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


831-UBC_1987_A7 D95.pdf [ 11.56MB ]
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

Full Text

THE EFFECTS OF TURBULENCE ENHANCEMENT ON THE PERFORMANCE OF A SPARK-IGNITION ENGINE  By ROBERT DYMALA-DOLESKY Mgr.Inz., The T e c h n i c a l U n i v e r s i t y of S z c z e c i n , Poland,  1980  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE  in THE FACULTY OF GRADUATE STUDIES Department of Mechanical  Engineering  We accept t h i s t h e s i s as conforming to the r e q u i r e d  standard  THE UNIVERSITY OF BRITISH COLUMBIA November, 1986  © ROBERT DYMALA-DOLE SKY  32  In  presenting  this  degree at the  thesis in  University of  partial  fulfilment  of  of  department  this or  thesis for by  his  or  requirements  for  an advanced  British Columbia, I agree that the Library shall make it  freely available for reference and study. I further copying  the  agree that permission for extensive  scholarly purposes may be granted her  representatives.  It  is  by the  understood  that  head of copying  my or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  McxW^cc^  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 i n 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 performance 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 i s shown  to be very important for the middle part of flame development.  The  simple squish design produces faster burning rate i n the f i r s t half of  - ii -  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 i s 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 i n the latter case. It appears that at lean operation mixture motion influences combustion process to a lesser degree than at stochiometric conditions.  - iii -  TABLE OF CONTENTS Page ABSTRACT TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES NOMENCLATURE ACKNOWLEDGEMENT 1.  2.  3.  4.  i i iv vi vii xi xiii  INTRODUCTION  1  1.1  1  General Discussion  1.2 Mixture Motion Effects i n I.C. Engines  3  1.3  7  Review of Previous Work on Turbulence in I.C. Engines ..  1.4 Objectives and Scope of this Work  11  EVALUATION OF THE SQUISH-JET DESIGN  13  2.1  Squish-Jet Design  13  2.2  Flow Visualization Experiment  15  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  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  5.  27  4.2.1  Test Conditions  27  4.2.2  Analytical Procedure  28  DISCUSSION AND EXPERIMENTAL RESULTS  31  5.1  31  Flow Experiments - iv -  CONTENTS (Continued) Page  5.2  5.3  6.  5.1.1  Results of Jet Velocity Measurements  31  5.1.2  Results of Turbulence Measurements  34  Firing Test Results  36  5.2.1  General Performance Parameters  37  5.2.2  Firing Pressure Analysis  37  Discussion of Experimental Technique  40  5.3.1  Flow Measuring Technique  40  5.3.2  Performance Evaluating Technique  45  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  Ensembled Peak Pressure and Standard Deviation of Cylic Peak Pressure for Different Piston Geometries, at 3000 and 2100 RPM  80  Ensembled Peak Pressures and Standard Deviation of Cyclic Peak Pressure for Different Piston Geometries, at 1200 RPM ..  81  Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries, at 3000 and 2100 RPM  82  Ensembled IMEP and Standard Deviation of Cyclic IMEP for Different Piston Geometries, at 1200 RPM  83  11.  12. 13. 14.  - 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  Comparison of Cyclic Jet Velocity Profiles for Pistons 5 and 6 at 3000 RPM  102  Comparison of Cyclic Jet Velocity Profiles for Pistons 1, 2, and 7 at 3000 RPM  103  5.2 5.3  - vii -  LIST OF FIGURES (Continued) Page 5.4 5.5 5.6 5.7 5.8 5.9  Comparison of Ensembled Jet Velocity Profiles for Pistons 1, 2 and 7 at 3000 RPM  104  Comparison of Ensembled Jet Velocity Profiles for Pistons 3, 4 and 5 at 3000 RPM  105  Comparison of Ensembled Jet Velocity Profiles for Pistons 5 and 6 at 2100 RPM  106  Comparison of Two Cycles of the Jet Velocity Profiles Measured i n Piston 5 at 300 RPM  107  Comparison of Cyclic Jet Velocity Profiles Measured in Piston 5 for Three Engine Speeds: 3000, 2100, 1200 RPM  108  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  Ill  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  - viii -  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  Area  a  Heat transfer coefficient  AjCe)  Instantaneous area of squish inflow,  CH  m  2  Surface of inflow to the bowl, Area of the leakage,  mm  2  mm  2  mm  2  Area of the contact of gas i n volume mm  with the cylinder head,  2  TOT BBDC  Total contact area of gas with the cylinder head, Before bottom dead center  BDC  Bottom dead center  BTDC  Before top dead center  BMEP  Brake mean effective pressure  BSFC  Brake specific fuel consumption  C  Piston velocity, m/s  CL  Clearance height, mm  CR  Compression ratio  D  Engine bore diameter,  d  Piston bowl diameter, mm  d n H h  Channel In the piston diameter, Depth of the piston bowl, mm  HWA  Hot wire anemometry  i  Particular cycle  IMEP  Indicated mean effective pressure, kPa  L  Length of the connecting rod, mm  MBT  Ignition - minimum advance for best torque  m  mass, kg  \  Leaking mass of the gas  P  Pressure, kPa  R  Crank radius, mm  RPM  Revolutions per minute  RAFR  Relative air fuel ratio  S  Distance of the piston from the top, mm  mm mm  Bowl height, mm  - xi -  mm  2  U(i,t)  Instantaneous velocity, i n cycle i , time t  U(i,t)  Mean velocity  V  Volume, m  ^TOT Vj  Total cylinder volume Volume above the squish area  V  Swept volume, m  3  3  g  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/m  v  Kinematic viscosity, m /s  x  Ratio of specific heats  3  2  - xii -  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 H i l l for their stimulating contributions i n various discussions.  - xiii -  1.  CHAPTER 1 INTRODUCTION  1.1 General Discussion At the present time internal combustion (I.C.) engines have undergone a history of improvements, almost a century long. Changes in economic and environmental conditions i n 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 i n I.C. engines.  The area of particular importance i s 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 i s particularly important for the viability of slow burning fuels, like natural gas.  Their potential  as alternate fuels i s being restricted by their considerably slower  2.  burning rates i n comparison to gasoline, which results i n a loss of power of engines converted from gasoline fuel to natural gas. It i s recognized that progress i n the direction of fast, lean burning 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. challenging of them i s mixture motion.  The most important and  This challenge i s 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 f i r s t class problem i n 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 d i f f i c u l t i s the lack of consistency with respect to experimental techniques used by different research groups.  Consequently i t i s often questionable to compare  absolute results published.  In addition there are some serious doubts  whether what i s frequently described as turbulence i s not a result of the experimental technique and method of data analysis used.  The question of  experimental procedure w i l l be dealt with i n detail, i n 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 i s 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 i n 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 f i r s t  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 f i r s t 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 i s 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 i t 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. positive effect was recorded.  In a C.I.  engine, however, a  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. squish combustion chamber.  Figure 1.2 shows a schematic of the  Experimental evidence on the role of squish  in combustion is not conclusive.  Some published results suggest that the  overall effect of squish motion i s 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 propagation.  A noticeable decrease of ignition delay in squish chambers i s  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 i s 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. problems arise from the unique nature of engine flows.  The  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 i s : t  1 U(i,t ) = i  / t  w  + T/2  w  U(i,t)dt w  - T/2  where T i s 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 f i t t i n g to data U ( i , t ) . w  The RMS velocity fluctuation, or turbulence intensity is represented by: t \l  w  + T/2  t -T/2  (u(i,t))2dt  w  The integral length scale is defined as: oo  L  x  = / R(r)dr o  where R(r) is a spatial autocorrelation  coefficient.  7.  The Taylor microscale i s : 11/2 (92R/9v2)  J 0  However, to evaluated these scales, assumptions of isotropy and relaxation of the turbulent fields are usually used.  The length scales are  then calculated from easier to measure time scales.  This procedure, i n  many engine applications, i s 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 i n 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 i n 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 i s the correlation between turbulence and the speed of the flame propagation. Measurements of the turbulent flow field in engines were f i r s t undertaken i n 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 experiments 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 i n the intake process, decay during compression, relaxation in i t s 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 f i r s t 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 i n conditions similar to that existing i n engines.  In their report Tindal  suggested a method for proper analytical evaluation of data from probes calibrated i n ambient conditions. More extensive projects were carried out during the "energy c r i s i s " 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 i n  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 i n 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. too  extensively.  None of them tried to explore the margin of error  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. scales.  