Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of turbulence enhancement on the performance of a spark-ignition engine Dymala-Dolesky, Robert 1986

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

Item Metadata

Download

Media
831-UBC_1987_A7 D95.pdf [ 11.56MB ]
Metadata
JSON: 831-1.0097090.json
JSON-LD: 831-1.0097090-ld.json
RDF/XML (Pretty): 831-1.0097090-rdf.xml
RDF/JSON: 831-1.0097090-rdf.json
Turtle: 831-1.0097090-turtle.txt
N-Triples: 831-1.0097090-rdf-ntriples.txt
Original Record: 831-1.0097090-source.json
Full Text
831-1.0097090-fulltext.txt
Citation
831-1.0097090.ris

Full Text

THE EFFECTS OF TURBULENCE ENHANCEMENT ON THE PERFORMANCE OF A SPARK-IGNITION ENGINE By ROBERT DYMALA-DOLESKY Mgr.Inz., The Technical University of Szczecin, Poland, 1980 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n THE FACULTY OF GRADUATE STUDIES Department of Mechanical Engineering We accept t h i s t h e s i s as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November, 1986 © ROBERT DYMALA-DOLE SKY 3 2 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 M c x W ^ c c ^ IB>vcf>'\sev\n The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 DE-6(3/81) ABSTRACT An attempt has been undertaken to enhance turbulence in an S.I. engine at the final stage of the compression stroke, without affecting the intake process. The method employed to control the turbulence level made use of an original design called the squish-jet combustion chamber. The design had potential to generate jets in the chamber before CTDC and thus create dramatically different turbulent flow patterns. Natural gas, a slow burning fuel, was used for performance tests, and different levels of turbulence were expected to markedly affect the combustion process. A flow visualization experiment was performed under conditions similar to a motored engine. As a result, the jet development in the squish-jet type combustion chamber was documented. A new type of a flat cylinder head, and a set of squish-jet pistons were designed and manufactured. Experiments conducted on the redesigned Ricardo Hydra, single cylinder research engine, evaluated the influence of the squish-jet chamber on the mixture motion and the engine perform-ance over a wide range of operating conditions. The jet velocities were measured with a hot wire probe located in the piston bowl, and turbulence parameters with a probe inserted through a cylinder head. The squish-jet design was evaluated for 6 different configurations. As a result i t has been established that the squish-jet design does not create jets strong enough to dramatically enhance the turbulent flow field. The design, however, diminished the squish effect which is shown to be very important for the middle part of flame development. The simple squish design produces faster burning rate in the first half of - i i -the combustion process and develops the highest peak pressures. Variabilities of both cyclic IMEP and peak pressure are found to be unaffected by the presence or absence of strong squish motion. This suggests that the most important phase of combustion for the cyclic variation is the i n i t i a l stage of the flame development. A comparison of ensembled pressure signals between combustion chamber designs, conducted at RAFR=1.00 and at RAFR=1.25 shows less dispersion in the latter case. It appears that at lean operation mixture motion influences combustion process to a lesser degree than at stochiometric conditions. - i i i -TABLE OF CONTENTS Page ABSTRACT i i TABLE OF CONTENTS iv LIST OF TABLES vi LIST OF FIGURES v i i NOMENCLATURE xi ACKNOWLEDGEMENT x i i i 1. INTRODUCTION 1 1.1 General Discussion 1 1.2 Mixture Motion Effects in I.C. Engines 3 1.3 Review of Previous Work on Turbulence in I.C. Engines .. 7 1.4 Objectives and Scope of this Work 11 2. EVALUATION OF THE SQUISH-JET DESIGN 13 2.1 Squish-Jet Design 13 2.2 Flow Visualization Experiment 15 3. EXPERIMENTAL APPARATUS AND METHOD 18 3.1 Experimental Method 18 3.2 Experimental Apparatus 19 3.3 Instrumentation and Data Acquisition System 22 3.3.1 Flow Measuring System 22 3.3.2 Instrumentation for Firing Tests 23 4. DATA ANALYSIS 25 4.1 Flow Field Analysis 25 4.1.1 Experimental Conditions 25 4.1.2 Analytical Procedure 26 4.2 Firing Tests Analysis 27 4.2.1 Test Conditions 27 4.2.2 Analytical Procedure 28 5. DISCUSSION AND EXPERIMENTAL RESULTS 31 5.1 Flow Experiments 31 - iv -CONTENTS (Continued) Page 5.1.1 Results of Jet Velocity Measurements 31 5.1.2 Results of Turbulence Measurements 34 5.2 Firing Test Results 36 5.2.1 General Performance Parameters 37 5.2.2 Firing Pressure Analysis 37 5.3 Discussion of Experimental Technique 40 5.3.1 Flow Measuring Technique 40 5.3.2 Performance Evaluating Technique 45 6. CONCLUSIONS AND RECOMMENDATIONS 47 6.1 Conclusions and Observations 47 6.1.1 Introduction 47 6.1.2 Observations 47 6.1.3 Conclusions 48 6.2 Recommendations 49 REFERENCES 51 APPENDIX A - SQUISH-JET ANALYTICAL MODEL 55 APPENDIX B - MASS FRACTION BURNED COMPUTER PROGRAM 63 - v -LIST OF TABLES Page Table 1. Rlcardo Engine Specifications 72 2. Hot Wire Probe and Anemometer Specifications 72 3. Natural Gas Properties 72 4. Comparison of the Engine Performance for Different Piston Geometries at WOT, 3000 RPM, AFR=1.00 and MBT Timing 73 5. Comparison of the Engine Performance for Different Piston Geometries at WOT, 3000 RPM, AFR=1.25 and MBT Timing 74 6. Comparison of the Engine Perforamnce for Different Piston Geometries at WOT, 2100 RPM, AFR=1.00 and MBT Timing 75 7. Comparison of the Engine Performance for Different Piston Geometries at WOT, 2100 RPM, AFR=1.25 and MBT Timing 76 8. Comparison of the Engine Performance for Different Piston Geometries at WOT, 1200 RPM, AFR=1.00 and MBT Timing 77 9. Comparison of the Engine Performance for Different Piston Geometries at WOT, 1200 RPM, AFR=1.25 and MBT Timing 78 10. Comparison of the Engine Fuel Consumption for Different Piston Geometries, at Part Load, BMEP=2.5 bar, 2000 RPM 79 11. Ensembled Peak Pressure and Standard Deviation of Cylic Peak Pressure for Different Piston Geometries, at 3000 and 2100 RPM 80 12. Ensembled Peak Pressures and Standard Deviation of Cyclic Peak Pressure for Different Piston Geometries, at 1200 RPM .. 81 13. Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries, at 3000 and 2100 RPM 82 14. Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries, at 1200 RPM 83 - vi -LIST OF FIGURES Page Figure 1.1 Obstructions on the Valves Generating Swirl 84 1.2 Squish Combustion Chamber 84 2.1 Squish - Jet Design 85 2.2 Flow Visualization Experimental Set up 86 2.3 Photograph of the Plexiglass Model 87 2.4 Photographs of the Jet Development 88 2.5 Photographs of the Jet Development 89 3.1 Cross-section of the Redesigned Ricardo Engine 90 3.2 Photographs of the Cast and Machined New Cylinder Head 91 3.3 Photograph of the Ricardo Hydra Engine 92 3.4 Photograph of the New HW Probe 93 3.5 Schematic of the Linkage Mechanism 94 3.6 Photographs of the Connecting Rod with the Linkage Mechanism and the Inside Wiev of the Piston 95 3.7 Photographs of the Probe Position in the Piston and the Connection Between Connecting Rod and the Piston 96 3.8 Schematics of the Data Acquisition System 97 4.1 Schematic of the Cases for Flow Experiments 98 4.2 Hot Wire Probe Positions Across the Piston Bowl 99 4.3 Schematic of the Cases for Performance Tests 100 5.1 Comparison of Cyclic Jet Velocity Profiles for Pistons 4 and 5 at 3000 RPM 101 5.2 Comparison of Cyclic Jet Velocity Profiles for Pistons 5 and 6 at 3000 RPM 102 5.3 Comparison of Cyclic Jet Velocity Profiles for Pistons 1, 2, and 7 at 3000 RPM 103 - v i i -LIST OF FIGURES (Continued) Page 5.4 Comparison of Ensembled Jet Velocity Profiles for Pistons 1, 2 and 7 at 3000 RPM 104 5.5 Comparison of Ensembled Jet Velocity Profiles for Pistons 3, 4 and 5 at 3000 RPM 105 5.6 Comparison of Ensembled Jet Velocity Profiles for Pistons 5 and 6 at 2100 RPM 106 5.7 Comparison of Two Cycles of the Jet Velocity Profiles Measured in Piston 5 at 300 RPM 107 5.8 Comparison of Cyclic Jet Velocity Profiles Measured in Piston 5 for Three Engine Speeds: 3000, 2100, 1200 RPM 108 5.9 Comparison of Cyclic Jet Velocity Profiles Measured in Piston 5 for Three Engine Speeds: 3000, 2100, 1200 RPM - 1 Graph 109 5.10 Comparison of Ensembled Jet Velocity Profiles Measured in Piston 5 at Speeds: 3000, 2100 and 1200 RPM 110 5.11 Comparison of Cyclic Jet Velocity Profiles Measured in Pistons 5 and 6 I l l 5.12 Comprison of Ensembled Mean Velocity Profiles for Four Piston Geometries at 3000 RPM 112 5.13 Comparison of Ensembled Turbulent Fluctuations for Four Piston Geometries at 3000 RPM 113 5.14 Comparison of Ensembled Mean Velocity Profiles for Piston 1, Top Probe Position, at 3000, 2100 and 1200 RPM 114 5.15 Comparison of Ensembled Mean Velocity Profiles for Piston 4, Top Probe Position, at 3000, 2100 and 1200 RPM 115 5.16 Ensembled Turbulent Fluctuations for Piston 4, Top Probe Position, at 3000, 2100 and 1200 RPM 116 5.17 Comparison of Ensembled Mean Velocities for Piston 4, at 3000 RPM, Measured Across the Piston Bowl 117 5.18 Comparison of Ensembled Turbulent Fluctuations for Piston 4, at 3000 RPM, Measured Across the Piston Bowl 118 5.19 Comparison of Ensembled Mean Velocities for Piston 1, at 3000 RPM, Measured Across the Piston Bowl 119 - v i i i -LIST OF FIGURES (Continued) Page 5.20 Comparison of Ensembled Turbulent Fluctuations for Piston 1, at 3000 RPM, Measured Across the Piston Bowl 120 5.21 Comparison of Cyclic Velocity Profiles for Pistons 1 and 4 at 3000 RPM and Bottom Probe Position 121 5.22 Comparison of Ensembled Turbulent Fluctuations for Pistons 1 and 4 at the Bottom of the Piston Bowl at 3000 RPM 122 5.23 Comparison of the Ensembled Velocity Profiles in the Middle of the Piston Bowl for Pistons 1 and for 4, at 3000 RPM 123 5.24 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 3000 RPM and AFR=1.00 124 5.25 Comparison of the Ensembled Pressure Traces for Piston Geometries 3, 4 and 5 at 3000 RPM and AFR=1.00 125 5.26 Comparison of the Ensembled Pressure Traces for Piston Geometries 2, 5 and 7 at 3000 RPM and AFR=1.00 126 5.27 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 3000 RPM and AFR=1.25 127 5.28 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 2100 RPM and AFR=1.00 128 5.29 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 2100 RPM and AFR=1.25 129 5.30 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 1200 RPM and AFR=1.00 130 5.31 Comparison of the Ensemble pressure Traces for Five Piston Geometries at 1200 RPM and AFR=1.25 131 5.32 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 3000 RPM and at Two AFR=1.00 and 1.25 132 5.33 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 2100 RPM and at Two AFR=1.00 and 1.25 133 5.34 Comparison of Mass Fraction Burned for Pistons 3 and 7 at 3000, 2100 and 1200 RPM 134 5.35 Comparison of Mass Fraction Burned for Five Piston Geometries at 3000 RPM and AFR=1.00 135 5.36 Comparison of Mass Fraction Burned for Five Piston Geometries at 3000 RPM and AFR=1.25 136 - ix -LIST OF FIGURES (Continued) Page 5.37 Schematic of the Suggested Experimental Set up for the Evaluation of Vibration Effects on Turbulence Measured with HWA 137 5.28 Geometry of the Engine for Analytical Evaluation 138 5.29 Schematics of the Geometry for the Jet Calculation 139 - x -NOMENCLATURE A a AjCe) CH TOT BBDC BDC BTDC BMEP BSFC C CL CR D d d n H h HWA i IMEP L MBT m \ P R RPM RAFR S Area m2 Heat transfer coefficient Instantaneous area of squish inflow, mm2 Surface of inflow to the bowl, mm2 Area of the leakage, mm2 Area of the contact of gas in volume with the cylinder head, mm2 Total contact area of gas with the cylinder head, mm2 Before bottom dead center Bottom dead center Before top dead center Brake mean effective pressure Brake specific fuel consumption Piston velocity, m/s Clearance height, mm Compression ratio Engine bore diameter, mm Piston bowl diameter, mm Channel In the piston diameter, mm Depth of the piston bowl, mm Bowl height, mm Hot wire anemometry Particular cycle Indicated mean effective pressure, kPa Length of the connecting rod, mm Ignition - minimum advance for best torque mass, kg Leaking mass of the gas Pressure, kPa Crank radius, mm Revolutions per minute Relative air fuel ratio Distance of the piston from the top, mm - xi -U(i,t) Instantaneous velocity, in cycle i , time t U(i,t) Mean velocity V Volume, m3 T^OT Total cylinder volume Vj Volume above the squish area Vg Swept volume, m3 U Horizontal velocity, m/s U„ Squish velocity calculated with heat transfer effects, m/s U . Squish velocity calculated with leakage, m/s a) Angular velocity of teh engine 1/s v Vertical velocity, m/s 9 Crank angle p Density, kg/m3 v Kinematic viscosity, m2/s x Ratio of specific heats - x i i -ACKNOWLEDGEMENT I would like to express my sincere gratitude to my supervisor Dr. R.L. Evans for his help and encouragement throughout the course of this work. I would also like to thank A. Jones for his work on the data acquisition system and L. Drakes whose machining talent contributed greatly to the excellent performance of the experimental equipment. Further thanks are due to Professors Hauptmann and Hil l for their stimulating contributions in various discussions. - x i i i -1. CHAPTER 1  INTRODUCTION 1.1 General Discussion At the present time internal combustion (I.C.) engines have under-gone a history of improvements, almost a century long. Changes in economic and environmental conditions in the last two decades, however, have brought issues such as environmental protection and global energy conservation to the focus of public attention. This has dramatically increased the need for a better understanding of fundamental processes in I.C. engines. The area of particular importance is lean operation. The last two decades of research on combustion have established that the advancement of I.C. engines depends on their ability to operate with a combustion process called fast and lean burning. This requirement appears contradictory because a lean mixture burns slower. Lean opera-tion offers two important benefits: higher efficiency, a consequence of lower temperature and at part load less pumping loss, and lower levels of exhaust emissions. A practical engine operation in the lean region introduces, however, large cyclic variations and misfiring. Fast burning makes lean operation viable practically by decreasing these negative effects. In addition fast burning increases efficiency due to approach-ing closer to a constant volume combustion and improves knock limits. Higher knock limits can lead to an increased compression ratio and consequently higher thermal efficiency. Progress towards fast, lean burning is particularly important for the viability of slow burning fuels, like natural gas. Their potential as alternate fuels is being restricted by their considerably slower 2. burning rates in comparison to gasoline, which results in a loss of power of engines converted from gasoline fuel to natural gas. It is recognized that progress in the direction of fast, lean burn-ing can be achieved by an optimization of three key factors: mixture motion, ignition system and chemical changes affecting reaction kinetics. In most cases the factors are interrelated. The most important and challenging of them is mixture motion. This challenge is a consequence of the extreme complexity of the problem. The processes in I.C. engines contain nearly every conceivable fundamental problem from thermodynamics and fluid mechanics interacting with chemistry. They are correlated with each other and strongly depend on the combustion chamber environment. Neither the turbulent flow field nor combustion are well understood. This makes i t a first class problem in science with implications for many areas not related to I.C. engines. Experimental data collected by different research groups, to date, show trends linking the flow field with combustion. However, what makes comparison of published experimental results very difficult is the lack of consistency with respect to experimental techniques used by different research groups. Consequently i t is often questionable to compare absolute results published. In addition there are some serious doubts whether what is frequently described as turbulence is not a result of the experimental technique and method of data analysis used. The question of experimental procedure will be dealt with in detail, in the next chapters. What has been firmly established is that qualitatively a higher level of turbulence increases turbulent flame speed and burn rate [1]. There are, however, many interesting problems related to the speci-fics of this interaction which are not understood. We don't know at what 3. point of flame development turbulence becomes really important, what is the exact role of scales, whether there is an optimum turbulence level in particular thermodynamic conditions for faster burning, at what point turbulence starts to affect the flame negatively, or how the flame affects turbulence in front of i t . These are only some of the questions which need to be answered. The existing theories link turbulence to two mechanisms. The first perceives turbulent fluctuations as element wrinkling and by this increasing surface of the propagating flame [2]. The other theory sees turbulence as the mechanism which increases the rate of unburned charge entrained by intermittent regi ons of activity In the flame front [3]. Some experiments suggest presence of both mechanisms except, at different Reynolds number. In order to advance the concepts, further experimental data is s t i l l needed, especially probing the microstructure of the turbulent flame. A l l of the published experimental evidence suggest that turbulence is one of the most important parameters affecting combustion. Turbulence appears to be the key factor which can increase the speed of the flame propagation, and bring us closer to a practical fast lean combustion process. 