignition Engine - A Literature Study", S A E Paper 810020, 1981. [2] Hancock, M.S. , Buckingham, D.J . , Belmont,M.R., "The Influence of Arc Parameters on Combustion in a Spark Ignition Engine", S A E paper 860321, 1986. [3] Anderson, R.W. , "The Effects of Ignition Power on Fast-burn Engine Combustion", S A E Paper 870549, 1987. [4] Soltau, LP . , "Cylinder Pressure Variation in Petrol Engines", Proceedings of the Institution of Mechanical Engineers, No.2, 1960 - 1961. [5] Patterson, D.J . , "Cylinder Pressure Variation, A Fundamental Combustion Prob-lem", S A E Paper 660129, 1966. [6] Evans, R .L . , "Internal Combustion Engines Squish-jet Combustion Chambers", U.S.A. Patent, #4,572, 123 Feb 25 1986. [7] Hinze, P.O., "Turbulence", 2nd Ed. McGraw-Hill, 1975. [8] Schlichting, H . , "Boundary-Layer Theory" 7th Ed. McGraw-Hill, 1979. [9] Tennekes, H . , and Lumley, J .L. , "A First Course in Turbulence" MIT Press, 1972. [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. 64 Bibliography 65 [11] Lancaster, D.R., "Effects of Engine Variables in a Spark Ignition Engine", S A E Paper 760159, 1976. [12] Fuller, D.E. , and Daneshyar, H . , "Definition and Measurement Parameters in Re-ciprocating LC.Engines", S A E Paper 861529, 1986. [13] Whitelaw, J .H. , et al, "Squish and Swirl-Squish Interaction in Motored Model En-gines", Transactions of the A S M E , Journal of Fluids Engineering Vol.105, 1983. [14] Heywood, J .B. , "Fluid Motion Within the Cylinder of Internal Combustion Engines-The 1986 Freeman Scholar Lecture", J . of Fluids Engineering Vol.109/3, March 1987. [15] Winsor, R .E . , and Patterson, D.J . , "Mixture Turbulence- A Key to Cyclic Combus-tion Variation", S A E Paper 730086, 1973. [16] Saxena, V . , and Rask, R .B . , "Influence of Inlet Flows on the Flow Field in an Engine", S A E Paper 870369, 1987. [17] Bracco, F . V . , and Hall, M . J . , "A Study of Velocities and Turbulence Intensities Measured in Firing and Motored Engines", S A E Paper 870372, 1987. [18] French, D.T. , and Fansler, T.D. , "Swirl, Squish and Turbulence in Stratified-Charge Engines: Laser- Velocimetry Measurements and Implications for Combustion", SAE Paper 870371, 1987. [19] Rask, R .B. , "Laser Doppler Annemometry Measurement in an Internal Combustion Engine", S A E Paper 790094, 1979. [20] Witze, P.O., "Application of L D V to Spark-Ignition Engines", Flow Lines, Summer 1989. Bibliography 66 [21] Matsuoka, S., et al, " L D A Measurement and a Theoretical Analysis of the In-Cylinder Air Motion in a DI-Diesel Engine", S A E Paper 850106,1985. [22] Cole, J .B. , and Swords, M.D. , "An Investigation of the Ignition Process in a Lean-Burning Engine Using Conditionally Sampled Laser Doppler Annemometry", S A E Paper 800043, 1980. [23] Fraser, R .A . , and Bracco, F .V . , "Cycle-Resolved L D V Integral Length Scale Mea-surements in an I.C.Engine", S A E Paper 880381, 1988. [24] Obokata et al, "Velocity and Turbulence Measurements in a combustion Chamber of S.I.Engine Under Motored and Firing Operations by L . D . A . with Fiber-Optic Pick-Up", S A E Paper 870166, 1987. [25] Witze, P.O., "A Critical Comparison of Hot Wire Annemometry and Laser Doppler Velocimetry for IC Engine Applications", S A E Paper 800132, 1980. [26] Lewis, B . , and Von Elbe, G. , "Combustion, Flames and Explosions of Gases", 3rd Ed. Academic Press Inc. 1987. [27] Reynolds, O., Phil . Trans. Roy. Soc, Ser. A,186, 123 (1895). [28] Townsend, A . A . , "The Structure of Turbulent Shear Flow", Cambridge University Press, London, 1956. [29] Bradshaw, P., "An Introduction to Turbulence and its Measurement", Pergamon Press, Oxford, 