His correlation did not incorporate any effect of turbulent 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 i s , however, very questionable. Up to the middle of the 1970's the only available measuring technique to probe the flow f i e l d i n 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 i n 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 thermocouples and thermometers [13,17,20].  Their claims of obtaining a proper  frequency response are s t i l l met with scepticism. An extensive evaluation 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 i n  many flow conditions. Measurements, however, in I.C. engines are s t i l l loaded with obstacles. A major difficulty encountered with LDV measurements In engines i s a necessity to redesign the engine to accommodate visual access. operation.  This imposes restrictions on parameters of the engine  Other reported d i f f i c u l t i e s 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 a l . [29].  Some very  interesting projects with LDV measurements i n engine like conditions are reported by Arcumanis, Bicen and Whitelaw [30,31,33].  They conducted  experiments i n 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 i s becoming has been  demonstrated in a recent paper by Fraser [33], which reports two point length scale measurements with LDV i n 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] i n their attempt did not  manage to detect any squish velocity. gible.  They claimed that i t was negli-  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 i n time and with cyclic variations. From the combustion stand point, however, squish effects were evaluated mostly i n 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 i s 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 i n i t s 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 i n the combustion chamber at about 40° BTDC. Turbulence generated by the breaking up of these jets would have excellent timing i n affecting the combustion process. The project was completed in three stages. A flow visualization experiment, conducted f i r s t , 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 i n 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 i t s 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 i s i n the incorporation of channels i n the piston of the bowl-in-piston type of combustion chamber.  The channels  were expected to create jets i n 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 substantial gains in engine performance.  The cross-section of the design i s  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 c r i t i c a l 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 i t 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 f i r s t 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 i n 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 i n 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 i n 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. dramatic.  Their dependence on compression ratio is less  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 i n 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 i n the model.  The cylinder liner of the model was made  of plexiglas, and the covering plate of Lexan polycarbonate.  The piston  was i n 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 conditions, 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 a l l the dimensions were scaled with the engine. Flow inside the model was seeded with microbaloons.  A l l 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 i t s 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 i n 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. photograph of the assembly. one of the films. horizontally.  Figure 2.3 displays a  Figures 2.4 and 2.5 present 6 frames from  Jets are seen developing from two channels located  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. electronic counter was located in view of the camera.  A display of an 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. then used.  Another simple method was  A stroboscope lamp was positioned in view of the camera and  triggered 4 times during the compression stroke. More frequent triggering 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, i n 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 i n a test c e l l  of the Alternative Fuels Laboratory of the Department of Mechanical Engineering at the University of British Columbia.  The c e l l i s 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 i n the engine while motored by a D.C. motor of the dynamometer, at WOT and three different speeds: 1200, 2100, 3000 RPM. operation. (CTA).  This speed range i s representative of a modern engine  The measuring technique was constant temperature anemometry  Hot wire probes were calibrated i n 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 a i r 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, i n 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 i n an unambiguous manner.  