1.2 Mixture Motion Effects in I.C. Engines Flow patterns generated in I.C. engines can be categorized into two groups: large scale motions and small scale effects. The first of these affect large parts of the combustion chamber volume. They are usually well developed and organized. Small scale motion is created by turbu-lence directly affecting the mixing process and the speed of flame 4. propagation. This division is to a certain extent simplistic, because turbulence is also generated on the boundaries of the large scale motions which results in an interaction of both patterns. The most common types of large scale fluid motion in engines are: swirl and squish. Swirl is generated during the inflow of mixture to the I.C. engine cylinder. It is created by intake port geometry or by obstructions located on the intake valves, which direct the mixture flow in a tangential rotating fashion, Figure 1.1. A very extensive study of the swirl effect on combustion in engines was performed at FORD in the mid-1970's. As a result it was reported that there is a considerable difference between the impact of swirl on combustion in S.I. and C.I. engines [4]. In a S.I. engine the presence of swirl affected neither performance nor exhaust emissions. In a C.I. engine, however, a positive effect was recorded. The gain in performance depended strongly on optimization of the swirl strength for particular engine geometry and operating parameters. The conclusion concerning swirl in S.I. engines was contradicted by the work of Witze [5,6] which showed the importance of other factors, like spark plug position and chamber geometry. More systematic studies similar to that of Witze are s t i l l needed to determine the detailed role of swirl in the combustion process in I.C. engines. The other type of large scale motion, squish, is generated before TDC of the compression stroke. It is induced by the geometry of a combustion chamber which directs mixture from the perimeter of the piston towards a centrally located bowl. Figure 1.2 shows a schematic of the squish combustion chamber. Experimental evidence on the role of squish in combustion is not conclusive. Some published results suggest that the overall effect of squish motion is negligible [7,8], while others [9] 5. find squish flow of considerable value. The most frequent opinion is that the mean squish motion doesn't increase the speed of flame propaga-tion. A noticeable decrease of ignition delay in squish chambers is attributed to a slightly higher level of turbulence generated by the squish. The combination of squish and swirl is reported to have a dramatic effect on combustion [10], It is suggested that squish helps to break up swirling motion and increases the turbulence level. Generally the effects of the large scale motions are strongly dependent on a particular engine geometry and operating conditions. The most important aspect of the fluid mechanics in I.C. engines is attributed to turbulence. Turbulence is characterized by a marked increase in the fluid transport properties. This in turn has significant consequences for the mixing and combustion processes. Most of the turbulent energy in I.C. engines in generated during break up of the inlet jet, created by the intake process. The small scale motion thus generated, decays during the compression stroke, leaving substantially less activity present just before, and during combustion. The turbulent flow field in an engine is usually characterized by a set of parameters used in the theory of isotropic turbulence. Their relevance to real engine flow parameters Is highly questionable. The problems arise from the unique nature of engine flows. In most cases the motion is highly unsteady, often inhomogeneous and nonisotropic, with considerable cyclic variations. The motion has also imposed on i t rapidly changing thermodynamic and geometric conditions. How to extract realistic information concerning turbulence in such conditions is a matter of debate. 6. To help at least quantify experimental data researchers turned to existing concepts like: mean velocity, intensity, integral and micro scales. In I.C. engines mean velocity, defined differently by many researchers, is regarded as some mean value the flow field contains. The best available definition is: t + T/2 1 w U(i,t w) = i / U(i,t)dt t - T/2 w where T is a period long enough to contain most of the turbulent frequencies and shorter than the time scale of the mean flow. The velocity fluctuation in a cycle i , is then defined as: u(i,t) = U(i,t) - U(i,t) where U(i,t) represents values obtained by curve fitting to data U(i,t w). The RMS velocity fluctuation, or turbulence intensity is represented by: t + T/2 w \l ( u ( i , t ) ) 2 d t t w - T / 2 The integral length scale is defined as: oo L = / R(r)dr x o where R(r) is a spatial autocorrelation coefficient. 7. The Taylor microscale is: (92R/9v2) 0 J 11/2 However, to evaluated these scales, assumptions of isotropy and relaxa-tion of the turbulent fields are usually used. The length scales are then calculated from easier to measure time scales. This procedure, in many engine applications, is very ambiguous. To draw more meaningful conclusions statistical analysis is usually performed on large sets of data. The presented definitions are not absolutely precise, but they offer a description of the flow field adequate for general analysis. It seems very unlikely that a better method will emerge in the nearby future. 1.3 Review of Previous Work on Turbulence in Engines The dramatic effect of turbulence on combustion in engines was demonstrated as early as 1911 by an experiment of Clerk [11], repeated afterwards by many others. A spark ignition engine which Clerk ran for three revolutions without intake and combustion, then ignited, had a combustion duration nearly twice as long as in the case when ignition followed immediately after the intake stroke. The engine generated also substantially less power. An experiment conducted recently by Dohring [12] in a rapid compression machine showed again how strong is the correlation between turbulence and the speed of the flame propagation. Measurements of the turbulent flow field in engines were first undertaken in the middle 1950's by Semenov [13]. The technique used was constant temperature hot wire anemometry (HWA). It was noticed that engine flow regimes were extremely complex to measure. Semenov's experi-ments were, however conducted in a simple disc shaped combustion chamber which allowed him to draw some basic conclusions. These were that the 8. source of turbulence generation was found in the intake process, decay during compression, relaxation in its final stage, and isotropy before TDC. His basic findings are s t i l l unchallenged. Experiments conducted in Great Britain between 1965 and 1975 were aimed first at the evaluation of the quality of the measuring technique [14,15]. Shortcomings of the hot wire anemometry (HWA) were exposed but i t was s t i l l the only available technique. A rare experiment was conducted by Tindal and coworkers [16]. They calibrated HW probes in conditions similar to that existing in engines. In their report Tindal suggested a method for proper analytical evaluation of data from probes calibrated in ambient conditions. More extensive projects were carried out during the "energy crisis" of the 1970's. In a relatively short period of time between 1970 and 1980 numerous results were published. In Great Britain Hassan, Dent and Derham completed a series of projects related to flows in diesel engine configurations [17,18]. In the US the research was concentrated in laboratories of auto manufacturers, like GM, and big national research centres, like Sandia. Published papers [19,20,21] analyzed effects of different parameters on the flow conditions and combustion. Their conclusions were in essence similar to Semenov's with further extensions for particualr geometries and conditions. The experimental technique was also refined. Some of the researchers went, however, a step further and tried to extract information about time scales of turbulent flow in engines [22,23]. Then using the assumption of isotropic turbulence they calcu-lated length scales. None of them tried to explore the margin of error too extensively. 9. More fundamental were projects aimed at obtaining quantitative correlation of turbulence with combustion. The most comprehensive of them done by Lancaster [24] obtained correlation of flame speed ratio (FSR) with turbulence intensity. Using the nondimensional FSR Lancaster avoided including effects of chemistry and thermodynamics on flame velocity. His correlation did not incorporate any effect of turbulent scales. This conclusion was contradicted by a report of Smith [25] who correlated turbulent Reynolds number with burn duration. Smith's method of obtaining Reynolds number is, however, very questionable. Up to the middle of the 1970's the only available measuring tech-nique to probe the flow field in engines was CTA. The technique mastered in wind tunnels proved to be very uncertain in engine-like conditions. Due to extensive difficulties with calibration of hot wire probes in high pressure and temperature, analytical techniques were chosen to obtain meaningful results. Directional ambiguity was overcome by using hot wires in better understood flow conditions. To obtain the necessary history of temperature, some of the reseachers used miniature thermo-couples and thermometers [13,17,20]. Their claims of obtaining a proper frequency response are s t i l l met with scepticism. An extensive evalua-tion of these temperature measuring techniques done by Wienke [26] concluded that generally they were showing considerable time lag and amplitude decrease in comparison to temperatures measured with optical methods. From the end of the 1970's the laser Doppler velocimeter (LDV) has been used to probe turbulence in engines. The technique promised many advantages: nonintrusive measurements independent from thermodynamic conditions of the flow, unambiguous results with respect to the flow 10. direction, and an opportunity to conduct measurements also during firing tests. At the present time experiments with LDV have become standard in many flow conditions. Measurements, however, in I.C. engines are s t i l l loaded with obstacles. A major difficulty encountered with LDV measure-ments In engines is a necessity to redesign the engine to accommodate visual access. This imposes restrictions on parameters of the engine operation. Other reported difficulties are: evaporation of seeding particles during compression and subsequent loss of signal intensity, fouling of optical windows and intermittency of the signal [27]. In the last 5 years LDV has allowed checks to be made on the quality of some measurements done with HWA, and the HWA technique has been refined. A very extensive comparison was conducted by Witze [28]. In conclusion he suggested an analytical method giving the best matching of HWA with LDV. A similar comparison was done by Monaghan et al. [29]. Some very interesting projects with LDV measurements in engine like conditions are reported by Arcumanis, Bicen and Whitelaw [30,31,33]. They conducted experiments in a transparent engine model and evaluated the very complex nature of the swirl-squish interaction. The model did not allow, how-ever, for measurements with engine like speeds and compression ratios. On the whole, experiments with LDV confirmed trends in flowfields measured with CTA. How powerful this new technique is becoming has been demonstrated in a recent paper by Fraser [33], which reports two point length scale measurements with LDV in conditions close to a normal engine operation. The last part of this review addresses one specific area, important for this thesis. This is the squish effect. There have been many projects carried out to evaluate squish motion. Reports are however 11. contradictory. Fitzgeorge and Allison [8] in their attempt did not manage to detect any squish velocity. They claimed that it was negli-gible. More detailed experiments conducted by Shimamoto [35] also involved development of a new measuring instrument. The squish velo-cities were measured and shown to be very close to those theoretically predicted. Unexpected large cyclic variations of the squish velocity were, however, also detected. A very complete evaluation of squish was performed by Woods and Ghirlando [36]. They solved a complete theoretical problem by a method of characteristics. Subsequent experiments showed squish velocities significantly different from those predicted; lower in values, shifted in time and with cyclic variations. From the combustion stand point, however, squish effects were evaluated mostly in relation to C.I. engines. There has been very l i t t l e work done on the effect of squish on combustion in S.I. engines. Interaction of flow with combustion is however of a quite diferent nature in diesel engines in comparison to S.I. engines. 1.4 Objectives and Scope of the Work The objective of this work was to gather more experimental data on the effects of turbulence on combustion in a modern type S.I. engine. The originality of the project lies in its attempt to control the character of turbulence in the engine at the final stage of the compression stroke, without affecting the nature of the intake process. A preliminary evaluation done by Cameron [37] suggested that a new bowl-in-piston type combustion chamber, called the squish-jet chamber, had potential to become a turbulence generator, in the I.C. engine environment. Analytical evaluation of the design showed a possibility of 12. obtaining high velocity jets in the combustion chamber at about 40° BTDC. Turbulence generated by the breaking up of these jets would have excellent timing in affecting the combustion process. The project was completed in three stages. A flow visualization experiment, conducted first, was expected to confirm qualitatively the jet effects in engine-like conditions. A transparent model was built and assembled on a rapid compression machine (RCM). Flow inside the model was seeded with microballoons, and high-speed cine films taken during single compression strokes. The RCM simulated one stroke of a motored engine at conditions of 1000 RPM and 9:1 CR. Subsequent experiments, conducted on a re-designed Ricardo Hydra single cylinder research engine, were divided into flow measurements and combustion tests. Flow experiments were conducted on the engine motored by a D.C. motor of a dynamometer. Histories of the jet development were measured with a hot wire probe located in the piston bowl. Turbulence parameters were obtained with a standard HWA probe through the cylinder head. The tests were conducted for 8 different piston geometries and at different operating conditions. A third set of tests was aimed at evaluation of combustion histories and engine performance for 5 different piston designs, over a wide range of operating conditions. An extensive analysis of experimental data included performance parameters according to SAE Power Test Code, J1349, pressure histories, standard deviation of cyclic peak, pressure, IMEP, standard deviation of cyclic IMEP and cumulative mass fractions burned. During the firing experiments the engine was fuelled with natural gas and the new method of turbulence enhancement was expected to increase the speed of flame propagation of this slow burning fuel. 13. CHAPTER 2 EVALUATION OF THE SQUISH-JET CONCEPT 2.1 Squish-Jet Design There has been very l i t t l e progress made in the search for a better S.I. engine design which would allow the generation of turbulence just before ignition and retain some degree of control over its parameters. An original idea, based on the above assumptions, however, was patented by Evans [38]. The so-called squish-jet piston can potentially generate controlled turbulence in a S.I. engine environment. The novelty of the design is in the incorporation of channels in the piston of the bowl-in-piston type of combustion chamber. The channels were expected to create jets in the combustion chamber during the late stage of the compression stroke, thus changing the flow pattern of the mixture. Introduction of this additional turbulence production source was expected to increase the burning speed and manifest itself in sub-stantial gains in engine performance. The cross-section of the design is shown in Figure 2.1. Preliminary work done by Cameron on a CFR engine brought mixed results. Hot wire measurements conducted through the cylinder head failed to detect the presence of jets, but effects on combustion duration have been documented. The conclusion was that the incorporation of the design shortened the ignition delay period. A critical evaluation of her work showed some weaknesses of the apparatus on which the work was conducted. Cameron used a piston altered to accommodate an insert with the squish-jet geometry. The constraints imposed by this, however, created a design unfavourable for jet creation. 