1971. [30] Taylor, G.I., "Statistical Theory of Turbulence", Book Extract, 1935. [31] Mallard, E. , and Le Chatelier, H.L. , Ann. Mines Paris 4, 343, (1883). Bibliography 67 [32] Damkohler, G. , "The Effects of Turbulence on the Flame Velocities in Gas Mix-tures", N A C A T M 1112, 1947. [33] Glassman, I., "Combustion", Academic Press Inc., 1977. [34] Lancaster, D.R., Krieger, R .B . et al, "Measurement and Analysis of Engine Pressure Data", S A E Paper 750026, 1975. [35] Shelkin, K. I . , "On Combustion in a Turbulent Flow", N A C A T M 1110, 1947. [36] Tabaczynski, R.J . , "Turbulence and Turbulent Combustion in Spark Ignition En-gines", Prog. Energy Combustion sci.Vol.2 pp 143- 165, 1976. [37] Herweg et al, "Flow Field Effects on Flame Formation in a Spark-Ignition Engine", S A E Paper 881639, 1988. [38] Chomiak, J . , "Flame Development From an Ignition Kerenel in Laminar and Tur-bulent Homogeneous Mixtures", Institute of Aeronautics, 02-256 Warszawa. [39] Keck, J . C , Heywood, J .B. , and Noske, G., "Early Flame Development and Burning Rates in Spark Ignition Engines and Their Cyclic Variability", S A E Paper 870164, 1987. [40] Tagalian, J . , and Heywood, J .B. , "Flame Initiatation in Spark-Ignition Engine", Combustion and Flame 64: 243-246 (1986). [41] Bauwens, L. , "Flames in Lean Swirl Gaseous Mixtures", M.A.Sc. Thesis, U B C , 1986. [42] Evans, R .L . , and Cameron, C , "A New Combustion Chamber for Fast Burn Appli-cations", S A E Paper 860319, 1986. [43] Dymala-Dolesky, R., "The Effects of Turbulence Enhancement on the Performance of a Spark Ignition Engine", M.A.Sc. Thesis, UBC,1986. Bibliography 68 [44] Tippett, E .C. , "The Effects of Combustion Chamber Design on Turbulence, Cyclic ariation and Performance in an SI Engine", M.A.Sc. Thesis, U B C , 1989. [45] Cole, D.E. , and Mersky, W., "Mixture Motion- Its Effect on Pressure Rise in a Bomb: A New Look at Cyclic Variation", S A E Paper 680766, 1968. [46] Klomp, E.D. , and Deboy, G.R., "The Effects of Fluid Motions on Combustion in a Pre-Chamber Bomb", S A E Paper 860162, 1986. [47] Nakamura, H . , Ohinouye, T., et al, "Development of a New Combustion System ( M C A - J E T ) in Gasoline Engine", S A E Paper 790016, 1976. [48] Abramovich, G.N. , "The Theory of Turbulent Jets", MIT Press, 1963. [49] Witze, P.O., and Vilchis, F.R., "Stroboscopic Laser Shadowgraph Study of the Effect of Swirl on Combustion in a Spark-Ignition Engine", S A E Paper 810226, 1981. [50] Gatowski, G.A. , Heywood,J.B., and Deleplace, O , "Flame Photographs in a Spark Ignition Engine", Combustion and Flame Vol.56 pp71-81, 1984. [51] Matekunas, F . A . , "Modes and Cyclic Variability", A S E Paper 830337, 1983. [52] Murase et al, "Plasma Jet Ignition in Turbulent Lean Mixtures", S A E Technical Paper Series, International Congress and Exposition Detroit, Michigan, 1989. [53] Drain, L . E . , "Coherent and Noncoherent Methods in Doppler Optical Beat Velocity Measurement", J . Phys. D: Appl. Phys.5, 481, 1972. [54] Adrian, R . J . and Earley, W.E . , "Analysis of L D V Performance Using Mie Scattering Theory", Proceedings of Minnesota Symposium on Laser Anemometry, Blooming-ton, M N , Oct. 1975. Bibliography 69 [55] Milane, R., "Combustion and Turbulence Structure in a Closed Chamber with Swirl". [56] Mayo, J . , "The Effects of Engine Design Parameters on Combustion Rate in Spark-Ignited Engines", S A E Paper 750355, 1975. [57] Lyn, W.T. , and Valdmanis, E. , "The Application of High Speed Schlieren Photogra-phy to Diesel Combustion Research", The Journal of Photographic Science, vol.10, 1962. [58] Daneshyar, H . , and Hi l l , P .G. , "The Structure of