The standard gasoline Ricardo Hydra engine configuration has a "bath-tub" combustion chamber located i n 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,  20.  • manufacturing of a packing plate to be placed under the cylinder block. The new cylinder head was required to have a f l a t 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. longer cylinder liner.  This also required a new,  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. consideration.  Properties of the material were an important  Automotive aluminum pistons are usually forged or cast  and fast cooled in permanent molds.  This procedure decreases the  material grain size and enhances i t s 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. university.  This was contracted outside the  Machining of the pistons, however, with especially d i f f i c u l t  21.  oval skirts was done i n 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 i n 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 i n 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. HWA  It was decided to develop a system based on a  probe located i n the piston with signals transferred out along a  linkage mechanism. The i n i t i a l plan called for a custom made probe by a specialized manufacturer, TSI. sive.  This option turned out to be too expen-  An original probe was then designed and built from the same  materials TSI uses.  It i s 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 i s shown i n Figure 3.4.  The most c r i t 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  22.  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 i n the piston. The engine modified as described above was the heart of the experiments.  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 i n 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 i n a constant temperature mode. The signals were filtered by a DANTEC 56N20 signal conditioner, set i n a low pass f i l t e r 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 i n the  piston was built according to the same sensor specifications. each velocity signal was accompanied by a pressure history.  Initially, 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 c e l l 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 i s 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 i s 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 i s used. be fuelled with natural gas or gasoline.  The engine can  During the experiments reported  in this thesis i t was only operated on natural gas from city mains. gas specifications are given in Table 3.  The  24.  The Data Translation data acquisition system operates with a frequency of 27.5 kHz and collects ten basic signals from analog transducers. are:  The monitored signals, i n addition to those already mentioned  natural gas differential pressure on a Meriam model 50 MWV-1.5  laminar flow element, pressure and temperature of a i r 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 collected  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 4.1.1  Flow Field Analysis 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 measurements at the same time.  The i n i t i a l plan was to conduct a l l  The first experiment showed, however,  that the short probe was exposed to very dynamic flow conditions. caused frequent breakage of the probe's sensor. wire the engine had to be dismantled.  This  To repair the sensing  It seemed advantageous then to  conduct the work in two steps: f i r s t 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 configurations: seven piston geometries and one different sensor orientation. cases are shown i n Figure 4.1.  The  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. broken.  Every time however the sensor was  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. failure, the probe was easily removed from the engine.  After sensor  The tests were  conducted for the same speed range as with the short probe.  A l l 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; i n 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 i n 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 i n the cylinder at the point of valve closure.  27. In the analysis of the jet effects only mean velocities were important. averaged.  Therefore, for each case, cyclic velocities were ensemble To evaluate turbulence characteristics a cycle by cycle  nonstationary window averaging technique was used.  This method i s  regarded as the most sound way of extracting fluctuations from HWA velocity measurements in engines.  The technique was evaluated by  Cattania [43] and i s also described i n detail by Boisvert [42].  In each  of the analyzed cases mean velocities, and RMS fluctuations for indiviudal cycles were extracted from raw velocities. then ensemble averaged.  Cyclic values were  The RMS values can be regarded as representative  of turbulence intensity.  No attempt to evaluate scales was made here  because, i t 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.  