14. The jet passages were located too close to the bowl entrance, thus diminishing the jet effects. The CFR engine used in the project was an older model, with heavy blow-by detected during the experiments. Its side-located spark plug, made it difficult to relate the situation to modern engines. The maximum possible engine speed was only 1200 RPM, well below normal range of an automotive engine operation. Finally, the experimental rig for her work did not allow any performance analysis to be conducted. As a result, a more complete project was undertraken by the author on a Ricardo Hydra, single cylinder research engine. The first step was an expansion of the analytical model describing flow in the squish-jet piston, with exposition of a l l the factors which could potentially affect jet velocities. The problem in itself turned out to be very complex, and only an approximate solution was used as guidance for the experimental work. The parameters which play an important role in squish-jet effects are: • degree of leakage from the area above the pistion either past the piston rings or valves, • inertia effects at high piston speeds, • adverse effects of heat transfer creating negative temperature gradients between the middle of the bowl and the bowl perimeter, • friction on the piston surface, and pressure losses in the channels. A simplified 2D analysis describes squish jet velocities in terms of geometric parameters of the engine and operating speed. The conclusions are discussed in the next chapter and the model in Appendix A. 15. 2.2 Flow Visualization Experiment The simplified analysis of the jet development allowed conclusions to be drawn with respect to trends of the mechanism. Both squish and jet velocities are strongly dependent on the piston velocity, clearance value and squish ratio. Their dependence on compression ratio is less dramatic. These facts have very important consequences. To create the desired effect the model had to obtain speeds comparable to the real piston. The standard technique of decreasing speed and using the medium of lower kinematic viscosity to keep the Reynolds number the same as in the real experiment, would not in fact generate squish and jet motions. It was decided to build a transparent model of the Ricardo engine block and assemble i t on the rapid compression machine [RCM]. The compression stroke of the RCM was transferred through a push rod to a plexiglas piston in the model. The cylinder liner of the model was made of plexiglas, and the covering plate of Lexan polycarbonate. The piston was in the form of a removable insert, which allowed for quick change of the geometry. It was sealed with two compressed teflon rings. The RCM simulates a compression stroke at 1000 RPM, relative to engine condi-tions, and has a compression ratio of 9:1. These were the parameters of the model operation. A clearance value of 1 mm was chosen similar to the real engine, and al l the dimensions were scaled with the engine. Flow inside the model was seeded with microbaloons. All the events during a compression stroke were filmed with a high speed camera at 1200 frames/sec. A set of experiments was performed with different geometries of the piston channels. The jet effect was documented but its strength was 16. much lower than expected. An important factor in the theoretical analy-sis was the effect of ring leakage on the squish. To check the quality of the model in this respect, a compressed state was left intact for a period of 3 minutes. The level of compression was checked afterwards, and found hardly changed. Figure 2.2 shows the set up of the experimental equipment for the flow visualization experiment. The model is seen positioned horizontally with a camera's view through the front plate. Figure 2.3 displays a photograph of the assembly. Figures 2.4 and 2.5 present 6 frames from one of the films. Jets are seen developing from two channels located horizontally. The total number of channels in this experiment was 8 but to avoid strong gravity effects only two were seeded. The flow visuali-zation experiment was expected to correlate the recorded events in the piston bowl with the angular position of the piston. A display of an electronic counter was located in view of the camera. The counter was triggered by an optical sensor located on the toothed wheel on the RCM. However, because of the adverse lighting effects the counter readout was not bright enough to be recorded on the film. Another simple method was then used. A stroboscope lamp was positioned in view of the camera and triggered 4 times during the compression stroke. More frequent trigger-ing was not attempted because of the possibility that some of the flash signals could be lost between the film frames. As a result only an approximate location of the film frames was established. In summary the experiment offered four basic conclusions: 1. The expected jet mechanism had been developing in practice but i t was much weaker than theoretically predicted. 17. 2. The jets didn't penetrate the middle of the piston bowl. 3. An advantageous change was to move the inlets of the channels, in the piston top, away from the entrance to the bowl. 4. The jet effect occurred even in the case of the channel geometry with maximum loss coefficient. 18. CHAPTER 3  EXPERIMENTAL APPARATUS AND METHOD 3.1 Experimental Method The most extensive part of the project was conducted in a test cell of the Alternative Fuels Laboratory of the Department of Mechanical Engineering at the University of British Columbia. The cell is equipped with a Ricardo Hydra, single cylinder research engine which has become a standard type used by many research centers. The experiments were divided into two phases: flow measurements and performance tests. Flow measurements were conducted in the engine while motored by a D.C. motor of the dynamometer, at WOT and three different speeds: 1200, 2100, 3000 RPM. This speed range is representative of a modern engine operation. The measuring technique was constant temperature anemometry (CTA). Hot wire probes were calibrated in a wind tunnel, and compensated analytically for different thermodynamic conditions present in the engine. To analyze jet development, measurements were done with a hot wire probe located in the piston bowl. It was decided to do the measure-ments for different channel geometries. Turbulence parameters were evaluated with a second probe inserted through the cylider head into the combustion chamber. The temperature of the mixture was calculated from the pressure history using the perfect gas law. Both HWA and pressure signals were digitized by a high-speed data acquisition system (DAS), ISAAC 2000, triggered by clock pulses every .2 degrees of crank angle. Performance tests were conducted for different piston geometries, chosen from those analyzed in the flow experiments. Test regimes included f u l l load operation at different speeds and air fuel ratios 19. (RAFR), part load at 2.5 BMEP, 2000 RPM and range of RAFR. During the tests the engine was fuelled with natural gas. The ignition timing was optimized to obtain minimum spark advance for best torque (MBT). A l l important parameters of the engine operation were stored in an IBM PC, and scaled to required SAE standards. Additionally pressure traces were collected at each test point. These were subsequently analyzed on a VAX mini-computer. Extensive analysis of the pressure histories was focused on peak pressure, IMEP and mass fractions burned. 3.2 Experimental Apparatus Two different areas of design work were required, in order to conduct the experimental program: 1. Major redesign of the existing configuration of the Ricardo Hydra engine. 2. Development of a system to measure jet velocities in an unambiguous manner. The standard gasoline Ricardo Hydra engine configuration has a "bath-tub" combustion chamber located in the cylinder head, and a flat piston. To conducted the planned experiments the configuration had to be changed. The re-design included: • design and manufacturing of a flat, aluminum cylinder head, • design and manufacturing of 10 aluminum pistons, • manufacturing of a new longer cylinder liner, • new and modified connecting rod, • modifications to the cylinder block, and timing drive system, 2 0 . • manufacturing of a packing plate to be placed under the cylinder block. The new cylinder head was required to have a flat bottom deck and a central position of the spark plug. To meet these conditions the valves position was moved from the center of the chamber, a smaller 12 mm spark plug accommodated close to the center, the exhaust port moved to the side of the head, and cooling manifolds re-directed. The new head was also fitted with a pressure transducer sleeve. The new pistons had to be much longer than standard designs In order to accommodate the combustion chamber bowl. This also required a new, longer cylinder liner. It was decided to manufacture pistons of casting aluminum alloy A356, heat treated to T6, the same type as for the cylinder head. The idea to use different inserts in the same piston was abandoned because of the requirements regarding geometry of the channels and material properties. Properties of the material were an important consideration. Automotive aluminum pistons are usually forged or cast and fast cooled in permanent molds. This procedure decreases the material grain size and enhances its strength in higher, operating temperatures. Financial constraints, however made only sand-casting a possible manufacturing option. As a result i t was estimated that the original piston material was 1/3 to 1/2 stronger at higher temperatures than the new one. This fact was taken into consideration in the choice of testing strategy. The manufacturing process of the cylinder head consisted of: pattern making, casting and machining. This was contracted outside the university. Machining of the pistons, however, with especially difficult 21. oval skirts was done in the machine shop of the Mechanical Engineering Department. The head was fitted with a pressure transducer sleeve. Valve seat inserts and guides were chosen from those available on the market. The geometry of the piston oval was completely new, because i t depends on the thickness of the piston wall in the wrist pin boss area. Piston rings and pin were chosen again from those available on the market. The rest of the modifications to the engine were done in the machine shop in the Mechanical Engineering Department. Figure 3.1 shows a cross section of the Ricardo engine in the new configuration. Figure 3.2 presents photographs of a cut casting of the new cylinder head and the final, machined product. Figure 3.3 shows photographs of the Ricardo Hydra engine and new pistons. The second area of necessary design work arose from a need to measure jet velocities. It was decided to develop a system based on a HWA probe located in the piston with signals transferred out along a linkage mechanism. The ini t i a l plan called for a custom made probe by a specialized manufacturer, TSI. This option turned out to be too expen-sive. An original probe was then designed and built from the same materials TSI uses. It is important to mention that the I.C. engine environment imposes very strict requirements on probes. They are exposed to higher temperatures and pressures and a considerable level of vibration. These constraints affect the choice of materials and manufacturing technique. The new probe is shown in Figure 3.4. The most crit i a l part of the system was, however, a linkage mechanism designed to transfer hot wire signals from the engine. The originality of the design is in the complete absence of wire deflection, even though the mechanism goes through a wide-range motion. The only 2 2 . part of the system where the deflection of the wires could not be avoided was in the connection between piston, and connecting rod. This joint was however optimized. As a result during the experiments there was not a single problem with the system. Figure 3.5 shows a schematic of the linkage mechanism. Figure 3.6 presents photographs of the connecting rod together with the linkage mechanism and inside view of the piston. Figure 3.7 shows photographs of the hot wire probe in the piston. The engine modified as described above was the heart of the experi-ments. The linkage system was used only for motoring tests during which a dc electic motor of the dynamometer was turning the engine. The engine specifications are given in Table 1. 3.3 Instrumentation and Data Acquisition System 3.3.1 Flow Measuring System Both of the hot wire probes were controlled by a DANTEC 56C17 bridge and CT01 anemometer operated in a constant temperature mode. The signals were filtered by a DANTEC 56N20 signal conditioner, set in a low pass filter mode, at 30 kHz. The signals were subsequently digitized and stored by a high-speed data acquisition system (DAS), ISAAC 2000. Clock pulses at .2 degrees of crank angle and trigger signals for the DAS, were generated by an AVL 360c/600 optical crank angle encoder coupled to the engine crank shaft. The stored data for 44 cycles, at particular engine operating conditions, were then transferred to the IBM PC, and VAX 11/750 mini-computer for analysis. Measurements with the two probes were carried out separately to better organize the acquisition process. The probe inserted through the cylinder head was a high temperature TSI probe, model 1226. Its specifications are shown in Table 2. The short probe located in the piston was built according to the same sensor specifications. Initially, each velocity signal was accompanied by a pressure history. Motoring pressure traces were, however, very repeatable. Average signals for different engine speeds were then used. Pressure was measured by a Kistler 6121A piezo-electric pressure transducer, amplified by a Kistler 5004 charge amplifier and digitized by the DAS at the same rate as the velocity signal. Control over the acquisition process was carried out by the IBM PC and over engine motoring conditions by the engine electronic control system made by CUSSONS. 3.3.2 Instrumentation for Firing Tests The test cell of the Alternative Fuels Laboratory has extensively developed instrumentation to assist engine testing. The system is designed around the Ricardo Hydra single cylinder research engine coupled to a McClure dynamometer. The engine is equipped with the Cussons electronic Control unit by the manufacturer. The unit monitors speed and load on the dynamometer, parameters of coolant, and lubricant, variable ignition timing, and air flow in a Meriam model 50MC2 - 4F laminar flow element. The air flow meter is located on the intake manifold and calibrated for pulsating flows. To monitor the engine performance, a Data Translation model DT2801A A/D converter controlled by the 1MB PC is used. The engine can be fuelled with natural gas or gasoline. During the experiments reported in this thesis i t was only operated on natural gas from city mains. The gas specifications are given in Table 3. 