Small Scale Turbulence and its Ef-fect on Combustion in Spark Ignition Engines", Prog. Energy Combust. Sci. Vol.13, P P 47-73, 1987. [59] Andrews, G.E. , Bradley, D. and Lwakabamba, S.B., "Turbulence and Turbulent Flame Propagation- A Critical Appraisal", Combustion and Flame 24, 285-304 (1975). [60] Dohring, K . , "The Relative Effects of Intake and Compression Stroke-Generated Turbulence in L C . Engines", M.A.Sc. Thesis, U B C , 1986. [61] Heywood, J .B. , "Internal Combustion Engines Fundamentals", McGraww-Hill, 1988. [62] Holder, D.W., and North, R.J . , "The Schlieren Methods" Notes on Applied Science No. 31, 1963. [63] "Laser Anemometer Systems", Thermo-Systems Inc., St. Paul, Minnesota, 1976. [64] Hung, J . "The Effects of Propane or Ethane Additives on Laminar Burning Velocity of Methane-air mixtures", U B C M.A.Sc. Thesis, 1986. Bibliography 70 [65] Pierik, R .J . , "Swirl Combustion in a Cylinderical Chamber", U B C , M.A.Sc. Thesis, 1987. [66] Namazian et al, "Schlieren Visualization of the Flow and Density Fields in the Cylinder of a Spark-Ignition Engine", S A E Paper 800044, 1980. Table 5-1: Maximum Pressure Rise and Time for Maximum Pressure Config. Dia.[mm] Max pres. [bar] T for M.pr.[ms] Quiescent Q 8.37 56 Offset 2 13.86 12.8 Offset 3 12.86 14 Offset 4 12.00 13 2 Radial 2 12.5 21.5 2 Radial 3 10.88 22 2 Radial 4 11.62 19 4 Radial 2 8.88 42 4 Radial 3 8.24 67 4 Radial 4 10.00 56 Table 5-2: Times for 10%, 50% and 90% Mass Burn Config. Dia. [mm] *io[ms] *so[ms] r.9o[ms] Quiescent Q 32.6 42.6 49.4 Offset 2 6 8.9 10.6 Offset 3 7.7 10.3 11.7 Offset 4 8 10.8 12.3 2 Radial 2 9.6 12.7 18.8 2 Radial 3 8.9 12.6 17 2 Radial 4 10.4 13.3 16.6 4 Radial 2 19.8 30 36.4 4 Radial 3 27.5 41 53.6 4 Radial 4 31.1 40 49.6 Fig.1.1: The Squish-jet Design, After Evans(6). .2-1: Growth of Jet Profiles(48). 75 n, za/t ec Velocity profiles In different sections of an axially symmetric submerged jet from Triipel'6 experimental data K . . M J * Profiles, Afte, Tn>pel «>d F o r t h — , From Abramovich's "The Theory of Turbulent Jets" (48). 76 Fig.2-3: Geometry of a Typical Measurement(63) Fig.2-4: Photogrnph of the Chamber Assembly. ARGON ION LASER COUNTER1 COUNTER2 ( — \ V f PC — 1 — Fig.2-5: Optical Arrangement of the LDV" System Fig.3-2: Photographs and Section Views of Ports 81 STORAGE T A N K 360 Kit CHARGE T A W K ?TECT*'C*J.Si6MAi A C H A I M S HOTDtf VAIV£  All? TAVlt - - C-D-I.S 12 V POWER SUPPLY PIPING AND CONTROL SYSTEM Fig.3-3: Control and Instrumentation Diagram of the Bomb. Fig.3-4: Calibration Curve for Propane Composition Analysis 83 NOTES: 1. CIRCUIT PATENT PENDING 2. CAPACITORS CB THRU CIZ ARE O.OI mfd +BATTERY tREOI • COIL (WHITE J POINTS (CREEN) -COIL (GREY) GROUND Fig.3-5: Schematic of Ignition System Electrical Wiring Fig.3-6: Photograph of Laser and Accessories 85 ANALOG -O INPUTS) Ext MUX T r i g Control Ext Clock Addr. Oecode Sample fi Hold 12-bit A/D c 3 IT Control Logic Clock Freq Div i d e r Address T Data S_4 I n t e r -rupt IBM PC BUS TPC00066 Fig.3-7: Block Diagram of the DT2811 I/O Board 220 140 1000 400 1470 1000 300 160 Fig.3-8: Schematic of the Schlieren Optical Bench 87 02 J Fig.3-9: Doppler Signal From the Photomultiplier Before and After Fil tering (*) Before filtering (b) After high pass filtering Fig.3-10: Section View of the Cyclone Particle Seeder 89 0 O w p u . Comparator IS V i Cony nor 4 V Fig.4-1: Basic Circuit Diagram of the Dual Mode Charge Amplifier 90 Fig.4-2: L D V Measurement Locations in the Central Plane of the Chamber Fig.5-4: Schlieren Photographic Sequences; 2-Radial Jets, Dia. = 2 mm. Al l Numbers Shown Are In Milliseconds. Fig.5-7: Schlieren Photographic Sequences; 4-Radial Jets, Dia. - 2 A l l Numbers Shown Are In Milliseconds. Fig.5-8: Schlieren Photographic Sequences; 4 -Radia l Jets, D i a . - 3 m m . A l l Numbers Shown Are l n Milliseconds. CO Fig.5-9: Schlieren Photographic Sequences; 4-Radial Jets, Dia. = 4 mm. ^ Al l Numbers Shown Are In Milliseconds. 