A l l 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 i n Figure 4.3. A l l six were analyzed with the same testing strategy: F u l l Load Operation: 1200 RPM: RAFR RAFR 2100 RPM: RAFR RAFR 3000 RPM: RAFR RAFR  = 1.00 = 1.25 = 1.00 = 1.25 = 1.00 = 1.25  28. 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. experiments.  Interesting problems were noticed in the course of the 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.  value used for motoring and firing tests was the same; 1 mm.  The clearance 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. before saving was averaged over 100 cycles.  The data  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 compression 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 i t s own exponent.  The mass trapped i n the cylinder i s essentially constant.  During combustion the polytropic exponent changes from i t s compression value to i t s expansion value.  Pressure increases partly due to burning  and partly due to piston motion.  The latter can be calculated i f the  change i n the polytropic coefficient i s approximated by a linear function.  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. to zero.  The end is assumed when the derivative comes  The whole combustion duration i s divided into steps i n 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. easy to use.  This method has appeal because i t Is very simple and  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 f i r s t describes the results of the flow measurements.  The second evaluates  combustion tests and the third analyzes the measuring techniques used i n the project.  5.1 5.1.1  Flow Field Measurements 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. piston number 4 is an ordinary squish design without channels.  The  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 i s , 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 i n 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 i n 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. cycles.  The averaging in each case was carried out over 44  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 i s 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. ences are rather small.  The differ-  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 i n 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 i s 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 i n comparison to other cases and i t i s 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 i s number 7.  It offers advantage i n  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 measurements i n 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 f i e l d with combustion.  A comparison of ensembled  mean velocities for four different piston geometries at the engine speed of 3000 RPM  i s shown in Figure 5.12.  clearly visible before TDC. t i a l l y diminished.  The mean squish flow in piston 4 i s  The other pistons have this motion substan-  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. mean motion.  The trend shown by the fluctuations follows the  In piston 4 the squish flow increases also substantially  the level of fluctuations before TDC. An interesting aspect of these results i s 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 agreement. 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. overestimation of the velocity measured with the  This can lead to an 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 i n 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. high energy motion.  The bottom part i s , however, f i l l e d with unusually 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  2.  TDC.  The reduction i n the squish motion i s proportional to the number of channels.  3.  An increased level of activity at the bottom of the piston bowl i s 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. WOT  The analysis of the results shows that at  there i s 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 i n the error analysis. load operation, 2.5 bar BMEP and 2000 RPM,  fuel consumption was  At part analyzed  for different RAFR. This i s an indication of the efficiency of the engine operation.  The results are shown in Table 10.  eable differences of the order of 5-10% rather favourable for the squish piston.  There are notic-  in the lean limit, but they are However, at these operating  conditions, errors associated with the evaluation of the engine performance are assessed to be the highest.  5.2.2  Firing Pressure Analysis The analysis of the combustion process i s 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. different elements of this evaluation.  Tables number 11-14 show  Peak pressures of the ensembled  pressure signals are displayed i n 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  38.  8 channels produce the lowest peaks, a l l i n 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, i t s value i s i n the  range of 4.5-5%. At RAFR=1.25 i t 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 i n 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 i s 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. result.  