24. The Data Translation data acquisition system operates with a frequency of 27.5 kHz and collects ten basic signals from analog trans-ducers. The monitored signals, in addition to those already mentioned are: natural gas differential pressure on a Meriam model 50 MWV-1.5 laminar flow element, pressure and temperature of air and natural gas. The engine operating parameters like: power, BMEP, BSFC, ignition timing, AFR and efficiency are continuously being calculated and updated every 4 seconds on the PC screen, ready to be saved. To analyze pressure history the previously mentioned high speed DAS is used and up to 100 cycles of pressure signals at a particular engine operating point can be saved on diskettes. During the performance tests the engine was driven to a set point, at which a l l the performance parameters were saved and 44 cycles of pressure data collected. The procedure was repeated for each testing point and piston geometry. Figure 3.8 shows a complete schematic of the data acquisition system, used both in the motoring and firing tests. 25. CHAPTER 4  DATA ANALYSIS The following two sections briefly describe conditions under whch the experiments were conducted and the techniques used to analyze collec-ted data. The methods to extract velocity from HWA signal, and turbulent fluctuations from velocity signal have been already well documented and references to them are only made. 4.1 Flow Field Analysis 4.1.1 Experimental Conditions During motoring tests three parameters were to be measured: jet velocities with the short probe, velocities in the piston bowl with a standard probe and pressure history. The i n i t i a l plan was to conduct a l l the measurements at the same time. The first experiment showed, however, that the short probe was exposed to very dynamic flow conditions. This caused frequent breakage of the probe's sensor. To repair the sensing wire the engine had to be dismantled. It seemed advantageous then to conduct the work in two steps: first jet measurements, then turbulence evaluation. As mentioned before the repeatability of the pressure signal required fewer measurements be made. The jet measurements were carried out for 8 different configura-tions: seven piston geometries and one different sensor orientation. The cases are shown in Figure 4.1. Most of the experiments were conducted at three different speeds: 1200, 2100, and 3000 RPM. Problems, however, emerged with the sensor oriented parallel to the direction of the jets, case number 6. Attempts were made three times to measure jet velocities at speeds higher than 2100 RPM. Every time however the sensor was broken. This can be explained by adverse dynamic flow conditions existing at the breakup region of the jets. To fully complete this part of the experiment the engine had to be dismantled 15 times. The measurements conducted with the standard hot wire probe inserted through the cylinder head were much less time consuming. After sensor failure, the probe was easily removed from the engine. The tests were conducted for the same speed range as with the short probe. All cases were evaluated with the hot wire sensor positioned just below the ignition point. For two piston geometries measurements were also taken in two additional positions across the piston bowl; in the middle and close to the bottom deck. The positions across the piston bowl are shown in Figure 4.2. 4.2.2 Analytical Procedure The analytical method used to calculate velocity from the anemometer signal is a fairly well-known technique. The theory based on King's law [39] was developed by Collins and Williams [40] and Davies and Fisher [41]. Over the years of testing in I.C. engine environment i t was refined by comparisons with LDV. The procedure used in this thesis includes suggestions by Witze [28] and is described in detail by Cameron [37], Dohring [12] and Boisvert [42]. Calibration constants were evaluated for each sensor in a wind tunnel at a velocity range 1-18 m/s. The temperature of the gas during engine experiments was calculated using the perfect gas law and assuming ambient temperature in the cylinder at the point of valve closure. 27. In the analysis of the jet effects only mean velocities were important. Therefore, for each case, cyclic velocities were ensemble averaged. To evaluate turbulence characteristics a cycle by cycle nonstationary window averaging technique was used. This method is regarded as the most sound way of extracting fluctuations from HWA velocity measurements in engines. The technique was evaluated by Cattania [43] and is also described in detail by Boisvert [42]. In each of the analyzed cases mean velocities, and RMS fluctuations for indiviudal cycles were extracted from raw velocities. Cyclic values were then ensemble averaged. The RMS values can be regarded as representative of turbulence intensity. No attempt to evaluate scales was made here because, it was regarded that in the particular flow conditions, and technique available no procedure would have a sound basis. 4.3 Firing Tests Analysis 4.3.1 Test Conditions Combustion tests were performed for five different piston geometries, chosen from those used in the motoring tests. All of them had the same most important geometric parameters: clearance value, squish area and compression ratio (CR). An additional case was evaluated for the squish piston at higher CR. The cases are shown in Figure 4.3. All six were analyzed with the same testing strategy: Full Load Operation: 1200 RPM: RAFR = 1.00 RAFR = 1.25 2100 RPM: RAFR = 1.00 RAFR = 1.25 3000 RPM: RAFR = 1.00 RAFR = 1.25 2 8 . Part Load Operation: 2000 RPM, 2.5 bar, RAFR = 1.00 - 1.30 During the tests the engine was fueled with natural gas, and ignition was optimized to obtain minimum spark advance for best torque (MBT). The temperature of the cooling water and lubricating o i l was kept between 70 and 80 °C. Interesting problems were noticed in the course of the experiments. Each piston had a new set of rings, creating a need for a break in period. During the first hour of the engine operation efficiency was only of the order of 23%, slowly increasing afterwards and levelling off after 6 hours. It was decided then to run each new configuration for about 10 hours under f u l l load conditions before conducting the tests. Another problem was valve floating. The clearance value used for motoring and firing tests was the same; 1 mm. This, however, caused a decrease of real gap in firing tests and consequently the valves were hitting the piston at higher engine speeds. It was decided then to make shallow cut-outs in the piston top to accommodate the valve float. 4.3.2 Analytical Procedure At each testing point data related to engine performance was saved. Amongst these were: natural gas and air flow rates, ignition advance, torque, speed, temperature and pressure of air and natural gas. The data before saving was averaged over 100 cycles. A computer program was subsequently used to evaluate performance parameters according to SAE.J1349 code requirements. Parameters like: power, BSFC, BMEP, AFR, brake thermal efficiency were calculated. At each testing point pressure traces for 44 cycles were also collected. These were transferred to the VAX minicomputer where analysis was undertaken. Pressure cycles were related to real values by assuming that pressure at the BDC, before compression, is equal to volumetric efficiency multiplied by ambient pressure. In each cycle peak pressure and IMEP were calculated. The traces were ensemble averaged, and peak pressure and IMEP were evaluated again. Statistical analysis was performed to calculate standard deviation of individual IMEP and peak pressure from the ensembled signal. Coeffi-cients of variance (COV) of both peak pressure and cyclic IMEP were calculated. A simple mass-burned fraction program was also developed to evaluate burning history. The code was based on a procedure, described by Rassweiler and Withrow [44], correlating cylinder-pressure development with the progress of flame front. The method is based on properties shown by a logarithmic P-V diagram and on some simplifying assumptions. The P-V diagram shows that once the intake valve is closed, the compres-sion process tends to be polytropic to the point of ignition, and the slope of the compression line in that region provides the polytropic exponent of compression. Similarly, from the end of combustion until the exhuast valve opens the process tends to be polytropic and exhibit its own exponent. The mass trapped in the cylinder is essentially constant. During combustion the polytropic exponent changes from its compression value to its expansion value. Pressure increases partly due to burning and partly due to piston motion. The latter can be calculated if the change in the polytropic coefficient is approximated by a linear func-tion. The problem with the procedure is to properly evaluate the end of 30. combustion. This was done here by calculating the second derivative of the polytropic coefficient. The end is assumed when the derivative comes to zero. The whole combustion duration is divided into steps in which pressure increase due to combustion is separated from that due to piston motion. Mass-burned fractions are then calculated as the pressure increase due to combustion divided by the total sum of pressure increase due to combustion. This method has appeal because i t Is very simple and easy to use. A comparison of essentially similar techniques with new sophisticated computer codes was done by Amann [45]. He found a high degree of agreement between results obtained from these simple methods and the advanced codes. Appendix B presents the printout of the computer program used to calculate mass burned fraction. 31. CHAPTER 5  DISCUSSION AND EXPERIMENTAL RESULTS This part of the thesis is divided into three sections. The first describes the results of the flow measurements. The second evaluates combustion tests and the third analyzes the measuring techniques used in the project. 5.1 Flow Field Measurements 5.1.1 Results of Jet Velocity Measurements The experiments conducted in this phase of the project were expected to show whether the predicted jet mechanism was developing during a real engine operation, and which channel geometry was the most advantageous for the jet strength. Figure 4.1 presents the types of piston geometries evaluated in these experiments. A comparison of cyclic velocity profiles for piston configurations number 4 and 5 can be seen In Figure 5.1. The piston number 4 is an ordinary squish design without channels. There Is a dramatic increase of velocity before TDC, for case No. 5, an indication of the jet motion. The flow detected before TDC in piston 4 is, however, also of considerable value. This result is understandable because in the squish design, horizontal flow above the piston changes into vertical inside the bowl. A stronger indication of the jet flow is displayed in Figure 5.2 which shows cyclic velocity profiles obtained in pistons number 5 and 6. The latter case was conducted with the probe's sensor oriented parallel to the jet direction and only at 2100 and 1200 RPM. The speed limit was caused by a repeated sensor breakage in case 6, during attempts to conduct measurements at 3000 RPM. There is a clear 32. difference in velocity profiles between case numbers 5 and 6. The substantially higher velocity before TDC in geometry 5 represents the jet flow. The probe indicates maximum jet velocity of 20 m/s which is probably overestimated because of three factors: the probe had not been calibrated in velocities higher than 18 m/s, the air in the jet was colder than in the bowl and because the flow conditions in the jet region were very complex. The cyclic velocity traces for some other geometries are shown in Figure 5.3. An interesting case is number 2, which was expected to generate the strongest jets. However, as the velocity profile for this case shows, there is a high velocity flow through the channels during the middle part of the compression stroke, but i t diminishes before TDC. It appears that geometry 2 reduces pressure gradient above the piston during the i n i t i a l stage of the compression stroke and prevents strong jet development at the end. A more general trend was expected to be displayed by ensembled velocity traces. The averaging in each case was carried out over 44 cycles. Figures 5.4, 5.5 and 5.6 show comparison of ensembled jet velocity profiles for different piston geometries. It is apparent that differences between pistons with channels are not substantial. The case number 4, Figure 5.5b, shows however, lower velocities than the others. The strongest jet effect displayed by the ensembled profile is for the piston number 7, Figure 5.4. Figure 5.6 presents ensembled velocities for cases 5 and 6 measured at the engine speed of 2100 RPM. The differ-ences are rather small. An analysis of two diferent cycles from one test point, displayed in Figure 5.7 suggests the reason why ensembled signals don't show stronger jet velocities. It appears that there is a consider-able variation in the jet timing between cycles. Their development is 33. also less similar to a steadily changing jet but rather to an irregular jet flow. This character was also noticed during the flow visualization experiment. The averaging process smoothes individual peaks and the final effect is less dramatic than that displayed by individual cycles. Another question posted was how the jets develop at different engine speeds. Figures 5.8 and 5.9 show cyclic velocity profiles for the geometry number 5 at three engine speeds. There is quite good propor-tionality displayed by the graphs with the jet velocity increasing gradually at higher engine speed. Figure 5.10 presents a similar comparison for ensembled signals. To establish which evaluated geometry gives the strongest jet effect a complete evaluation was conducted. It was concluded that channels of 5/32 inch diameter generated higher energy jets than 3/16, Figure 5.11. A change from 8 to 4 holes didn't seem to create much stronger jets, Figure 5.5. It was anticipated, before the tests, that the geometry number 2 would be very advantageous. The channel shape have the lowest coefficient of pressure loss in comparison to other cases and i t is easy to manufacture. It was recognized, that the geometry number 2 would create jets directed at the bottom of the piston bowl but i t was expected that an increased jet strength could create interesting effects. The experimental data for case 2, Figure 5.3, does show interesting results, but as mentioned before not the desired ones. The other interesting geometry is number 7. It offers advantage in manufacturing and additionally It directs jets at the spark plug location. The geometry has also a disadvantage - the highest coefficient of pressure loss in comparison to the other cases. The measured cyclic jets are, however, not much weaker than for the other piston geometries. 