2.50 H TIME (ms) Fig.5-10: Comparison of Charge Pressure Development: 2 Offset Jets o o TIME (ms) Fig.5-11: Comparison of Charge Pressure Development: 2 Radial Jets 2.50 H TIME (ms) Fig.5-12: Comparison of Charge Pressure Development: 4 Radial Jets 16.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 TIME (ms) Fig.5-13: Cyclic Variation Of Combustion Pressure 2 Offset .lets, Din. - 2 mm PRESSURE (bar) ,5-15: Cyclic Variation Of Combustion Pressure 2 Offset Jets, Dia. - 4 TIME (ms) Fig.5-16: Cyclic Variation Of Combustion Pressure 2 Radial Jets, Din..= 2 mm o PRESSURE (bar) ZOT Fig.5-18: Cycl ic Variation Of Combustion Pressure 2 Radinl Jets, Dia . = 4 mm 16.00 14.00 H 12.00 H 10.00 -\ o § 8.00 H I/) in a 6.00 H 4.00 H 2.00 0.00 0.00 10.00 30.00 TIME (ms) 60.00 Fig.5-19: Cyclic Variation Of Combustion Pressure 4 Radial Jets, Dia. = 2 m ni o co 16.00 TIME (ms) Fig.5-20: Cyclic Variation Of Combustion Pressure 4 Radial Jets, Dia. = 3 mm 16.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 TIME (ms) Fig.5-21: Cyclic Variation Of Combustion Pressure Rise. 4 Radial Jets, Dia. = 4 mm TIME (ms) Fig.5-22: Combustion Pressure Rise: 2 Offset Jets 16.00 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 TIME (ms) Fig.5-23: Combustion Pressure Rise: 2 Radial Jets TIME (ms) Fig.5-24: Combustion Pressure Rise: 4 Radial Jets 16.00 TIME (ms) Fig.5-25: Comparison Of Combustion Pressure Rise Dia. = 2 mm o » o o € i Offset Jets 2 Rodial Jets 4 Radial Jets 0.00 10.00 20.00 - i 1 " 1 1 r 30.00 40.00 50.00 TIME (ms) i 1 1 r 60.00 70.00 Fig.5-26: Comparison Of Combustion Pressure Rise Dia. = .1 mm PRESSURE (bar) 80.00 TIME (ms) Fig.5-28: Mass Fraction Burned : 2 Offset Jets 80. TIME ( m s ) Fig.5-29: Mass Fraction Burned : 2 Radial Jets Fig.5-31: Comparison of M . F . B . For Different. Configurations, Fig.5-32: Comparison of M .F .B. For Different Configurations, Din. - 3 mm 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 TIME (ms) Fig.5-33: Comparison of M . F . B . For Different Configurations, Dia . -• 4 mm u en HMC r/R - 0M3. M t «T Crnnfctr o I S i » IE m ilo tin jjo 3o m «J» Jo Hue (m) r / R - n « 6 . Ajrii of M A u en °°" 1 5 2 t?o IE jJo tin lh 3o X «4> »A> TWC (m») r/R • ( M l M a e» M Fig.5-34: U-Component velocity on Axis of Jet, Offset Configuration, Din. - 2 mm r/R • 0.89S. tab ot Chombar r/R • O.MA. tail of J*t JO.CO a t Ven "I w ilo ik no i2o ilo dt M w IE <E~ TOrt (m.) 90.00 m . § . 4 0 . 0 0 -t > 3000 -Ven v en 1 to ilo IE no ><o no J» >Jo «o «o <i» TIME ( I T H ) r/R • 0665. tail of M r/R • 0.266. tall of Jd TlUt I n ) 3» 3M 400 440 480 S » o.oo r — i 1 1 1 i i r i l l r~ _ •0 to i n iao n o 140 n o 3» MO 400 440 4«o s » TlUt (ms) Fig.5-35: V-Component velocity on Axis of Jet, Offset Configuration, Dia. = 2 mm. Fig.5-36: U-Component. velocity on Axis of Chamber, Offset Configuration, Dia. = 2 mm. r/R - 0.440. Axl« of CHombw r/n m 0.89ft. A** o* Oomb«r Fig.5-37: V-Component Velocity on Axis of Chamber, Offset Configuration, Dia.. = mm. r/R • 0 89B. A«li c« M r/R • 0.440 Aib of Jal Fig.5-38: U-Component Velocity on Axis of Jet, Offset Configuration, Dia . = 4 mm. r/R - 0B9B * » * <* • WOO -•7 I > 30 OO I IMS fOSO O.OB Ken v en i ,], .lo *lo t!o r l . do sit. « Jo •*> a» TIME (ml) r/R - 0 « S » * » * ot J»t 0 Ven w en . A i -5. >lo J» . J . J. J . A 4. «i. 