An evaluation of expanded pressure traces helps to explain this The piston without channels produces faster rate of combustion  at the i n i 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, i s within 1-1.7%. For RAFR=1.25 i t doubles to 2.5-3%. There i s 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 i n the tables.  It is somehow surprising that such small increas in CR.  substantially changes peak pressure and IMEP. insight at the importance of CR.  This gives, however, an  The performance of the engine in terms  of power etc., i s not much different for case 41 i n comparison to case 4. The tests were, however, carried out f i r s t with piston 41. The same piston, with altered bowl, was then used for test 4.  It i s 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 i s 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 i s i n the piston number 2. A similar trend is observed at different engine speeds and RAFR. An unexpected result i s shown i n 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 i s 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 i n 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. quite well.  The model represents trends i n burning  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 presented by the pressure curves.  The fastest burning i s 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 dispersion at RAFR=1.25 than at RAFR=1.00. It i s interesting to notice that the model based on so many assumptions 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 i s an assessment of the limitations of the employed measuring technique. the evaluation of the quality of the obtained data.  This is a basis for The following  sections are aimed at conducting such analysis.  5.3.1  Flow Measuring  Technique  The low precision of the flow f i e l d measurements i n engines i s associated with the limitations of hot wire anemometry. Many researchers in the past tried to estimate error limits related to the technique. however, d i f f i c u l t to assign an exact number to i t .  It i s ,  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 i s recognized that errors i n the data  obtained with HWA are caused by three factors: lack of directional sensit i v i t y of the probe, assumptions used i n the analytical scaling of the results to the conditions of probe calibration and by a limited information about mean fluid temperature. the  The assessment of the error caused by  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 i n velocity.  This type of error i s ,  however, not the most important in comparative studies presented in this thesis.  Rarely a more physical nature of the problem i s 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 combustion chamber, both temperature and pressure are varying dramatically.  At  the final stage of the compression, the temperature of the gas i s 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 i n the mixture are i n fact carrying not only velocity but also temperature gradients. effects.  The HWA,  however, cannot differentiate these  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 i n 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 i s 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 i n 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 i s LDV. It i s recognized that the results of the measurements presented i n 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-  b i l i t y was small because total wire resistance was only .25 Ohm in comparison to the sensor resistance of 15 Ohms. This was checked, however, by conducting one experiment with a shortcircuited probe located i n 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. A l l 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 i n turbulence measurements i n engines.  This i s 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 i s whether the probe's body vibration can induce oscillations of the prongs.  The materials used for prongs, in high temperature probes,  are usually extremely s t i f f 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 i s 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 i n 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 i t s 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%. evaluation.  Consequently 1% i s the error margin set for the pressure  45.  5.3.2  Performance Evaluating Technique The accuracy of the results obtained during the performance tests i s  d i f f i c u l t 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 i t s 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. however, very different. test was completed. was powered up.  Dynamic behaviour was,  The meter rarely returned to zero value after a  It showed consistently some reading every time i t  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. A l 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 f u l l load, prior to taking any measurements. It i s 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 i t s parameters. The work began with a flow visualization experiment which confirmed jet development i n 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 i t s 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 i n 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 1.  