34. The ensembled velocity profile for case 7, Figure 5.4 shows even more pronounced jet motion. This result suggests a potential advantage of piston 7 in combustion. A very important question was how far the jets were penetrating the piston bowl. The answer was expected to be given by measurements conduc-ted through the cylinder head. The flow visualization experiment suggested, however, that the effect is rather local, contained in the region close to the bowl walls. On the whole, this part of the experimental work offered mixed conclusions. The jet mechanism was confirmed to develop in the real engine operation. However, taking into consideration the shortcomings of the measuring technique and the results shown by the ensembled velocity profiles i t was concluded that the jet motions were relatively weak. This conclusion had to be confirmed, however, by more detailed measure-ments in the piston bowl. 5.1.2 Results of Turbulence Measurements The measurements of turbulence in most of the evaluated cases were performed in the region about 3 mm below ignition point. This position was close to the bottom deck of the cylinder head, however, i t was felt that flow measurements in that region were the most important for the correlation of the flow field with combustion. A comparison of ensembled mean velocities for four different piston geometries at the engine speed of 3000 RPM is shown in Figure 5.12. The mean squish flow in piston 4 is clearly visible before TDC. The other pistons have this motion substan-tially diminished. Even piston number 7, which has upward directed channels does not generate higher level of activity near the spark plug 35. location. Figure 5.13 presents a similar comparison of ensembled RMS turbulent fluctuations. The trend shown by the fluctuations follows the mean motion. In piston 4 the squish flow increases also substantially the level of fluctuations before TDC. An interesting aspect of these results is the level of agreement with respect to the velocity trends measured In different piston geometries. Both mean velocities and turbulence fluctuations follow nearly the same curves in a l l cases, except for the region close to TDC. Figure 5.14 shows mean velocity profiles at different engine speeds for piston 1. Figure 5.15 presents the same comparison for piston 4. The latter case displays a strong squish motion before TDC decreasing from 28 m/s at 3000 RPM to 18 m/s at 2100 RPM. Figure 5.16 shows similar data for turbulent fluctuations, which decrease from 8 m/s at 3000 RPM to 6 m/s at 2100 RPM. Comparison of measured mean squish velocity for case number 4 with that predicted by the analytical model shows a good agree-ment. According to the model squish velocity should be of the order of 30 m/s, at the engine speed of 3000 RPM. It should be kept in mind, however, that at the top probe location the squish motion is changing into a downward flow carrying also cooler air. This can lead to an overestimation of the velocity measured with the HWA. In two different geometries, number 4 and 1, measurements were also conducted down Into the combustion chamber bowl, Fig. 4.2. The results of these measurements represent probably the most important insight into the flow character in the piston bowl. Figure 5.17 shows ensembled mean velocities measured down into the piston bowl for piston 4. Figure 5.18 displays fluctuations for the same experimental conditions. Both the 36. mean velocity and turbulent fluctuation decrease down in the piston bowl. Interesting elements are bumps displayed by both traces at the points where the piston's top surface moves across the sensor position. The analysis down into the combustion bowl for piston 1 is shown in Figures 5.19 and 5.20. It appears that the flow activity at the bottom part of the piston bowl is much higher for piston 1 than 4, Figures 5.21 and 5.22. Piston number 1 shows less activity in the middle part of the bowl, Figure 5.23. The bottom part is, however, filled with unusually high energy motion. This pattern suggests that developing jets were pushed down into the bowl, thereby increasing levels of activity. The analysis suggests the following conclusions: 1. Channels in the piston reduce squish motion in the upper part of the combustion chamber, which considerably decreases the level of turbulent fluctuation and mean motion at the spark plug position, before TDC. 2. The reduction in the squish motion is proportional to the number of channels. 3. An increased level of activity at the bottom of the piston bowl is the only positive effect of the jets. 4. Jet effects are generally weak and do not influence flowfield in the middle part of the piston bowl. 5.2 Firing Test Results During the firing tests the engine was fuelled with natural gas and the performance parameters were evaluated over a complete regime of operating conditions. It has been already mentioned that the testing was carried out only up to 3000 RPM, the medium-range speed available on the 37. engine assembly. This range was chosen because of a considerable concern over the properties of the piston material. 5.2.1 General Performance Parameters The calcualted performance parameters are presented in Tables 4 to 10. The performance evaluation at wide open throttle (WOT) operation was conducted for three speeds, 1200, 2100 and 3000 and two relative air fuel ratios (RAFR), 1.00 and 1.25. The analysis of the results shows that at WOT there is no clear difference in the performance of the engine for different piston geometries. To a large extent this is surprising, because of factors which are mentioned in the error analysis. At part load operation, 2.5 bar BMEP and 2000 RPM, fuel consumption was analyzed for different RAFR. This is an indication of the efficiency of the engine operation. The results are shown in Table 10. There are notic-eable differences of the order of 5-10% in the lean limit, but they are rather favourable for the squish piston. However, at these operating conditions, errors associated with the evaluation of the engine perform-ance are assessed to be the highest. 5.2.2 Firing Pressure Analysis The analysis of the combustion process is the most objective when based on the pressure history. For that reason i t was the most extensive part of the analytical work in the project. Tables number 11-14 show different elements of this evaluation. Peak pressures of the ensembled pressure signals are displayed in Tables 11 and 12. They are the high-est, at a l l operating conditions, for the piston without channels. The piston with four channels shows the second highest maximum. Pistons with 3 8 . 8 channels produce the lowest peaks, a l l in the same range. Standard deviation of individual peaks for different geometries does not show any noticeable trend. At a l l speeds and for RAFR=1.00, its value is in the range of 4.5-5%. At RAFR=1.25 it increases to 7.5-10%. IMEP calculated for individual cycles and than ensemble averaged turned out to be different by only 1 kPa from IMEP calculated over the ensembled pressure trace. This shows the high quality of the pressure traces taken with the resolution .2 CA. The traces were digitally filtered before performing calculations. Trends in the IMEP for different piston geometries are well pronounced. They are presented in Tables 13 and 14. The IMEP calculated for the ensemble average taken over 44 cycles seems to be consistently higher for the pistons with channels. The lowest value is for the squish piston, no 4. At the same compression ratio the best IMEP's are for pistons with 8 holes directed upwards or to the middle of the combustion chamber. An evaluation of expanded pressure traces helps to explain this result. The piston without channels produces faster rate of combustion at the ini t i a l stage of the process, before TDC, which results in a higher negative work. This decreases the value of the cumulative IMEP. An analysis of the standard deviation of cyclic IMEP does not reveal too much variation between the designs. Over the whole speed range and RAFR=1.00 standard deviation, for different geometries, is within 1-1.7%. For RAFR=1.25 i t doubles to 2.5-3%. There is a slight trend, noticed at different engine speeds. The deviation decreases with speed but only by a marginal value of .5%. To check the effect of higher CR. on the combustion process piston number 4, with lower bowl volume, was tested. This change increased CR. 39. from 9:1 to 9.1:1. Results of the tests are shown as for piston 41 in the tables. It is somehow surprising that such small increas in CR. substantially changes peak pressure and IMEP. This gives, however, an insight at the importance of CR. The performance of the engine in terms of power etc., is not much different for case 41 in comparison to case 4. The tests were, however, carried out first with piston 41. The same piston, with altered bowl, was then used for test 4. It is recognized that because of this, friction loss ln case 4 was lower than in the rest of the tested geometries. This could have lead to a slightly better performance. The most interesting aspect of the analysis is shown by ensembled pressure traces. Figures 5.24-5.33 present traces obtained for a l l piston geometries, over a wide range of operating conditions. The trends are more visible at higher engine speeds at which the squish velocity has the highest value. Figure 5.24 shows ensembled pressure traces for five piston geometries at 3000 RPM and RAFR=1.00. Figures 5.25 and 5.26 present exposed details of the curves from Fig. 5.24. The trends are obvious. The squish piston without channels produces the fastest burning. Piston with 4 channels has slower combustion and pistons with 8 channels the slowest. Amongst the pistons with 8 channels piston 7 seems to be slightly better than 5. The slowest combustion is in the piston number 2. A similar trend is observed at different engine speeds and RAFR. An unexpected result is shown in Figures 5.32 and 5.33. They compare pressure cycles for 3000 and 2100 RPM for two RAFR. It seems that dispersion of pressures for different pistons is higher at RAFR=1.00 than at RAFR=1.25. The last part of the analysis was carried out with the developed model for mass fraction burned. It has been already described in section 40. 4.32. It was interesting to see what trend this simple technique would show. The model was run over some characteristic cases and the results are shown in Figures 5.34-5.36. Figure 5.34a shows mass-burned fractions for piston number 7 at three different speeds, Figure 5.34b displays similar comparison for case 3. The model represents trends in burning quite well. A decrease in engine speed slows down, in a proportional way, the burning rate. Figures 5.35 and 5.36 show mass fractions burned for different pistons. Again the trend is showing the direction pre-sented by the pressure curves. The fastest burning is produced for pistons with no channels, slower pistons with 4 channels and the slowest for the geometry with 8 channels, directed at the bottom of the bowl. A comparison of pressure dispersion between pistons at different RAFR presents the same trend noticed from pressure histories, lesser disper-sion at RAFR=1.25 than at RAFR=1.00. It is interesting to notice that the model based on so many assump-tions gives results which can be viewed as good, even for sophisticated codes. 5.3 Evaluation of Measuring Techniques A very important part of every experimental work is an assessment of the limitations of the employed measuring technique. This is a basis for the evaluation of the quality of the obtained data. The following sections are aimed at conducting such analysis. 5.3.1 Flow Measuring Technique The low precision of the flow field measurements in engines is asso-ciated with the limitations of hot wire anemometry. Many researchers in the past tried to estimate error limits related to the technique. It is, however, difficult to assign an exact number to i t . Published uncer-tainty values range from 10-15% [43] to 50-70% [42]. The latter range should be, however, a motivation for abandoning the technique rather than conducting further analysis. It is recognized that errors in the data obtained with HWA are caused by three factors: lack of directional sensi-tivity of the probe, assumptions used in the analytical scaling of the results to the conditions of probe calibration and by a limited informa-tion about mean fluid temperature. The assessment of the error caused by the last factor is usually performed by sensitivity checks. Dohring [12] estimated uncertainty limits on the mean velocity of the order of 23%. He used different temperature profiles for the velocity evaluation and assessed obtained differences in velocity. This type of error is, however, not the most important in comparative studies presented in this thesis. Rarely a more physical nature of the problem is evaluated. HW probes used in I.C. engines are exposed to a complex environment. During the compression stroke which rapidly decreases the volume of the combus-tion chamber, both temperature and pressure are varying dramatically. At the final stage of the compression, the temperature of the gas is on the order of 300 °C and the wall temperature on the coolant side below 100 °C. This introduces large temperature gradients within the gas contained in the cylinder. It is fairly easy then to imagine a situation where eddies in the mixture are in fact carrying not only velocity but also temperature gradients. The HWA, however, cannot differentiate these effects. As a result an eddy with certain positive velocity gradient carrying mixture of higher than mean temperature would not be detected by the anemometer. The fluctuation would be in turn amplified for eddies carrying temperatures lower than ambient. Such problem can potentially, in the most disadvantageous situation, blur the character of the turbulent fluctuations obtained from the experiment. 42. This error can be avoided i f the exact temperature at the point of velocity measurement is also known. Practically, however, this is impossible on the present level of the experimental technique. The error can be minimized by conducting measurements in areas distant from the chamber walls. In I.C. engines, however, at the end of compression there is not much choice because dimensions decrease rapidly. An improvement can be made by keeping the coolant temperature at a maximum possible level and by conducting measurements in engines motored at lower speeds. The latter would allow enough time for the thermal boundary formation on the chamber walls. The engine speed of 1000 RPM might be, however, already too fast for the boundary layer development. Assuming that the thermal layers are formed there are s t i l l problems with streaks and bursts, normal elements of boundary layer activity. It appears then, that HWA in engine application should be used only for qualitative measurements and rather in the central regions of simple shaped combustion chambers. The velocity trends in those areas are, however, already well known. To gain more information regarding complicated aspects of flow dynamics in engines the only precise technique is LDV. It is recognized that the results of the measurements presented in this thesis have large margins of error. Jet measurements with the short probe are probably indicating much higher velocities than the real values. This is caused by cooling of the gas in the piston channels. Additional errors might have been introduced by a more complicated flow condition existing in the measured region, like vertical or recirculating flows. Considerable attention was drawn to the effect the connecting 43. wire deflection, between the piston and the connecting rod, could have on its change of resistance. There was a possibility that the varying wire resistance could have been interpretated as flow velocity. The possi-bility was small because total wire resistance was only .25 Ohm in comparison to the sensor resistance of 15 Ohms. This was checked, how-ever, by conducting one experiment with a shortcircuited probe located in the piston. No variation of resistance was detected. The confidence limits set by the author on the mean jet velocity is of the order of 30-40%. Turbulence parameters measured with the standard probe carry the ambiguities mentioned before. Most of the measurements were conducted about 5 mm from the bottom deck of the cylinder head because that region was the most important in the project. The analytical technique to scale data to calibration conditions was based on the method suggested by Witze [28] which gave best matching with LDV. To extract turbulence fluctuations from velocity a cycle by cycle nonstationary analysis described by Cattania [43] was used. Window sizes of 5, 10, 12, 15 and 20° crank angle were checked with 12° chosen. During the measurements coolant temperature was kept within 70-80 °C. All sensor repairs were done by the author by spot welding under a microscope. The new welds were carefully analyzed, sensors heat treated at 600 °C for 5 hours, those which showed further decrease of resistance were discarded. It is important, however, to expose another potential source of error in turbulence measurements in engines. This is vibration. The HW probes are usually inserted to the combustion chambers through spark plug fittings. These, depending on their quality, may cause transmission of the engine vibration to the probe's body. An important question at this 44. point is whether the probe's body vibration can induce oscillations of the prongs. The materials used for prongs, in high temperature probes, are usually extremely stiff superalloys. This makes i t rather unlikely that any prong vibration could be induced by a dynamic flow or the probe's body movement. Consequently, the only vibration mode which can be potentially transmitted to the probe is the engine motion. During these experiments the maximum rotational speed of the engine was 50 Hz, substantially below frequency level evaluated as turbulence. An experi-ment is however suggested to confirm the lack of the vibration effect on the HWA turbulence indications. Figure 5.37 presents the sketch of a possible experimental setup. The limit of confidence with respect to the turbulent fluctuations is set at 20-30%. The other measured parameter during the flow experiments was pressure. Pressure measurements is the most precise from measurements used in the I.C. engine environment. In these particular experiments information by Brown and Lancaster [46,47] was taken into consideration. The pressure transducer was recessed from the chamber by half of its diameter. Motored pressure traces were analyzed and peak pressures found 1° crank angle BTDC. No shift was detected. The only questionable point in the pressure evaluation was the assumption of the reference pressure value. It was calculated here as a result of volumetric efficiency and ambient pressure. The error Introduced by this procedure is very small because i t affects only vertical shifting of the pressure curve. Both the pressure transducer and charge amplifier have error specification lower than 1%. Consequently 1% is the error margin set for the pressure evaluation. 