'T» TIMt (im) r/R «• 0.440 Arts of M tlMC (rm)) r/R - O.MS fcrta ol J * 440 400 MO Fig.5-39: V-Component Velocity on Axis of Jet, Offset Configuration, Dia. = 4 mm. r/R • 0891 Avh of Chamber r/R - 0.440 A»l» of OomNr en "5 so ilo IE im 3*0 no 3o j»e ' ' ' r/R • 0 MS tew, of OsMxr i A .Jo . i , n, >lo X ,» no TIME (mi) Figure 5-40: U-Component Velocity on Axis of Chamber, Offset Configuration, D 4 mm. r/R - 089*. ««>• ot CfwmMr TMC (m<) r/R - 0S8S. Art Bl CWrneor nut (mt) r/R - 0.440. A*t« of Cftomeor W O O TIME (rm>) Fig.5-41: V-Component Velocity on Axis of Chamber, Offset Configuration, Dia. = A mm. Fig.5-42. U-Component Velocity on Axis of Jet, 2 Radial Configuration, Dia . = 2 133 r/R - 0 096. Aril of Cfwnpf r/R » 0.440. fell of chombar "3 2 Ho lift no iE X do jio *3o «E 3 O M O TWC (rm) G "5 2 iE IE iE 3> iE J » jio «E 3o «E~ Tmr (mt) r/R • 0663. Mo of Crwmbor r/R - O.Mt. Mo of Oomkor * — ilo IE iE iE iE jE 3o no iE «lo »lo 0 40 oo TIMt (m.) Fig.5-44: U-Component Velocity on Axis of Chamber, 2 Radial Configuration, Dia. = 2 mm. r/B « 0098. of Oombor r/f» - 0.440. AJU* of Chorntw Fig.5-4 5 V-Component. Velocity on Axis of Chamber, 2 Radial Configuration, Dia . = 2 mm. r/R - 0.898. "fc* of J«l r/R - 0 865. taia of Mt SM r/R - 0.440. tah of J«l r/R • 0.366. Art ot JM, Fig.5-46: U-Component Velocity on Axis of Jet, 4 Radial Configuration, Dia. = 2 mm. Fig.5-47: V-Component Velocity on Axis of Jet, 4 Radial Configuration, Dia. = 2 mm. r/R • 0.998. AXHJ ol Oomtwr r/H - 0.440. 'Ajtft of Ctartibtr 0>M G u en TWC (ma) r/H - 0.S6S. Mt a) CftortMr trUt (mt) G ~tlo iC iS 3o JC 3o So 3o *C «To sis IMC (mt) r/N • UO.2W.Ailt of Ctiomtjor G 0 ~40 BO TJo IM »0 MO 3»" MO 3o IC ••> TWC (mt) Fig.5-48: U-Component Velocity on Axis of Chamber, 4 Radial Configuration, Dia . = 2 mm. r/R • 0099. >t»t of Ovymfer r/R • 0410, An!* of Chamber TMC (ms) ~ 5 M Hii I S no >4o no l ie J4O «Jo *le « k 5» IMC (ms) r/R - 0 « « 5 . "»«•• 0» r/R • 0 . 2 M . •»*» of Cke ra to i.4000 f * 3000 G Ven en "5 SO ito J00 Ho no 3M jJo *00 **0 <2o sjo n u t (ma) G Ven « en o «b 00 IJo l5 100 »lo 35 370 MO 400 4I0 tm TMC (ms) Fig.5-49: V-Component Velocity on Axis of Chamber, 4 Radial Configuration, Dia. = 2 mm. Fig.5-51: Distribution of Tangential Velocity, Offset Configuration, Dia. = All Numbers Are In ms 4 mm. Fig.5-52: Distribution of Tangential Velocity, 2 Radial Configuration, Dia. = All Numbers Are In ms 2 mm. Appendix A D E R I V A T I O N OF J E T F L O W P A R A M E T E R S In general a jet may be assumed to issue from a circular orifice into a stream of uniform velocity or into a quiescent fluid, with a uniform outlet velocity. For turbulent flow of high Reynolds number, the flow becomes fully developed shortly after the orifice outlet. The longitudinal fluctuating velocity is expressed by: .du u ~ l-dy The transverse fluctuating velocity is: , db v ~ — dt where / / du —v ~ u ~ I— dy and b is the transverse length or width of jet. But db T t ~ U m , um is axial velocity at the center. On the other hand, rate of growth of the jet can be expressed by: db db dx db dt dx dt dx m Comparing the above expressions, the rate of growth of jet width is obtained as db —— = const dx 144 Appendix A. DERIVATION OF JET FLOW PARAMETERS 145 Therefore, integration gives b = cx (A.31) where c is a constant. Usually the half width bi/2 is considerd as a characterstic length for jet flows. The velocity variation along the axis was derived from the conservation of momentum equation. In this derivation, it is assumed that the pressure in the jet is equal to the pressure in the surrounding space. Hence, For a round jet; the condition for constant