Conclusions 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.  ii)  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 turbulent fluctuation and has indeed positive influence on combustion. It does not affect, however, cyclic variation i n 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 i n the i n i t i a l stage of the flame development.  This conclusion i s  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 i s suggested that further work be conducted i n 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 i n 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 i n the cylinder head, which would close the squish area and force flow through the channels.  This feature may offer substantially stronger jet  patterns. It i s 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 C r i t i c a l 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 i n 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 i n 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.  52.  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 i n 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 i n 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 i n an I.C. Engine Cylinder", SAE 750005, 1975.  23.  HAGHG00IE, M., KENT, J.C. and TABACZYNSKI, R.J., "Turbulence TimeScale Measurement i n 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 i n 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 i n 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 i n 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 i n 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. S o c , 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 i n 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 i n 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)  a)  l-D squish analysis i n ideal conditions  b)  The effect of gas leakage on squish velocity.  c)  The effect of heat transfer on squish velocity.  d)  Jet velocity evaluation.  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 i t s TDC position:  S(p) = R(l+cosp) + L(l-^l-e/sin/p  (1)  V (P)  (2)  Volume Vg(p):  9  Total volume V  T Q T  =  - "4  -*/S(p)+CL  L  (p): v  T 0 T  (6)  =  .  [ s ( e )  +  C L ]  +  ( 3 )  Surface A (6): 1  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  ^  dV„re)  TOT /a\ = P(9) • dt  TOT d6  1  u7W> +  dp(6)  p(e) *  6  dp(9) d6 d6 * dt -  dv^re)  de  v  T Q T  (e)  ( 5 )  (6)  TOT  =  o  de  From continuity for volume V^:  dm^ dF  dv^e) =  co  [P(6)  = -A^O)  dp(e)de  + V.(9)  de  • p(6) • U(6)  dv  VjCe)  dv^e)  to U(6) = - AjCe) L ae  (7)  m  re)  TOT  i  de  J  v  v^Ce)  *  (8)  TOT  Substituting;  (1-Sq )«d«h 2  U(6) =  -  A[S(e)+CL] • [S(6)+CL+h.Sq ] 2  ds(8) de  (9)  57. Continuity equation for the piston chamber: dm. CH dt  dV =  0)  CH  CH p(9)-A v  v(6)  =  d  dp(6) _ de  ,ai  CH  CH  dv^re) TOT  CH A  CH* T0T V  di)  5e  (e)  Substituting: v(9)  =  Sq + CL + Sq 2  b)  ds(e) de  2  (12)  The Effects of Leakage on Squish Velocity  Assumptions: • uniform flow past the rings can be modelled as a flow through a hole at c r i t i c a l conditions. • flow through the hole is adiabatic. • properties of the gas i n the chamber are uniform and change polytropically.  No effect of mass change on polytropic  coefficient.  Decrement of squish velocity because of leakage i s : dm < > " L< >  u  6  u  e  = A l  <8)  - p(e)  • dt-  (13)  58.  Leaking rate, using c r i t i c a l flow conditions: (  drn^  x  +  1  ^2(x-iy  )  dt  • /x«p «p c c  r * i 2(x-l) + 1  L  L dt  d m  - Vsr>  V„,™ r TOT -i  J  /x»p«p o 0  Lv  r  (6)-  x+1 2  1  TOT  Substituting into equation 13:  .2 2 (D'-D') • /x.R.T U  < >-V > = e  e  d  r i 2(x-l) x  l  +  1  J  rR+L+CL+h'Sq -j  . [s(eHCLj° '  2  CL+h.Sq^SO)  (14) where c)  V^-V^,  is the leaking diameter of the rings.  The Effects of Heat Transfer on Squish Velocity  Assumptions: • gas in the cylinder changes polytropically. • wall temperature i s uniform and constant during compression. • heat transfer rate through the wall is uniform. • cooling of the gas i s by the chamber walls. The equation of energy for the gas i n volume Vj^ i s :  59.  Substituting for dn^/dt: dm -j-i= - T T • d • [S(9)+CL] • ^ R.T(6)  U (6) u  d t  H  We can obtain from 15:  V  6)  =  x-1 ~ p(e).ir.d^x.[S(6)+CL]  d(c T ) ™1 d t A  A  V  [  _  P  +  l dQ dF TOT +  P  i dt~^  ( 1 6 )  P  For uniform cooling the above equation w i l l be: d(C U ( 9 )  " " p(9). T.d"x.[S(9)+CL] t 1  vV dT~  V, dV, M  +  V ^ "^" T  T  f P  dT ] -  Subtracting (17) and (16):  K " V> e)  e  - p(8).,.d.;. (6)+cLi [S  [(r\-Sr) • f1  < 1 8 >  TOT  P  The rate of heat loss can be calculated:  dt  = A • a(T -T ) p ™ q V TOT r  m  (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 a = 265 • D  , 0.786 -.525 (CP) • T  (20)  where a: kcol/m hr°C, D: m, C: m/s, P: kg/cm , T: °K. 2  d)  2  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 i n volume V  and i n 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.  