45. 5.3.2 Performance Evaluating Technique The accuracy of the results obtained during the performance tests is difficult to assess because of the complexity of the measuring system. The results depend here on many values being measured with various instruments. Flow rates of natural gas and air were measured using laminar flow elements to which calibration curves were supplied by the manufacturer. The air flow element was exposed, however, to a pulsating flow regime and its calibration curve was suspect. An extensive check was carried out and the calibration was confirmed to within 1%. A bigger source of error was found to be caused by the differential pressure transducer which changes pressure into analog voltage. The constant was checked and found to be varying within 3-4% for the natural gas transducer. The air flow transducer was more precise and fluctuated only within 1%. The speed indication was checked with a tachometer and confirmed to be within less than 1% discrepancy. The electronic variable ignition timing system was checked and set to within 1° crank angle over the range of operating conditions. The Data Translation DAS showed some stray offset voltage values of the order of 3 mV returning even after zeroing the system. These values constituted errors of the order of 3%. The most questionable part of the measuring system was the torquemeter. A static check of the dynamometer showed a 1% agreement with the transducer indication, well within the acceptable limits. Dynamic behaviour was, however, very different. The meter rarely returned to zero value after a test was completed. It showed consistently some reading every time i t was powered up. Zeroed before the test the meter displayed torque read-ing after the engine was stopped. The reading varied within +-1.5 Nm. This behaviour raised questions about the accuracy of the instrument 46. during dynamic operation. The torque measurements were very important for the proper evaluation of the performance parameters of the engine. Al l the values like power, BMEP, BSFC, efficiency were evaluated on the basis of torque measurement. Generally, prior to conducting any measurements the system had to be powered up for about 1 hr to warm up the electronic components, then zeroed. As mentioned before each new piston geometry was allowed for 10 hr break in operation, under ful l load, prior to taking any measurements. It is recognized, however, that i t might have been insufficient for some tests, which could have resulted in performance parameters not directly reflecting the quality of the combustion process. Most of the additional errors were of a random nature and time constraints prohibited more careful evaluation of their nature. The limit of confidence on the performance data is estimated within 7-10%. 47. CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS 6.1 Conclusions and Observations 6.1.1 Introduction The objective of this project was to investigate the influence of turbulence enhancement on the performance and combustion behaviour of a modern S.I. engine. It was expected that the use of the squish-jet type combustion chamber would increase the level of turbulence before TDC with some degree of control over its parameters. The work began with a flow visualization experiment which confirmed jet development in the squish-jet combustion chamber. The project was subsequently continued by extensive flow measurements and performance tests on a single cylinder research engine. During the firing tests the engine was fueled with natural gas and the turbulence enhancement was expected to improve the rate of combustion of this slow burning fuel. 6.1.2 Observations A considerable effort was made to evaluate the true nature of the squish-jet effect. The flow visualization experiment and cyclic measure-ments confirmed its development, however, the detected velocities were lower than those predicted. The measurements conducted at the bottom of the piston bowl showed higher values of mean velocities and turbulent fluctuations for the squish-jet piston than for the pure squish design. However, the squish design generated a much higher level of flow activity in the upper part of the piston bowl, close to the spark plug location, before TDC. At 3000 RPM the maximum mean squish velocity in the squish 48. design was found to be 30 m/s and turbulent RMS fluctuation at 8 m/s. The squish-jet design decreased the mean velocity to 15 m/s and turbulent fluctuations to 6 m/s. The ensembled pressure traces obtained during firing tests show slower combustion process and lower peak pressure for the squish-jet combustion chamber in comparison to the ordinary squish design. The well projected trend shows that for higher flow activity at the spark plug location, combustion process proceeds faster and leads to a higher peak pressure. The analysis of IMEP and mass fraction burned suggests that the second half of the combustion process may be advantageous for the squish-jet design, which can lead to improvement of exhaust emissions. Calculated coefficients of variance (COV) of cyclic peak pressure and IMEP were in the same range for a l l designs. Both of these coeffi-cients were influenced by the mixture strength, doubling their values from operation at AFR=1.00 to AFR=1.25 at WOT. The COV of peak pressure was twice the COV of the IMEP at a l l operating conditions. There was no apparent trend detected in the values of these coefficients at different speeds. The effect of squish-jet design on the performance of the engine at WOT was small and within margin of error. At part load operation, however, pure squish design was more efficient by 7% at lean limit. 6.1.3 Conclusions 1. The experimental results suggest that the squish-jet design is not effective in promoting fast combustion because: i) It reduces squish motion and flow activity near the spark plug location before TDC; 49. i i ) It generates jets of low velocity; i i i ) It increases the flow activity only at the bottom part of the piston bowl. 2. The results show that the mixture motion at the spark plug location in an S.I. engine has a pronounced effect on the combustion process. 3. The squish motion increases values of both mean velocity and turbu-lent fluctuation and has indeed positive influence on combustion. It does not affect, however, cyclic variation in combustion. 4. The experimental data shows independence of cyclic variation in peak pressure and IMEP from presence or absence of squish motion. The squish motion introduces variations in the flow field at 20° CA BTDC. During the relevant experiments ignition was set at 30° CA BTDC. This suggests that cyclic variation in combustion originates in the i n i t i a l stage of the flame development. This conclusion is supported by a comparison of mean pressure traces and mass fractions burned for different piston geometries at RAFR=1.00 and at RAFR=1.25. 6.2 Recommendations It is suggested that further work be conducted in order to establish a l l features of the squish-jet design. The future experiments should be concentrated on the following problems: 1. Exhaust emissions should be measured for a squish-jet configuration and squish geometry. It is suggested that the improved second half of the combustion process can reduce levels or exhaust emissions. Another aspect is a potential negative effect the squish-jet design 50. can have through increased quenching areas in the channels. These can also carry a larger fraction of the residuals to the next cycles affecting combustion. 2. Flow measurements should be conducted with a re-entrant-bowl type combustion chamber to check i f the increased squish area has a strong effect on the strength of the jets. 3. Experiments should be performed using a step in the cylinder head, which would close the squish area and force flow through the channels. This feature may offer substantially stronger jet patterns. It is recommended that the last two sets of experiments be conducted on the plexiglass model assembled on the RCM which should simplify the measuring process. Additionally i t is suggested that hot wire measurements be conducted on the Ricardo engine, with a set up similar to the proposed in Chapter 5.31, to confirm no effect of engine vibration on the turbulent spectrum measured with the HWA. 51. REFERENCES 1. ANDREWS, G.E., BRADLEY, D. and LWAKABAMBA, S.B., "Turbulence and Turbulent Flame Propagation - A Critical Appraisal", Combustion and Flame, Vol. 24, pp. 285-304, 1975. 2. DAMKOHLER, G., "The Effects of Turbulence on the Flame Velocities in Gas Mixtures", NACA TM 1112, 1947. 3. TABACZYNSKI, R.J., FERGUSON, CR. and RADNAKRISHNAN, K., "A Turbulent Entrainment Model for Spark Ignition Engine Combustion", SAE 770, 1975. 4. MA, T.H., "Effects of Cylinder Charge Motion on Combustion", Proc. I. Mech. E. C81/75. 5. WITZE, P.O., "The Effect of Spark Plug Location on Combustion in a Variable Swirl Engine", SAE 820044, 1982. 6. WITZE, P.O. and VILCHIS, F.R., "Stroboscopic Laser Shadowgraph Study of the Effect of Swirl on Homogeneous Combustion in a Spark-Ignition Engine" SAW 810226, 1981. 7. ALCOCK, J.F. and SCOTT, W.M., "Some More Light on Diesel Combustion", Proc. I. Mech. E., 1962-63. 8. FITZGEORGE, D. and ALLISON, J.L., "Air Swirl in a Road-Vehicle Diesel Engine", Proc. I. Mech. E., (A.D.), 1962-63. 9. ZIV, A., "Squish Velocity in the Combustion Chamber of a 2-Stroke Cycle Engine", SAE 730186, 1973. 10. NAGAYAMA, I., ARAKI, Y. and LI0KA, Y., "Effects of Swirl and Squish on S.I. Engine Combustion and Emission", SAW 770217, 1977. 11. CLERK, D., "Cylinder Actions in Gas and Gasoline Engines", SAE Journal, Vol. 8, June 1921. 12. D0HRING, K., "The Relative Effects of Intake and Compression Stroke Generated Turbulence in I.C Engine Combustion Duration", UBC Thesis, Report AFL-86-01, 1986. 13. SEMENOV, E.S., "Studies of Turbulent Gas Flows in Piston Engines", English translation: NASA Technical Translation, F97, 1958. 14. HASSAN, H. and DENT, J.C, "Correction for Temperature Loading and High Gas Pressure Effects for the Constant Temperature HWA", J. Phys. D., 1969, 2. 5 2 . 15. DENT, J.C. and DERHAM, J.A., "A Method for Checking the Consistency of Velocity Computations from Hot Wire Anemometer Measurements In Variable Density Flow", Journal of Physics E: Scientific Instruments, 1972, Vol. 5. 16. TINDAL, M.J. and WILLIAMS, N.J., "An Investigation of Cylinder Gas Motion in the Direct Injection Diesel Engine", SAE 770405, 1977. 17. HASSAN, H. and DENT, J.C, "The Measurement of Air Velocity in a Motored Internal Combustion Engine Using Hot Wire Anemometer", Proc. I. Mech. E., 1970-71, 185. 18. DENT, J.C. and DERHAM, J.A., "Air Motion in a Four Stroke Direct Injection Diesel Engine", Proc. I. Mech. E., 1974, V. 188. 19. WITZE, P.O., "Measurements of the Spatial Distribution and Engine Speed Dependence of Turbulent Air Motion in an I.C. Engine", SAE 750886, 1975. 20. LANCASTER, D.R., "Effects of Engine Variables on Turbulence in a Spark-Ignition Engine", SAE 760159, 1976. 21. TABACZYNSKI, R.J., FERGUSON, D.R. and RADHAKRISHNAN, K., "A Turbulent Entrainment Model for Spark Ignition Engine Combustion, SAE 770647, 1977. 22. DENT, J.C. and SALAMA, N.S., "The Measurement of the Turbulence Characteristics in an I.C. Engine Cylinder", SAE 750005, 1975. 23. HAGHG00IE, M., KENT, J.C. and TABACZYNSKI, R.J., "Turbulence Time-Scale Measurement in a Spark Ignition Engine Using Hot Wire Anemometry and Fast Response Ion Probes", Symp. on Flows in I.C Engines, ASME WAM, 1982. 24. LANCASTER, D.R. and KRIEGER, R.B., "Effects of Turbulence on Spark -Ignition Engine Performance", SAW 760160, 1976. 25. SMITH, J.R., "The Influence of Turbulence on Flame Structure in an Engine", ASME, 1982. 26. WIENKE, H.J., MYERS, P.S. and UYEHARA, O.A., "A Resistance Thermometer for Engine Compression Temperatures", SAE 701A, International Summer Meeting, June 1963. 27. MORSE, A.P., WHITELAW, J.H. and YIANNESKIS, M., "Turbulent Flow Measurements by Laser - Doppler Anemometry in Motored Piston -Cylinder Assemblies", ASME, 1979, V. 101. 28. WITZE, P.O., "A Critical Comparison of Hot Wire Anemometry and Laser Dopper Velocimetry for I.C. Engine Applications", SAE 800132, 1980. 29. M0NAGHAN, M.L. and PETTIFER, H.F., "Air Motion and Its Effect on Diesel Performance and Emissions", SAE 810255, 1981. 53. 30. ARCOUMANIS, C, BICEN, A.F., and WHITELAW, J.H., "Measurements in a Motored Four - Stroke Reciprocating Model Engine", Journal of Fluids Engineering, 1982, Vol. 104/235. 31. ARCOUMANIS, C, BICEN, A.F. and WHITELAW, J.H., "Squish and Swirl-Squish Interaction in Motored Model Engines", Journal of Fluids Engineering, 1983, Vol. 105/105. 32. VAFIDIS, C, "Influence of Induction Swirl and Piston Configuration on Air Flow in a Four-Stoke Model Engine", Proc. I. Mech. E., Vol. 198, 1984. 33. FRASER, R.A. and FELTON, P.G., BRACO, F.V. and SANTAVICCA, D.A., "Preliminary Turbulence Length Scale Measurements in a Motored I.C. Engine", SAE 860021, 1986. 34. LUCAS, G.G., "The Effects of Squish on Charge Turbulence and Flame Propagation in a S.I. Engine", I. Mech. E. Conference, London, 1979. 35. SHIMAM0T0, Y. and SKIYAMA, K., "A Study of Squish in Open Chambers of a Diesel Engine", JSME, No. 63, Vol. 13, 1970. 36. WOODS, W.A. and GHIRLANDO, R., "Radial Flow in an Engine Cylinder Near the End of Compression", Proc. I. Mech. E., C65, 1975. 37. CAMERON, C, "An Investigation of Squish Generated Turbulence in I.C. Engines", UBC Thesis, Report AFL-85-02, 1985. 38. EVANS, R.L., "Internal Combustion Engine Squish Jet Combustion Chamber", USA Patent, No. 4, 572, 123, Feb. 25, 1986. 39. KING, L.V., "On Convection of Heat From Small Cylinders in a Stream of Fluid in Determination of hte Convective Constants of Small Platinum Wires with Application to Hot Wire Anemometry", Proc. Roy. Soc, Vol. 214A, No. 14, 1974. 40. C0LLIS, D.C. and WILLIAMS, M.J., "Two Dimensional Convection From Heated Wires at Low Reynold's Numbers", Journal of Fluid Mechanics, Vol. 6, 1959. 41. DAVIES, P.O.A.L. and FISHER, M.J., "Heat Transfer From Electrically Heated Cylinders", Proc. Roy. Soc. A., Vol. 280, 1964. 42. B0ISVERT, J., "Turbulent Combustion of Gas-Air Mixtures in a Spark Ignition Engine", UBC Thesis, Report AFL-86-05, 1986. 43. CATANIA, A.E. and MITTICA, A., "A Contribution to the Definition of Turbulence in a Reciprocating I.C. Engine", ASME 85-DGP-12, 1985. 44. RASSWEILER, G.M. and WITHROW, L., "Motion Pictures of Engine Flames Correlated with Picture Cards", SAE Trans., Vol. 42, No. 5, 1983. 54. 45. AMANN, A.C., "CYlinder Pressure Measurement and Its Use in Engine Research", SAE 852067, 1985. 46. BROWN, W.L., "Methods for Evaluating Requirements and Errors in Cylinder Pressure Measurements", SAE 670008, 1967. 47. LANCASTER, D.R., KRIEGER, R.B. and LIENESCH, J.H., "Measurements and Analysis of Engine Pressure Data", SAE 750026, 1975. 55. APPENDIX A Analytical Evaluation of Squish-Jet Motion a) l-D squish analysis in ideal conditions b) The effect of gas leakage on squish velocity. c) The effect of heat transfer on squish velocity. d) Jet velocity evaluation. a) l-D Squish Analysis in Ideal Conditions Figure 5.38 presents geometry of the Ricardo engine assembly and the dimensions used in these calculations. Instantaneous distance of the piston top from its TDC position: S(p) = R(l+cosp) + L(l-^l-e/sin/p (1) Volume Vg(p): V9(P) = -L"4 -*/S(p)+CL (2) Total volume V T Q T(p): v T 0 T(6) = . [ s ( e ) + C L ] + ( 3 ) Surface A 1(6): A^e) = ird • [S(8) + CL] (4) 56. Assumptions • ideal gas • no friction, heat transfer, leakage • uniform gas density in the combustion chamber From continuity equation dni ^ d V „ r e ) TOT /a\ TOT = P(9) • dt d6 u7+W6> dp(9) d6 d6 * dt - o ( 5 ) 1 dp(6) = p(e) * de v T Q T ( e ) d v ^ r e ) TOT de (6) From continuity for volume V^ : dm^  dF = co [P(6) d v ^ e ) de + V.