total momentum is expressed [48] as: where um : velocity at the center of a given section x : distance of a section from the origin y : radius b : radius of outer boundary of jet From the universal nature of the velocity profiles, it was found that the dimensionless velocity is a function of dimensionless coordinates. u u, •m and Therefore for an axially symmetric jet, const (A.32) u. X Appendix A. DERIVATION OF JET FLOW PARAMETERS 146 On the other hand for a plane jet; .2 _ fb/a,_.,.. x2dV ux m / (u/um) — = const Jo x [h'x< / \2dv * I (u/um) — = const Jo x Therefore const um = — p - A.33) y/X The laws governing counter flows are basically the same as long as proper directional signs are taken care of in the analysis. Mutual interactions of oppositely oriented jets result in flow fields different from that of a single jet stream. Appendix B PRINCIPLES OF T H E SCHLIEREN SYSTEM The schliern system is a device used to visualize flows with density gradient by forming varying intensity images on the film plane. By passing light through the medium, flow can be visualized from the change in refractive index which is a linear function of the density. If the fluid in the test section is completely uniform in density, the intensty of light on the film plane will also be uniformly lit or uniformly dark depending on the amount of light intercepted by the knife edge. However, if there is density gradient in the flow light beams at different points will deflect from their original path, and the obstruction will intercept more light from some points than from others giving rise to light and dark regions on the screen. L x * oerlecUorv o* c> beam \n a Tes^ . b e ^ o n A n incoming ligh* my is deflected through an angle a. The deflection for small density 147 Appendix B. PRINCIPLES OF THE SCHLIEREN SYSTEM 148 gradients is L dn Lf3 dp where L is width of the flow field, p is local fluid density, p, is reference density, and n is index of refraction. If it is assumed that the point source is of rectangular shape with a width of b and height h units where the focal lengths of Lensl and Lens2 are equal to f, illumination IQ without interference of the knife edge will be _ Bbh  Io~rnJp where B = luminance and m = magnification of image on screen [62]. If a is the height after the knife edge is adjusted, the illumination on the film plane will be 56a When the light through the test section is defelected by an angle 8a, illumnation on the film plane becomes . . BbSa Then, the contrast C is SI fSa , and the sensitivity S is S = ^~ = £ Holder [62] showed that the maximum range of displacement of the image of the source for retention of sensitivity is equal to the height of the image, and the corresponding range of angular deflection is 8a = ~. Appendix C M I X T U R E P R E P A R A T I O N A N D C A L C U L A T I O N In a chemically balanced ratio, the reaction equation for propane and air is given by: C3H8 + 5(02 + 3.76N2) —> 3C02 + 4H20 + 18.8iV2. The sum of the partial pressures of the reactants is equal to total container pressure, i.e. Ppropane + Pair = Ptotal- Hence the partial pressure of propane is P 1 1 x propane x x Ptotal 1 + 5(1 + 3.76) 24.80 Since the cylinder working pressure is 17.3 MPa , the pressure before accidental com-bustion in the bottle may not exceed 2.47 M P a with a safety factor of 7.In this work the total pressure chosen was 2 MPa. At stoichiometry, the partial pressure of propane in the bottle was P t o t a l 2000fcPq 8 n R . , , p Ppropane = = ^ g = 80.645fcPa . The desired equivalence ratio was then used to determine the final partial pressure of propane. To prepare a premixed mixture, an evacuated bottle was filled with the calculated partial pressure of propane and the bottle temperature was allowed to return to ambient. The bottle was then filled slowly with extra dry air to 2000 kPa. The mixture was allowed to homogenize before it was analysed by gas chromatogrphy. 