(9) dp(e)-de = -A^O) • p(6) • U(6) (7) U(6) = - to AjCe) L ae d v ^ e ) VjCe) dv m re) TOTv i v ^ C e ) * de J TOT (8) Substituting; U(6) = - (1-Sq2)«d«h A[S(e)+CL] • [S(6)+CL+h.Sq2] ds(8) de ( 9 ) 57. Continuity equation for the piston chamber: dm. CH dV dt = 0) CH d CH v ( 6 ) = vCH,ai dp(6) _ CH p(9)-A CH de ACH* VT0T ( e ) d v ^ r e ) TOT 5e d i ) Substituting: v(9) = -Sq2 + CL + Sq2 ds(e) de (12) b) The Effects of Leakage on Squish Velocity Assumptions: • uniform flow past the rings can be modelled as a flow through a hole at critical conditions. • flow through the hole is adiabatic. • properties of the gas in the chamber are uniform and change polytropically. No effect of mass change on polytropic coefficient. Decrement of squish velocity because of leakage is: dm u<6> " uL< e> = A l <8) - p(e) • dt - (13) 58. Leaking rate, using critical flow conditions: drn^  dt d m L dt ( x + 1 ) ^2(x-iy • /x«p «p c c - Vsr> r * + 1 i L2(x-l) J /x»p«p o r 0 x+1 V„,™ 2 r TOT -i Lv (6)-1 TOT Substituting into equation 13: .2 2 (D'-D') • /x.R.T U<e>-Ve> = d . [ s ( e H C L j ° ' r x + 1 i l2(x-l) J rR+L+CL+h'Sq2 -j CL+h.Sq^SO) (14) where V^-V^, is the leaking diameter of the rings. c) The Effects of Heat Transfer on Squish Velocity Assumptions: • gas in the cylinder changes polytropically. • wall temperature is uniform and constant during compression. • heat transfer rate through the wall is uniform. • cooling of the gas is by the chamber walls. The equation of energy for the gas in volume Vj^  is: 59. Substituting for dn^/dt: dm -j-i= - T T • d • [S(9)+CL] • ^  Uu(6) d t R.T(6) H We can obtain from 15: d(c T ) A x-1 V ™1 P l dQ i V 6 ) = ~ p(e).ir.d^x.[S(6)+CL] [ _ d t + A dF + P dt~^ ( 1 6 ) PTOT For uniform cooling the above equation will be: d(CvV V, M dV, U ( 9 ) " " p(9).1T.d"x.[S(9)+CL] t dT~ + V T^ T"^" f P dT -] Subtracting (17) and (16): Ke)" Ve> - p(8).,.d.;.[S(6)+cLi [ ( r\-Sr) • f 1 < 1 8 > PTOT The rate of heat loss can be calculated: = A • a(T -T ) dt p m ™ q V rTOT (19) 60. Where can be calculated from the polytropic compression, T w assumed constant, and a calculated from available formulas like Woschni or Annand. Using Woschni formula: -.214 , 0.786 -.525 a = 265 • D (CP) • T (20) where a: kcol/m2hr°C, D: m, C: m/s, P: kg/cm2, T: °K. d) Jet Velocity Evaluation Figure 5.39a presents geometry of the squish-jet combustion chamber. Figure 5.39b shows simple model used in this analysis. Assumptions: • uniform gas density in the combustion chamber • gas undergoes polytropic compression • no effects of uneven cooling and leakage • uniform gas pressure in volume V and in the volume above the piston bowl • no relative effects of squish and jet velocities • gas flow from the squish volume-V^ can be treated as flow through parallel pipes. From continuity equation: J(9): A. • p(6) + U (9) • A.(9) • p(9) = U(9) • p(9) • A,(9) i s i 61. A 1 J(8) • j^Qj + U8(9) = U(0) (21) From assumption of steady flow through parallel pipes: H = H. s T i T Loss Loss Calculating pressure loss from Darcy-Weisbach formulae for: • turbulent flow conditions through the pipes • one pipe representing channel in the piston • no additional losses except friction on the walls f !jL . ! i i = f , L j , J 2 s ' DR ' 2 g i * d 4 ' 2 g Us r f i L i V 1 /2 — = r_i . -i . —2.1 ( 2 2 ) J L f L d.J K ' s s j U (6) f. L. D (8) , / 2 J(8) L f L d. / s j (23) 2Trd[S(e)+CL] W h e r e °H " SoJ-rCL+Ttd The equation (23) suggest that for the same friction coefficient f = f , o(L ) = o(d ), jet velocity is of order similar to squish velocity. However when additional losses are taken into consideration like: entrance loss, elbow loss the equation (23) changes into: U (8) f L D D , _ J L _ = f _ l . _1 .JL + 5 Hn 1 / 2 J ^ L fs Ls d j fs L s J where K is the coefficient of pressure loss. 62, Factors which can potentially decrease jet velocity are: • decrease of pressure along squish surface • inertia of the gas A factor which can increase the jet velocity is: • dynamic effect of the piston motion. APPENDIX B PROGRAM HB (input,output); {$1 FRAME.DEF ) {$1 FRAME.10 ) PROCEDURE Writelntroduction; {- -) BEGIN Cirscr; writelnC MASS BURN RATE ANALYSIS'); writelnC ====r=============rr===») • writein; wr i telnC Introduction'); writelnC '); writeln('This prograi reads cylinder pressure traces recorded froi the ') writelnCRicardo engine and analyses thei to obtain estimates of the iass writelnCburning rates.'); wri te lnC); writeCPress 'return' to continue.... '); read(kbd,ch); END; PROCEDURE Wr iteMenuAndSelectOpt ion; { ) BEGIN cirscr; writein; writelnCThe following options are availab writelnC writelnC writelnC writelnC writelnC writelnC writelnC writelnC writelnC writelnC writelnC e: '); writein; writein; Load Pressure Data Scale Pressure Data Siooth Pressure Data List Pressure Data Align Pressure Data 6: Calculate 6aaia Values 7: List 6aua Values 8: Save Sana Values 9: Differentiate Gaiias 10: MassBurnRate 11: Exit writein; writein; writeCSelect one of the options C l . . 111: '); REPEAT read(option); writeC '); UNTIL option IN E1..111; writein; END; FUNCTION voKca: integer): real; { . ) CONST stroke = 0.0689; bore = 0.0803; length = 0.1580; clrv = 56.50E-6; VAR SinTheta^osTheta^istFroiTDC^rnk^rad^yljadJwoL^weptVol: real; BEGIN. SinTheta:=sin(int(ca+lB0)t0.0175); CosTheta:=cos(int(ca+180)to.0175); crnk_rad:=stroke/2.0; cyl_rad:=bore/2.0; TwoL:=2.Otlength; 0istFroiT0C:=(crnkjad)»(1.0-CosTheta*(crnk.ra(J/TwoL)*SinTh*ta*SinTheta)f SweptVol:=pit(cyljad)t(cyl_rad)tDistFroiTDC; vol:=SweptVol+drv; END; PROCEDURE ReadlnData; { } VAR FileType: integer; BEGIN Clrscr; writeCEnter type of file to be loaded: Binary(l), Bsaved(2) or AsciiO) REPEAT read(FileType); writeC '); UNTIL FileType IN (1,2,31; writeln; writeln; writeln; CASE FileType OF 1: BEGIN uritelnCEnter naie of file to be loaded:'); writeln; LoadBinaryDataCx'.'N1); IF ExitFlag=true THEN exit; writeln; END; 2: BEGIN writelnCEnter naie of file to be loaded:'); writeln; LoadBinaryData('x','B'); IF ExitFlag=true THEN exit; writeln; END; 3: BEGIN writelnCEnter naie of file to be loaded:'); writeln; LoadAsciiDataCx'); IF ExitFlag=true THEN exit; writeln; END; END; delay(2000); END; PROCEDURE ScalePressureData; { } VAR diff,scale,BDCianPres: real; BE6IN clrscr; writelnC SCALE PRESSURE DATA '); writeln; writeCEnter scale factor in bar/volt: '); readln(scale); writeln; writelnC CALCULATING.... »); CASE NoOfDataPoints OF 720: FOR i:=l TO 360 DO axA[i]:=(axA[i+360]/204.8-10)tscale; 1440: FOR i:=l TO 360 DO axA[i]:=(axA[i*2+7201/204.8-10)tscale; 3600: FOR i:=l TO 360 DO axAti]:=(axAC1799+i«]/204.8-10)tscale; END; writeCEnter BDC tanifold pressure in 8ar:');readln(BDCianPres); diff:=BDCianPres-axA[l]; FOR i:=l TO 360 DO axACi]:=axA[il+diff; NoOfDataPoints:=360; END; PROCEDURE SioothPressureData; { . } VAR Nweights,ri^: integer; F1,PAR : real; weights : shortarray; yval,xval : ARRAYC0..440] OF real; PROCEDURE data_stooth_weights(sioothnui: integer; weights : shortarray); { - - } VAR s»oothdeg,niiidat, startconv,stopconv,nuicoef: integer; BEGIN nuidat:=440; s80othdeg:=ssoothnui DIV 2; startconv:=sioothdeg+l; stopconv:=nuidat-s«oothdeg; FOR i:=0 TO nutdat DO xvalli]:=0; FOR i:=startconv TO stopconv DO FOR j:=0 TO saoothnui-1 DO xval[i3:=xval[i]+yvalti-sioothdeg+j]tweightstj+13; j:=sioothdeg; k:=nutdat-sioothdeg+1; FOR i:=l TO saoothdeg DO BEGIN xvalCj]:=xval[j+l]; xvalCkl:=xval[k-11; j:=j-l; k:=k+l; END; END; PROCEDURE DesignFilter(ncoe: integer; fl,par: real; VAR H: shortarray); { } VAR aa,bb,argO,argl,argw, d,ak,afl,psdsui : real; psd,wcoef : shortarray; ncoel,l,nodd,kk : integer; BE6IN nodd:=ncoe-(ncoe DIV 2)t2; ncoel:=ncoe DIV 2; 1:=ncoel+l; aa:=0.5tint(l-nodd); bb:=int(ncoel)-aa; argO:=pi/bb; argw:=1.5tpar/bb; IF noddOO THEN Htll:=2.0tfl; FOR i:=l TO ncoel DO BE6IN k:=l-i; ak:=int(i)-aa; afl:=2lpitaktf1; d:=pitak; HCkl:=sin(afl)/d; argl:=argOlak; wcoef[k]:=exp(parlln(abs(sin(argl)/argl))); HCk]:=Htk]twcoef[k]; END; FOR k:=l TO ncoel DO BEGIN kk:=ncoe+l-k; H[kkl:=HCk]j END; END; BEGIN { StoothPressureData ) cirscr; writelnC DIGITAL FILTERING '); writein; Fl:=0.035; writelnCEnter filter frequency as a percentage of the satple frequency') writeCeg 0.02,0.035,0.060...0.50) [0.035]:.'); readln(Fl); Nweights: =81; writein; writelnC CALCULATING... '); writein; { extend pressure data at both ends so that after filtering the nuiber ) { of data points will be the saie as before. > FOR i:=0 TO 40 DO yvaUi]:=axAtl]; FOR i:=41 TO 400 DO yvalti3:=ax*Ci-401; FOR i: =401 TO 440 DO yval[il:=axA[360J; DesignFilter(Nweights,Fl,1,weights); data_stooth_weights(Nweights, weights); FOR i:=l TO 360 DO BEGIN axA[i]:=xval[i+40]; ayA[i]:=yval[i+40J; END; writelnC CALCULATIONS FINISHED '); delay(1500); { SsoothPressureData ) END; PROCEDURE ListPressureData; { } BEGIN clrscr; writelnC LIST PRESSURE DATA '); writeln; FOR i:=l TO 360 DO BEGIN urit^':((i+360):4,' ',ayA[i]:12:3,' ',axACi]:12:3); IF (i HOD 20)=0 THEN BEGIN writeCPress "return'..'); read(kbd,ch); IF ord(ch)=27 THEN exit; clrscr; writelnC LIST PRESSURE DATA '); writeln; END; END; END; PROCEDURE AlignPressureData; { ) VAR n: integer; BEGIN clrscr; writelnC ALIGN PRESSURE DATA '); writeln; writelnCshift data to right(+) or left(-) by 'n" ' ) ; writeCenter n : '); readln(n); writeln; IF n'M THEN BEGIN FOR i:=360 DOMNTO-1 DO axACi+n]:=axACil; FOR i:=l TO n DO axACil:=axACn+l]; END; IF n<l THEN BEGIN FOR i:=l TO 360-n DO axA[il:=axA[i+n]; FOR i:=360-n+l TO 360 DO axACil:=axA[360-nl; END; writelnC CALCULATING...'); delay(lOOO); END; PROCEDURE CalculateGauas; <-- ) VAR Numerator,Denominator,Gana: real; BEGIN clrscr; writelnC CALCULATE GAMMA VALUES '); writeln; writelnC CALCULATING... '); FOR i:=l TO 359 DO BEGIN IF axAm<>0 THEN BEGIN Numerator:=(axACi+ll/axACil)j Denominator:=(vol(i)/vol(i+l)); IF (Nuierator)O) AND (Denominator>0) AND (DenominatorOl.O) THEN Gana:=ln(Nuierator)/ln(Denoiinator) ELSE 6ana:=0; END ELSE 6aaaa:=0; IF Gaaaa<0 THEN azACi+l]:=0 ELSE IF 6aaaa>3 THEN azACi+ll:=3 ELSE azA[i+l]:=6aaaa; END; writelnC CALCULATIONS FINISHED. '); delay(1500); END; PROCEDURE ListGaiias; { } BEGIN cirscr; writelnC LIST GAMMA VALUES '); writein; FOR i:=l TO 360 DO BEGIN writeln((i+360):4,' ',(vol(i)*iE6):12:3,' », az'ti1:12:3); IF (i MOD 20)=0 THEN BEGIN writeCPress •return"..'); read(kbd,ch); IF ord(ch)=27 THEN exit cirscr; writelnC LIST PRESSURE DATA '); writein; END; END; END; PROCEDURE SaveGanas; { ) BEGIN cirscr; writelnC SAVE GAMMA VALUES »); writein; SaveBinaryData('z'); delay(2000); END; PROCEDURE Differentiate; { } VAR saoothnua,n: integer; PROCEDURE data saooth sg(saoothnua,derivnua: integer); { : : } TYPE saoothcoeftype = array[0..2,1..5,1..13] of integer; noracoeftype = array[0..2,1..5I of integer; CONST saoothcoef : saoothcoeftype = (((-3,12,17,12,-3,0,0,0,0,0,0,0,0), (-2,3,6,7,6,3,-2,0,0,0,0,0,0), (-21,14,39,54,59,54,39,14,-21,0,0,0,0), (-36,9,44,69,84,89,84,69,44,9,-36,0,0), (-11,0,9,16,21,24,25,24,21,16,9,0,-11)), ((-2,-1,0,1,2,0,0,0,0,0,0,0,0), (-3,-2,-1,0,1,2,3,0,0,0,0,0,0), (-4,-3,-2,-1,0,1,2,3,4,0,0,0,0), (-5,-4,-3,-2,-1,0,1,2,3,4,5,0,0), (-6,-5,-4,-3,-2,-1,0,1,2,3,4,5,6)), ((2,-1,-2,-1,2,0,0,0,0,0,0,0,0), (5,0,-3,-4,-3,0,5,0,0,0,0,0,0), (28,7,-8,-17,-20,-17,-8,7,28,0,0,0,0), (15,6,-1,-6,-9,-10,-9,-6,-1,6,15,0,0), (22,11,2,-5,-10,-13,-14,-13,-10,-5,2,11,22))); noncoef : noncoeftype = ((35,21,231,429,143), (10,28,60,110,182), (7,42,462,429,1001)); VAR 5»oothdeg,nuidat, startconv,stopconv, nuicoef : integer; BE6IN IF (sioothnui>=l) AND (sioothnu»<=5) AND (derivnui>=0) AND (derivnui<=2) THEN BEGIN nuadat:=360; nuicoef:=2*sioothnui+3; sioothdeg:=(nuicoef-l) div 2; startconv:=sioothdeg+l; stopconv:=nuidat-siootrideg; FOR i:=0 TO nuidat DO ayAti]:=0; FOR i:=startconv TO stopconv DO BEGIN FOR j:=0 TO nuicoef-1 DO ayA[i]:=ayA[i]+azA[i-sioothdeg+j]*s«oothcoef[derivnut,si>oothnui, j+11; ayAf.i]:=ayA[i]/noricoef [derivnui,sioothnui]; END; j:=sioothdeg; k:=nuidat-s»oothdeg+l; FOR i:=l TO sioothdeg DO BEGIN ayA[j]:=ayA[j+l]; ayA[k]:=ayACk-l]; j:=j-l; k:=k+l; END; END; PROCEDURE ListDiffGanas; { } BEGIN cirscr; writelnC LIST DIFFERENTIATED GAMMA VALUES •); writein; FOR i:=l TO 360 DO BEGIN writeln((i+360):4,' »,azA[ih 12:3,r ',ayA[i]:12:3); IF (i MOD 20)=0 THEN BEGIN writeCPress "return"..*); read(kbd.ch); IF ord(ch)=27 THEN exit; cirscr; writelnC LIST DIFFERENTIATED GAMMA VALUES '); writein; END; END; END; BEGIN { Differentiate ) cirscr; writelnC DIFFERENTIATE GAMMA VALUES '); writein; writeCEnter degree of snoothing to be done before differentiating:'); readln(sioothnui); writeln; writeCDifferentiate once(l) or twice(2): '); readln(n); writelnC CALCULATING.. '); writeln; data_s>ooth_sg(sioothnui,n); writelnC CALCULATIONS FINISHED. '); writeCList the results?, (y/n):'>; readln(ch); IF upcase(ch)='Y' THEN ListDif fGanas; { Differentiate > END; PROCEDURE MassBurnRate; { ) VAR startcoib,endcoib,duration: integer; CoibGana,CoibPress,r1assBurned: ARRAYC1..360] OF real; PROCEDURE FindStartCoib; { J BEGIN writeCEnter spark tise in degrees before TDC: '); readln(StartCoib); StartCoab:=180-StartCotb; END; PROCEDURE FindEndComb; { ) BEGIN FOR i:=250 DOUNTO 190 DO IF abs(ay*m)>0.0t THEN BE6IN endcoib:=i; exit; END; END; PROCEDURE CalcCoabGawa; { ) VAR slope: real; BEGIN duration:=endcoib-startcoib; slope:=(az*Cendcotb]-a2A[startco«bl)/int(duration); FOR i:=startcoab TO endcotb DO BE6IN Co«b6aiiaf.i]:=int(i)tslope+azA[startco«b]; writeln(i:3,' '(Co«bGana[i]:10:4); END; writei'Press return to continue..'); readln(ch); END; PROCEDURE CalcHassBurned; { ) VAR Pl,P2,VlfV2,SuiDeltaP,DeltaP,FinalPress: real; BEGIN SutDeltaP:=0.0; FOR i:=startcomb TO endcoib DO BE6IN Vl:=vol(i); V2:=vol(i+l); Pl:=axA[il; P2:=axAti+ll; DeltaP:=P2 - PUexp(CotbGaMaCiltln(Vl/V2)); writeln(i:3,' ',DeltaP:10:5); SuiDeltaP:=SutDeltaP+DeltaP; CoibPressU]:=SuiDeltaP; END; FinalPress:=Co»bPressti]; writelnCFinalPress = '.FinalPress:12:5); FOR i:=startcoib TO endco»b DO HassBurned[i]:=CoibPress[i3/FinalPress; END; PROCEDURE ListHassBurned; { ) BEGIN cirscr; writelnC LIST MASS FRACTION BURNED '); writein; j:=0; FOR i:=startcoib TO endcoab DO BEGIN writeln(i:4,* \Co«bPress[i3:12:3,' •,HassBurned(i]:t2:6); j:=j+l; IF (j=20) OR (i=endcoab) THEN BEGIN writeCPress •return*..'); read(kbd.ch); IF ord(ch)=27 THEN exit cirscr; writelnC LIST MSS FRACTION BURNED '); writein; END; j:=0; END; writeCPress return to continue..'); readln(ch); END; PROCEDURE SaveHassBurned; { --) BEGIN cirscr; writelnC SAVE MASS BURNED VALUES '); writein; FOR i:=l TO startcoib-1 DO ayA[ih=0.0; FOR i:=startcoib TO endcotb DO ayA[i3:=HassBurned[i]; FOR i:=endcoib+l TO 360 DO ayA[ih=1.0; SaveBinaryData('y'); delay(2000); END; BEGIN { NassBurnRate ) cirscr; writelnC CALCULATE MASS FRACTION BURNED'); writein; FindStartCoib; FindEndCoib; writeCEnd of coabustion occured at ',endcoib-180,' degrees ATDC); writeCPress return to continue..'); readln(ch); CalcCoibGana; CalcMassBurned; writeCList the results? (y/n):')( readln(ch); IF upcase(ch)='Y' THEN ListHassBurned; writeCSave the results? (y/n):'); readln(ch); IF upcase(ch)='Y' THEN SaveNassBurned; { NassBurnRate } END; BEGIN { -MB-new(ax); new(ay); new(az); Writelntroduction; Writein; vriteCSelect »enu options(l) or run auto»atically(2): '); readln(i); IF i=i THEN BEGIN REPEAT ExitFlag:=false; WritetlenuAndSelectOption; CASE Option OF 1: ReadlnData; 2: ScalePressureData; 3: SioothPressureData; 4: ListPressureData; 5: AlignPressureData; 6: CakulateGaaias; 7: List6aaaas; 8: S a v e 6 a n a s ; 9: Differentiate; 10: NassBurnRate; 11: BEGIN uriteCare you sure? '); readln(ch); IF upcase(ch)='Y' THEN ExitFlag:=true; END; END; UNTIL ExitFlag; END ELSE BEGIN ReadlnData; ScalePressureData; SioothPressureData; CalculateGaiaas; Differentiate; NassBurnRate; END; { n B > 72. Table 1. Engine Specifications Bore Stoke Swept Volume Compression Ratio Clearance Value Squish Area Intake Valve Opens Intake Valve Closes Exhaust Valve Opens Exhaust Valve Closes Spark Plug Type 80.26 mm 88.90 mm 450 ml 9:1 1 mm 70 % 12 BTDC 56 ABDC 56 BBDC 12 ATDC Champion Type AGYC Table 2. Hot Wire Probe Specifications Model Sensor Material Sensor Diameters Sensor Length Temperature Coefficient of Resistance Sensor Operating Temperature Bridge Anemometer Signal Conditioner TSI Model 1226 Platinum Iridium 6.3 urn 1.25 mm 0.0009/ °C 600 °C DANTEC CT01 DANTEC Model 56C17 DANTEC Model 56N20 Table 3. Natural Gas Composition Component Vol. % Methane 94.00 Ethane 3.30 Propane 1.00 Butane 0.30 Nitrogen 1.00 Carbon Dioxide 0.30 Lower Heating Value 48,558 kJ/kg : C o m P a r i s o n of the Engine Performance for Dif f e r e n t Piston Geometries at WOT , 3000 RPM , RAFR =1.01 , MBT Enrjl ne_sneed_=_3000_RPM RAFR = 1.01 Piston No. Speed ( r / s ) Power (IcW) BMEP (bar) BSFC (g/kWhr E f f . ) (%) Torque (Nm) Ign.Ad\ (deg) A l r f l . (g /s ) N .G. f l (g /s ) RAFR 2 49.51 8.67 7.79 253.64 29.23 27.92 30 10.27 .61 1.01 3 49.75 8.19 7.37 265.34 27.94 26.41 30 10.22 .60 1.01 5 49.32 8.35 7.53 264.93 28.04 26.95 29 10.32 .61 1.01 7 49.50 8.31 7.46 262.93 28.21 26.74 30 10.33 .61 1.01 4 49.46 8.42 7.67 260.82 28.31 27.46 31 10.35 .61 1.01 41 49.07 8.58 7.77 261 J? 28.43 27.8? 28 10.45 .62 1.01 Table 5 : Comparison of the Engine Performance for Different Piston Geometries, at WOT , 3000 RPM , RAFR = 1.26 , MBT Engine_speed_=_3000_RPM RAFR =1.26 Piston No. Speed ( r / s ) Power (kW) BMEP (bar) BSFC (g/kWhr E f f . ) % Torque (Nm) Ign.Ad\ (deg.) A i r f l . (g /s ) N.G. f l (g /s ) AFR 2 49.58 7.62 6.85 242.1? 30.6? 