149 Appendix D GAS C H R O M A T O G R A P H A Hewlett Packard model 5750B Research Gas chromatograph was used for mixture strength determination. The unknown composition sample of a known volume was in-jected into a port in the machine where the sample would travels through a thin, long column which separates the hydrocarbons and air. To each component a specific reten-tion time was identified. This was the time taken by the component to travel through the column depending on the column temperature [64]. When a component goes through a Wheatstone bridge, a peak was produced on the recorder due to the residence time where the area under this peak was proportional to the volume injected. The volume percent of the fuel was then determined from the caliberation curve for the fuel used. The following procedure was followed for mixture analysis. a. The chromatograph, the bridge current in machine #2, the electronic panels and the oven were turned on and left to warm up for some time. b. Oven temperature for propane was set to 130°C and the oven remained closed. c. A vial was evacuated and flushed with the unknown mixture a few times; then a sample of a sample of the mixture was prepared in the vial. The mixture pressure in the vial was left higher than atmospheric so that the mixture would leak outward and nint diluted by ambient pressure if there was any leakage through the rubber seal of the vial. d. The integrator power was turned ON and the zero adjusted on the bridge. The chart was set to A U T O . e. A one milliliter sampling syringe was flushed a few times with the unknown mixture 150 Appendix D. GAS CHROMATOGRAPH 151 ratio. Exactly one milliliter of the mixture at atmospheric pressure was injected in the " B " column of the oven. f. The S T A R T / S T O P button was pressed on the integrator immediately after injec-tion. The sample was tested several times until the consistency and accuracy satisfied the operator. Finally the volume percent of the fuel was used to determine the equivalence ratio. Appendix E E Q U I P M E N T SPECIFICATIONS 1. C O M B U S T I O N C H A M B E R Type: Cylindrical, Multi-jet ports, Central Spark. Chamber Size: Length = 43 mm Diameter= 80.8 mm Material: Body: 304 Stainless Steel Window: Quartz Maximum Pressure: 200 bars 2. P I E Z O E L E C T R I C P R E S S U R E T R A N S D U C E R : Kistler Instrument Corporation Model: 6123 A2 Type: Quartz Pressure Transducer, Miniature Profile Range: 0...200 bar Calibrated Partial Rang e 0...20 bar Overload 300 bar Sensitivity -15.95 pc/bar Natural Frequency > 110 kHz Frequency Response ± 1 % 6 kHz Linearity < ±1.0 %FSO Hysteresis < 0.8 %FSO Acceleration Sensitivity 152 Appendix E. EQUIPMENT SPECIFICATIONS 153 Axial < 0.003 bar/g Transverse < 0.0002 bar/g Shock Resistance 2000 g Thermal Sensitivity Shift 20...100°C < +0.5 % 200..±50°C < ± 1 % 20...350°C < ± 3 % Calibrated in Range 20...300 °C Operating Temperature Range -196...350 °C Transient Temperature Error < 0.02 bar Insulation: at 20°C > 10 Teraohms at 350°C > 10 1 0 Ohms Ground Insulation > 108 Ohms Weight 9.5 g 3 A B S O L U T E P R E S S U R E T R A N S D U C E R : Sensym Model: LX1830AZ Type: Absolute, mid-pressure range, signal conditioned, strain guage pressure transducer Operating Pressure: 0...300 psia Maximum overpressure:300 psi Offset Calibration: 2.5+0.30 V Offset Shift With Temperature: ±1.10 %FS Span Shift With Temperature: ±1.10 %FS Offset Stability: ±1.0 %FS Appendix E. EQUIPMENT SPECIFICATIONS Span Sensitivity Calibration: 33.3±0.67 m V / P s i Span Stability ±0.3 %FS 4 C H A R G E A M P L I F I E R : Kistler Instrument Corporation Model: 5004 Type: Dual Mode Amplifier Scale Settings: 12 Steps in 1 - 2 - 5 Sequences Selectable Scales for Transducer Ranges: 0.01 to 0.11 Pc/or mV per M . U . 100 to 500,000 M . U . / V 0.1 to 1.1 Pc/or mV per M . U . 10 to 50,000 M . U . / V 1 to 11 Pc/or mV per M . U . 1 to 5,000 M . U . / V 10 to 110 Pc/or mV per M . U . 0.1 to 500 M . U . / V 100 to 1100 Pc/or mV per M . U . 0.01 to 50 M . U . / V Input Voltage < ±125 V Output, Unlimited Short Circuit Proof Voltage ±10 V Current ± 5 mA Impedance 100±0.5 Ohms Calibration Input Sensitivity 1±0.5 % Time constant, r = Rg * Cg Setting, Long up to 100,000 s Setting, Medium 1 to 5,000 s Setting, Short 0.01 to 50 s Accuracy of Ranges < ± 1 % Noise Input (Cable) 0.01 Pcrm,/nf Drift (Due to Leakage Current) < ±0.03 Pc/s Appendix E. EQUIPMENT SPECIFICATIONS Operating Temperature Range 0 to 350 °C Power 100 to 130 V A C 50 to 60 Hz, 8 V A 5 IGNITION S Y S T E M : Heath Company. Model : CP-1060 Type : Capacitive Discharge Ignition System Input : 12-14 V D C Maximum Input: 18 V D C Trigger Source: Ignition Points Capable of Switching 40 ohm to Ground Within Volt. Output Duration:0.6 ms at Low Voltage and rpm. 0.4 ms at Normal Voltage and rpm. 0.2 ms at High Voltage and rpm. Maximum Allowable Ignition: Point Contact Resistance: 7 ohms Minimum Allowable Ignition: Point Shunting Resistance: 100 ohms 6 V A C U U M P U M P : Cenco Model : Hyvac 7 Type : Rotary Vane Pump Pump Speed: 525 rpm Capacity at Atmospheric Pressure: 79 l/min Capacity at 1 micron of Hg: 35 l/min guaranteed Vacuum : 0.1 micron of Hg Appendix E. EQUIPMENT SPECIFICATIONS Merit Factor : 44% Motor Power : 0.25 HP 7 SOLENOID V A L V E : V E R S A Model : ESM-2301-40 Type : 2-Way, Direct-Solenoid Actuated Valve Maximum Pressure: 200 psi Minimum Orifice Between Ports: 1/16 in Oprerating Voltage: 12 V D C Power : 10 W 8 SOLENOID V A L V E : ASCO Electric Ltd, Model : 8321 A l Type : 3-Way Direect-Solonoid Actuated Valve Operating Pressure: 10 - 200 Psi Minimum Orifice Between Ports: 9/16 in Pipe Size: 1/4 in Operating Voltage: 12 V D C Power: 9.7 W 9 A I R V A L V E : ASCO Electric Ltd. Model : P210c94 Type : Air-Operated Spring Valve Working Pressure: 30 - 125 Psia Pipe: 0.5 in Orifice 5/8 in Appendix E. EQUIPMENT SPECIFICATIONS 10 HIGH SPEED C A M E R A : Red Lake Laboratories, Inc. Model : H Y C A M K2001 Type : 16 mm, High Speed Motion Picture Camera Frame Rate: 100 - 8500 fps 11 L A S E R : Spectra Physics Model : 165 Type : Argon-Ion Maximum Power: 5 watts Power Supply: 115 Volts (AC) Cooling Water Supply Pressure: > 30 Psia 12 A N A L O G A N D DIGITAL I /O B O A R D : Data Translation, Inc. Model : Dt 2811 P G H Number of Analog Inputs: lQSe/BDi Input Ranges: Jumper Selectable Output Data Codes: Offset Binary Gain : 1 Common Mode Input Voltage, Maximum: ±11 V Resolution: 12 bits System Accuracy: Within ±0.03% FSR Channel Acquisition Time: 20 ps (average) A / D Conversion Time: 30«us A / D Convertor Thoroughput: 20 kHz Number of Output Channals: 2 Input Data Codes: Offset Binary Appendix E. EQUIPMENT SPECIFICATIONS D A C Thoroughput: 50 kHz max. Digital Input Lines: 8 Digital Output Lines: 8 Connector: J l , 3M-3596-5002, 3m-3425 Compatible Bus: I B M P C / X T / A T Number of I /O Registers: 8 

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.831.1-0098487/manifest

Comment

Related Items