24.22 35 10.72 .51 1.25 3 49.34 7.23 6.55 247.7? ?9.96 23.46 33 10.51 .50 1.26 5 49.15 7.49 6.77 242.3t 30.61 24.23 34 10.66 .50 1.26 7 49.65 7.33 6.57 247.or 29.96 23.54 32 10.62 .50 1.26 4 49.6? 7.28 6.65 251 .?f 29 .43 23.6S 32 10.63 .50 1.26 41 49.33 7 .49 6.75 244.31 30.3? 24.19 34 10.68 .51 1.26 Table__6_ : Comparison of the Engine Performance for Di f f e r e n t Piston Geometries, at WOT , 2100 RPM , RAFR = 1.02 , MBT Engine speed = 2100 RPM RAFR = 1.02 P i s t o n Speed Power BMEP BSFC E f f Torque Ign .Adv A1r f l . N . G . f l . AFR No. ( r / s ) (kw) (bar) g/kWhr! (%) (Nm) (deg. ) (g / s ) (g / s ) 2 34.92 5.85 7.45 247.17 30 12 26.75 27 6.90 .40 1.02 3 34.75 5.63 7.13 258.59 28 .78 25.54 26 6.81 .40 1.02 5 34.62 5.73 7.36 251.43 29 54 26.36 28 6.82 .40 1.02 7 34.84 5.76 7.35 254.55 29 24 26.31 26 6.90 .41 1.02 4 34.76 5.62 7.50 265.32 2R 92 26.92 28 6.90 .41 1.02 41 34.62 5.86 7.53 249.60 29 73 27.23 28 6.91 .41 1.02 '' Comparison of the Engine Performance for Different Piston Geometries , at WOT , 2100 RPM , RAFR = 1.27 , MBT Engine speed : 2100 RPM RAFR = 1.27 P i s t o n No Speed ( r / s ) Power (MO BMEP (bar) BSFC (g/kWhr E f f . ) (%) Torque (Nm) Ign.Adv (deg. A1r f 1 . ( g / s ) N . G . f l ( g / s ) AFR 2 34.82 5.07 6.47 238.56 31.12 23.22 33 7.12 .34 1.27 3 34.75 4.86 6.23 247.43 29.97 22.31 32 7.11 .33 1.27 5 34.74 5.03 6.44 238.61 31.10 23.13 31 7.07 .33 1.27 7 34.82 5.18 6.62 233.65 31.73 23.74 29 7.12 .34 1.27 4 34.82 4.96 6.48 242.54 30.24 23.12 30 7.14 .33 1.27 41 34.76 5.11 6.55 239.37 31 .00 23.56 32 7.21 .34 1.27 Table_8 : Comparison of the Engine Performance for Different Piston Ceometries , at WOT , 1200 RPM , RAFR = 1.03 , MBT Engine speed = 1200 RPM RAFR = 1 .03 Piston No. Speed ( r / s ) Power (kw) BMEP (bar) BSFC (rj/kwhr E f f . ) U) Torque (Nm) Ign.Adv (deg) A1r.fl (g/s) N.G. f l . (g/s) AFR 2 19.83 3.07 6.87 261.63 28.34 24.60 20 3.85 .22 1.03 3 19.82 2.96 6.66 270.50 27.41 23.75 17 3.90 .22 1.03 5 19.74 3.01 6.78 262.70 28.22 24.32 18 3.79 .22 1.03 7 19.86 3.09 6.92 258.7? 28.66 24.77 17 3.82 .22 1.03 4 19.68 2.98 6.77 26B.4? 28.02 24 .42 17 3.83 .22 1.03 41 19.84 3.07 6.88 264.43 28.04 24.65 19 3.89 .23 1.03 Table 9 : Comparison of the Engine Performance for D i f f e r e n t Piston Ceometries , at WOT , 1200 RPM , RAFR = 1.28 , MBT Engine speed_=_1200_RPM RAFR = 1.28 P i s t o n No. Speed ( r / s ) Power (kW) BMEP (bar) BSFC (g/kWhr E f f . ) (%) Torque (Nm) Ign .Ad' (deg) A i r . f l . (G/s) N . G . f l ( g / s ) AFR 2 19.92 2.60 5.82 255.3] 29.04 20.83 24 3.99 .18 1.28 3 19.96 2.62 5.83 261.0'. 28.40 20.86 22 4.07 .19 1.28 5 19.82 2.67 5.98 250.3! 29.63 21.41 23 3.98 .19 1.28 7 19.72 2.64 5.94 248.8^ 29.80 21.29 21 3.95 .19 1.28 4 19.72 2.48 5.50 274.2' 27.82 20.02 22 4.01 .19 1.28 41 19.93 2.61 5.59 267.6 27.74 20.04 23 4.03 .19 1.28 Table 10 : Comparison of the Engine Fuel Consumption for Different Piston Geometries , at Part Load Encj2ne_speed__: 2000 RPM Par t load BMEP = 2.5 bar X ^ A F R P i s t o n noSv 1.01 1.05 1.10 1.15 1.19 1.25 1.30 2 355 350 346 342 341 345 348 3 360 367 375 367 367 365 370 5 369 367 350 346 342 334 341 7 346 347 340 335 330 320 325 4 347 345 336 327 330 347 336 41 355 360 331 320 329 320 335 lableJA : Ensembled Peak Pressure and Standard Deviation of Cyclic Peak Pressure for Different Piston Geometries, at 2100 RPM and 3000 RPM Engine Speed : 3000 RPM , WOT , MBT Piston No. Ensembied Peak Press (kPa) Std .Dev. o f c y c l . Peak(kPa) % Ensenbled Peak Press, (kPa) Std .Dev. of c y c l . Peak(kPa) (V /c RAFR = 1.00 RAFR = 1.25 2 4575.6 278.8 6 4184.2 291.3 7 3 4655.3 251.5 5.4 4241.4 331.1 7.8 5 4621.6 241.2 5.2 4341.2 270.3 6.2 7 4616.1 216.9 4.7 4260.2 282.8 6.6 4 4736.6 196.8 4.1 4319.9 271.8 6.3 41 4887.5 231.8 4.7 4461.2 481.2 10.8 Engine Speed : 2100 RPM , WOT , MBT Piston No. Ensembled Peak Press, (kPa) Std .Dev. of cycl . Peak(kPa) % Ensembied Peak Press . (kPa) Std.Dev. of cycl . Peak(kPa) % RAFR = 1.00 RAFR = 1 . 2 5 2 4329.9 246.4 5.7 3852.7 385.7 10.0 3 4165.6 214.1 5.1 3917.7 284.9 7.3 5 4318.3 212.0 4 .9 3837.2 285,9 7.5 7 4314.9 208.5 4.8 3948.0 300.0 7.6 4 4361.6 211.8 4.8 3920.0 285.9 7.3 41 4614.0 208.3 4.5 4229.8 282.0 6.7 Table 12 : Ensembled Peak Pressures and Standard Deviations of Cyclic Peak Pressures for Different Piston Geometries Engine Speed 1200 RPM , WOT , MBT P i s t o n Mo. Ensembied Peak P re s s . (kPa) S td .Dev . o f c y c l . Peak(kPa % Ensembied Peak P re s s . (kPa) S td .Dev . o f c y c l . Peak(kPa) RAFR = 1.00 RAFR = 1.25 2 3900.9 207.1 5.3 3290.3 338.9 10.3 3 3910.5 171.8 4 .4 3518.9 283.3 8.1 5 4006.3 182.5 4 .5 3523.5 243.8 6.9 7 3893.9 200.5 5.1 3566.1 291.3 8.2 4 3911.9 179.7 4 .6 3618.5 167.3 4 .6 41 4231.3 221.0 5.2 3472.5 284.2 8.2 82. Table J_3_: Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries , at 3000 and 2100 RPM Engine Speed : 3000 RPll , WOT , MBT P i s t o n No. Ensembied IMEP (kPa) S td .Dev . o f cycl. IMEP(kPa) % Ensembled IMEP (kPa) S td .Dev . of cycl . IMEPCkPa! % RAFR = 1.00 RAFR = 1.25 2 811.2 8.5 1.1 729.2 21.8 3.0 3 781.9 9.2 1.2 716.5 25.6 3.6 5 816.1 12.9 1.6 744.4 19.5 2.6 7 798.5 9.3 1.2 720.0 19.0 2.6 4 778.5 12.6 1.6 715.8 23.3 3.2 41 823.3 10.2 1.2 733.6 25.6 3.5 Engine Speed : 2100 RFM , WOT , MBT P i s t o n Mo. Ensembied IMEP (kPa) Std.Dev o f cycl. IMEP(kPa; % Ensembi ed IMEP (kPa) S td .Dev . of cycl. IMEP(kPa) 0/ ,} RAFR = 1.00 RAFR = 1.25 2 748.4 11.8 1.6 668.8 26.0 3.9 3 738.6 9.0 1.2 660.7 16.6 2.5 5 760.1 11.6 1.5 682.3 22.0 3.2 7 759.7 7.5 1.0 686.7 20.9 3.0 4 720.7 10.8 1.5 665.1 21.3 3.2 41 762.9 13.0 1.7 687.4 15.8 2.3 lableJA ' Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries, at 1200 RPM Engine Speed : 1200 RPM , WOT , MBT Piston No. Ensembled IMEP (kPa) Std.Dev. of cycl . IMEP(kPa) % Ensembied IMEP (kPa) Std.Dev. of cycl . IMEP(kPa) % RAFR = 1.00 RAFR = 1.25 2 692.5 13.6 2.0 604.3 17.9 2.8 3 685.6 10.7 1.6 617.1 15.7 2.5 . 5 695.0 11.2 1.6 633.4 13.5 2.1 7 706.6 a . i 1.1 627.5 13.4 2.1 4 666.2 4.8 n.7 603.1 11.9 2.0 41 705.1 13.2 1.9 621.1 17.5 2.8 84. F i g . 1.1 Obstructions on the valves Generating S w i r l a. Shrouded Valve b. Vortex Valve F i g . 1.2 Squish Combustion Chamber F i g . 2 . 1 Squish-Jet Design Original steel c y l i n d e r U n e r and piston of R.C.M. Electronic counter High Speed camera F i g . 2.2 Flow V i s u a l i z a t i o n E x p e r i m e n t a l Ret-up co cn F ig . 2.3 Photograph of the Plex ig las Model F ig . 2.4 Photographs of the Jet Development F ig . 2.5 Photographs of the Jet Development 90. F i g . 3.1 Cross Section of the Redesigned Ricardo Engine F ig . 3.2 Photograph of the Cast and Machined New Cyl inder Head 92. F ig . 3.3 Photograph of the Ricardo Engine and New Pistons 9 3 . F ig . 3 . 4 Photograph of the New Hot Wire Probe 94. F i g . 3.6 Photograph of the Connecting Rod with the Linkage Mechanism and the Inside View of the Piston F ig . 3.7 Photograph of the Probe Posi t ion in the Piston and the Connection between the connecting rod and the Piston Exhaust Press .Trans Amp. HI.' Anemometer & F i l t e r D .A .S . DATA TRANSLATION D.A .S . ISAAC 2000 14: I sr ! PC VAX 11/750 1. Temp. of exhaust 2. P ress , of exhaust 3. Temp, of intake 4. P ress , of intake 5. Intake a i r f lowrate 6. Temp, of nat . gas 7. Press . of nat . gas 8. Mat. gas f lowrate 9. Var iab le i g n . t i m i n g 10. Tr igger . 2 c . a . & BDC 11. Speed 12. Torque 13. H.W.A.signal 14. H.W.A.signal 15. Pressure 15. AVL crank angle meter 17. D i f f . p ress , t r a n s . 18. D i f f . p r e s s , t r a n s . 3.8 Schematic of the Data A c q u i s i t i o n System Piston n o . l Piston no.2 4 channels 05/32 Piston no.3 4 channels t> 5/32 Piston no.5 4 channels 0 5/32 Piston no.4 4 channels 0 5/32 Piston no. 6 8 channels 0 5/32 P iston no.7 8 channels p 5/32 Piston no.8 4 channels 0 5/32 4 channels 0 12/64 F i g . 4.1 Schematic of the Geometries for the Flow Experiments 99. F i g . 4.2 Hot Wire Probe Positions Across the Piston Bovl 100. Piston no.2 P iston no.3 8 channels CR = 9 .0:1 4 channels CR = 9.0:1 Piston no.5 Piston no.7 8 channels CR = 9 .0:1 8 channels CR = 9.0:1 Piston no.4 Piston no.41 CR = 9 .0:1 CR = 9 .1 :1 F i g . 4.3 Schematic o f the Geometries f o r the Combustion T e s t s Fig. 5.1 Comparison of Cyclic Velocity Profiles for Piston 5 and 4 at 3000 RPM 102. 25-i 25 20 6 0 8 0 1 0 ° '20 140 160 180 2 0 0 2 2 0 24 0 260 2 8 0 3 0 0 320 CRANK ANGLES, 180=TDC PISTON 6 , 3000 RPM F i g . 5.2 Comparison of C y c l i c V e l o c i t y P r o f i l e s for Piston 5 and 6 at 3000 RPM 103. *5l 40 H F i g . 5.3 Comparison of C y c l i c Jet V e l o c i t y P r o f i l e s for Pistons 1 , 2 and 7 , at 3000 RPM 104. Fig. 5.4 Comparison of Ensembled Jet velocity Profiles for Pistons 1 , 2 and 7 , at 3000 RPM 105. 30 F i g . 5.5 Comparison of Ensembled Jet V e l o c i t y P r o f i l e s for Pistons 3 , 4 and 5 , at 3000 RPM Legend PISTON 5 PISTON 6 - i 1 1 1 i 1 1 1 1 1 1 r i i 60 80 100 120 140 160 180 200 220 240 260 280 300 320 CRANK ANGLES, 180=TDC F i g . 5.6 Comparison of Ensembled Jet Velocity P r o f i l e s for Pistons 5 and 6 at 2100 RPM 45-1 OH 1 1 1 1 1 I 1—••—i 1 1 1 r — T ' 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 CRANK ANGLES 180=TDC F i g . 5.7 Comparison of Two Cycles of the Jet Velocity Measured i n Piston 5 at 3000 RPM o 108. S3 A g 2o-> 15-120 CRANK ANGLES. 180=TDC 1200 RPM Fi g . 5.8 Comparison of C y c l i c Jet Ve l o c i t y P r o f i l e s Measured i n Piston 5 at Speeds: 3000 , 2]00 and 1200 RPM 45-1 F i g . 5.9 Comparison of C y c l i c Jet Velocity P r o f i l e s for Piston 5 at Speeds 3000 , 2100 and 1200 RPM o 25-1 O i 1 i i i 1 i 1 i 1 1 i i 1 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 CRANK ANGLES, 180=TDC F i g . 5 . ] 0 Comparison of Ensembled Jet Velocity P r o f i l e s Measured i n Piston 5 at Speeds 3000 , 2]00 and ]200 RPM oH 1 1 1 1 1 1 1 1 1 1 1 1 1 1 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 2 6 0 2 8 0 3 0 0 3 2 0 CRANK ANGLES, 180=TDC F i g . 5.11 C o m p a r i s o n o f C y c l i c J e t V e l o c i t y P r o f i l e s M e a s u r e d i n P i s t o n 5 and 8 a t 3000 RPM 5-0 I i 1 1 1 1 I I —I 1 1 !—I 1 1 4 0 6 0 8 0 100 120 140 160 18 0 2 0 0 2 20 240 26 0 2 8 0 3 0 0 320 CRANK ANGLES, 180=TDC F i g . 5.12 Comparison of Ensembled Mean Velocity P r o f i l e s for Four Piston Geometries at 3000 RPM , Top Probe Position 12-1 u 1 1 \ 1 1 1 1 1 1 1 1 1 1 I [ 40 6 0 8 0 100 120 140 160 18 0 2 0 0 2 2 0 240 2 6 0 2 8 0 3 0 0 320 CRANK ANGLES, 180=TDC F i g . 5.13 Comparison of Ensembled Turbulent Fluctuations for Four Piston Geometries at 3000 RPM , Top Probe Position CRANK ANGLES, 180=TDC F i g . 5.14 Comparison of Ensembled Mean Velocity P r o f i l e s for Piston 1 , Top Probe P o s i t i o n , at 3000 , 2100 and 1200 RPM 35 5 -0-| 1 1 1 r 1 1 i 1 1 1 1 1 i 1 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 CRANK ANGLES, 180=TDC F i g . 5.15 Comparison of Ensembled Mean Velocity P r o f i l e s for Piston A , Top Probe P o s i t i o n , at 3000 and 2100 RPM 12 F i g . 5.16 Comparison of Ensembled Turbul en t Fluctustions for Pis ton 4 , Top Probe Pos i t ion > at 3000 and 2100 RPM 35-i Fig. 5.17 Comparison of Ensembled Mean Velocity for Piston 4 Measured Across the Piston Bowl at 3000 RPM 12-1 140 160 180 200 220 CRANK ANGLES, 180=TDC 320 F i g . 5.18 Comparison of Ensembled Turbulent Fluctuations for Piston 4 Measured Across the Piston Bowl at 3000 RPM co 35-1 u n ! ( ! ! j , , , , j j ( ( f 4 0 6 0 8 0 100 120 140 160 18 0 2 0 0 2 2 0 24 0 2 6 0 2 8 0 3 0 0 320 CRANK ANGLES, 180=TDC F i g . 5.19 Comparison of ensembled mean v e l o c i t i e s f o r p i s t o n 1 at 3000 RPM, measured a c r o s s the p i s t o n bowl 12 F i g . 5.20 Comparison of Ensembled Turbulent Fluctuations for Piston 1 Measured Across the Piston Bowl at 3000 RPM ro o F i g . 5.21 Comparison of C y c l i c Velocity P r o f i l e s for Piston 1 and 4 at the Bottom Probe P o s i t i o n at 3000 RPM 12 140 160 180 200 220 CRANK ANGLES, 180=TDC 240 260 280 300 320 F i g . 5.22 Comparison of ensembled f l u c t u a t i o n s f o r p i s t o n 1 and 4 a t 3000 RPM, measured a t the bottom of the p i s t o n bowl ro ro 35 u n 1 1 i 1 1 1 1 1 1 1 1 1 1 i 4 0 6 0 80 100 120 140 160 18 0 2 0 0 2 2 0 24 0 2 6 0 2 8 0 3 0 0 320 CRANK ANGLES, 180=TDC F i g . 5.23 Comparison of the Ensembled Velocity P r o f i l e s for Pistons 1 and 4 Middle Probe Position at 3000 RPM 124. 5000 0-J 1 1 1 — T — 1 | f 120 140 160 180 200 220 240 260 C R A N K A N G L E S F i g . 5.24 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at WOT , 3000 RPM , RAFR = 1.00 , MBT 125. 5 0 0 0 0 - | , , - , T= , , 1 1 2 0 1 4 0 1 6 0 1 8 0 2 0 0 2 2 0 2 4 0 2 6 0 CRANK ANGLES Fig. 5.25 Comparison of the Ensembled Pressure Traces for Piston Geometries No. 3 , 4 , 5 at WOT , 3000 RPM , RAFR = 1.00 1 2 6 . 5000 OH 1 1 1— r- 1 1 — T 120 140 160 180 200 220 240 260 CRANK ANGLES F i g . 5.26 Comparison of the Ensembled Pressure Traces for Pistons 2 , 5 , 7 at WOT , 3000 RPM , RAFR = 1.00 , MBT 127. 5000-r 450CH 4 0 0 0 H 3500-1 ^ 3000-1 LxJ o : ZD 2 5 0 0 - | Ul Ul U J cn Q L 2000-1 1500-1 1000 500 Legend PISTON 2  PISTON 3 PISTON 4 o-r-120 140 160 180 200 220 240 260 C R A N K A N G L E S F i g . 5.27 Comparison of the Ensembled Pressure Traces for Five Pistons at 3000 RPM , RAFR = 1.25 , WOT , MBT 128. 5 0 0 0 4500 4 0 0 0 3 5 0 0 H p 3 0 0 0 LxJ cn ZD 2 5 0 0 CO oo UJ Q _ 2000 1500 1 0 0 0 H 5 0 0 H Legend PISTON 4  PISTON 2  PISTON 5 PISTON 5 PISTON 7 T 120 140 160 180 200 220 CRANK ANGLES 240 260 F i g . 5.28 Comparison of the Ensembled Pressure Traces for Five Pistons at 2100 RPM , WOT , RAFR = 1.00 , MBT 129. 5000 4500 120 140 160 180 200 220 CRANK ANGLES 240 260 F i g . 5.29 Comparison of the Ensembled Pressure TRaces for Five Piston Geometries at 2100 RPM , WOT , RAFR =1.25 , MBT 130. 5000 4500 4000 3 5 0 0 H O 3 0 0 0 H CH ZD 00 00 2500 Cr: •_ 2000 1500H 1000 500 120 140 Legend PISTON 2  PISTON 3  PISTON 4 PISTON 5 PISTON 7 I I 1 I I ' 160 180 200 220 240 260 CRANK ANGLES F i g . 5.30 Comparison of the Ensembled Pressure Traces for Five Piston Geometries at 1200 RPM , WOT , RAFR = 1.00 , MBT 131. 5000 4500 4000 Legend PISTON 2  PISTON 3  PISTON 4 PISTON 5 PISTON 7 3500H O 3 0 0 0 H Q_ _¥ Ld cn 2500H 00 00 cn Q_ 2 0 0 0 H 1500H ioooH 500H 0 _ | _ ! , , j ! ! f 120 140 160 180 200 220 240 260 C R A N K A N G L E S F i g . 5.31 Comparison of the Ensembled Pressure TRaces for Five Piston Geometries at 1200 RPM , WOT , RAFR = 1.25 , MBT o D-U J ZD (/> (/> Ul CZ CL 5 0 0 0 n 4500 H 4000 3500 3000H 2500H 2000 1500 H 1000 500 180 200 220 CRANK ANGLES 3000 RPM, RAFR = 1.00 24 0 260 O CL -X UJ CZ ZD in to u i cz Q-5000 4500 4000 3500 3000 2500 H 2000 1500 1000 500H o-t Legend PISTON 2 '20 140 160 180 200 220 CRANK ANGLES 3000 RPM, AFR = 1.25 240 260 F i g . 5.32 Comparison of Ensembled P r e s s u r e T r a c e s f o r F i v e P i s t o n G e o m e t r i e s a t 3000 RPM, WOT, MBT, f o r RAFR = 1.00 and RAFR = 1.25 GO ro 5000 - r 4500H 4000 3500H p 3000 H 0+- — i i I l 1— 120 140 160 180 200 220 CRANK ANGLES 5000 4500 240 2fi0 120 140 160 180 200 220 240 260 CRANK ANGLES 2100 RPM, RAFR = 1.00 2100 RPM, RAFR = 1.25 F i g . 5.33 Comparison of Ensembled P r e s s u r e T r a c e s f o r F i v e P i s t o n G e o m e t r i e s at 2100 RPM, WOT, MBT, f o r RAFR = 1.00 and RAFR = 1.25 O J O J 134. Fig. 5.34 Comparison of Mass Fraction Burned for Piston 3 and 7 at WOT , RAFR = 1.25 , MBT and Speeds: 3000 , 2000 and 1200 RPM 135. 120 140 160 180 2 0 0 2 2 0 2 4 0 2 6 0 CRANK ANGLES F i g . 5.35 Comparison of Mass Fraction Burned for Five Piston Geometries at 3000 RPM , WOT , MBT , RAFR =1.00 Fig. 5.36 Comparison of Mass Fraction Burned for Five Piston Geometries at 3000 RPM , RAFR = 1.25 , W O T , M B T F i g . 5.3 7 Schematic of the Suggested Experiment f o r the E v a l u a t i o n of V i b r a t i o n E f f e c t s on T u r b u l e n c e Measured v i t h HWA 138. A l - s u r f a c e above p e r i m e t e r of the p i s t o n bowl A C H " s u r f a c e of the i n l e t to the p i s t o n bowl B - d i s t a n c e between c e n t e r and top of the p i s t o n CL - c l e a r a n c e between top of the p i s t o n and c y l i n der head (6D - b o r e d iame te r 0d - p i s t o n bowl d i a m e t e r L c o n n e c t i n g r o d l e n g t h R c r a n k r a d i u s S s t r o k e S ( 6 ) - i n s t a n t a n e o u s s t r o k e U - s q u i s h v e l o c i t y V v e r t i c a l v e l o c i t y J - j e t v e l o c i t y 8 - c r a n k a n g l e s from BDC F i g . 5.38 Geometry of the E n g i n e f o r A n a l y t i c a l E v a l u a t i o n 139. F i g . 5 .39 S c h e m a t i c s o f the g e o m e t r y f o r the j e t c a l c u l a t i o n : a) c o m b u s t i o n chamber g e o -m e t r y b) model f o r c a l c u l a t i o n s 

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.831.1-0097090/manifest

Comment

Related Items