1 J(8) • j^Qj A  +  U (9) = U(0)  (21)  8  From assumption of steady flow through parallel pipes:  H sLoss  = H. i Loss  T  T  Calculating pressure loss from Darcy-Weisbach formulae for: • turbulent flow conditions through the pipes • one pipe representing channel i n the piston • no additional losses except friction on the walls  f  !jL . ! i i = f , j , J s 'D '2g i * d ' 2g L  R  4  s = r _ii . -ii — f  U  .V —2.1  L  r  J  L  U (6)  W  h  e  r  e  f  L  s  f.  J(8)/  2  L  f  d.  s  K  j  L. L  1/2  J  D (8) ,  / 2  d.j  s  ( 2 2 )  '  (23)  2Trd[S(e)+CL] °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)  _ J L _ = J  ^  f  L  D  D , Hn 1 / 2  f _ l . _1 .JL + 5 L f  s  L  s  d  j  f  s  L  s  where K is the coefficient of pressure loss.  J  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 i s : • dynamic effect of the piston motion.  APPENDIX B PROGRAM HB (input,output);  {$1 FRAME.DEF ) {$1 FRAME.10 ) PROCEDURE Writelntroduction; {-) BEGIN Cirscr; writelnC writelnC  MASS BURN RATE ANALYSIS'); ====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.'); writelnC); writeCPress 'return' to continue....'); read(kbd,ch); END; PROCEDURE Wr iteMenuAndSelectOpt ion;  {  ) BEGIN cirscr; writein; writelnCThe following options are availab e: '); writein; writein; writelnC Load Pressure Data writelnC Scale Pressure Data writelnC Siooth Pressure Data writelnC List Pressure Data writelnC Align Pressure Data writelnC 6: Calculate 6aaia Values writelnC 7: List 6aua Values writelnC 8: Save Sana Values writelnC 9: Differentiate Gaiias writelnC 10: MassBurnRate writelnC 11: Exit writein; writein; writeCSelect one of the options C l . . 111: '); REPEAT read(option); writeC '); UNTIL option IN E1..111; END;  FUNCTION voKca: integer): real;  .  {  writein;  )  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) SweptVol:=pit(cyljad)t(cyl_rad)tDistFroiTDC; vol:=SweptVol+drv; END;  f  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'.'N ); 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; 1  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 ax [i]:=(ax [i+360]/204.8-10)tscale; 1440: FOR i:=l TO 360 DO ax [i]:=(ax [i*2+7201/204.8-10)tscale; 3600: FOR i:=l TO 360 DO ax ti]:=(ax C1799+i«]/204.8-10)tscale; END; writeCEnter BDC tanifold pressure in 8ar:');readln(BDCianPres); diff:=BDCianPres-ax [l]; FOR i:=l TO 360 DO ax Ci]:=ax [il+diff; NoOfDataPoints:=360; END; A  A  A  A  A  A  A  A  A  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]:=ax tl]; FOR i:=41 TO 400 DO yvalti3:=ax*Ci-401; A  FOR i: =401 TO 440 DO yval[il:=ax [360J; A  DesignFilter(Nweights,Fl,1,weights); data_stooth_weights(Nweights, weights); FOR i:=l TO 360 DO BEGIN ax [i]:=xval[i+40]; ay [i]:=yval[i+40J; END; A  {  A  writelnC CALCULATIONS FINISHED '); delay(1500); SsoothPressureData  )  END; PROCEDURE ListPressureData; {  }  BEGIN clrscr; writelnC FOR i:=l TO 360 DO BEGIN  LIST PRESSURE DATA ');  urit^':((i+360):4,' ',ay [i]:12:3,' IF (i HOD 20)=0 THEN BEGIN A  writeln;  ',ax Ci]:12:3); A  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 '); writelnCshift data to right(+) or left(-) by 'n" writeCenter n : '); readln(n); writeln; IF n'M THEN BEGIN FOR i:=360 DOMNTO-1 DO ax Ci+n]:=ax Cil; A  FOR i:=l END;  writeln; ');  A  TO n DO ax Cil:=ax Cn+l]; A  A  IF n<l THEN BEGIN FOR i:=l TO 360-n DO ax [il:=ax [i+n]; FOR i:=360-n+l TO 360 DO ax Cil:=ax [360-nl; END; writelnC CALCULATING...'); delay(lOOO); END; A  A  A  A  PROCEDURE CalculateGauas;  <--  )  VAR Numerator,Denominator,Gana: real; BEGIN clrscr; writelnC CALCULATE GAMMA VALUES '); writelnC CALCULATING... '); FOR i:=l TO 359 DO BEGIN IF ax m<>0 THEN BEGIN  writeln;  A  Numerator:=(ax Ci+ll/ax Cil)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; A  A  END ELSE 6aaaa:=0; IF Gaaaa<0 THEN az Ci+l]:=0 ELSE IF 6aaaa>3 THEN az Ci+ll:=3 ELSE az [i+l]:=6aaaa; END; writelnC CALCULATIONS FINISHED. '); delay(1500); END; A  A  A  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))); : noncoeftype = ((35,21,231,429,143), (10,28,60,110,182), (7,42,462,429,1001));  noncoef  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 ay ti]:=0; FOR i:=startconv TO stopconv DO BEGIN FOR j:=0 TO nuicoef-1 DO ay [i]:=ay [i]+az [i-sioothdeg+j]*s«oothcoef[derivnut,si>oothnui, j+11; ay f.i]:=ay [i]/noricoef [derivnui,sioothnui]; END; j:=sioothdeg; k:=nuidat-s»oothdeg+l; FOR i:=l TO sioothdeg DO BEGIN ay [j]:=ay [j+l]; ay [k]:=ay Ck-l]; j:=j-l; k:=k+l; END; A  A  A  A  A  A  A  A  A  A  END;  PROCEDURE ListDiffGanas; { } BEGIN cirscr; writelnC LIST DIFFERENTIATED GAMMA VALUES •); writein; FOR i:=l TO 360 DO BEGIN writeln((i+360):4,' »,az [ih 12:3, ',ay [i]:12:3); IF (i MOD 20)=0 THEN BEGIN writeCPress "return"..*); read(; IF ord(ch)=27 THEN exit; cirscr; writelnC LIST DIFFERENTIATED GAMMA VALUES '); writein; END; END; END; A  r  A  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: '); StartCoab:=180-StartCotb; END;  readln(StartCoib);  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]-a2 [startco«bl)/int(duration); FOR i:=startcoab TO endcotb DO BE6IN Co«b6aiiaf.i]:=int(i)tslope+az [startco«b]; writeln(i:3,' ' Co«bGana[i]:10:4); END; writei'Press return to continue..'); readln(ch); END; A  A  (  PROCEDURE CalcHassBurned; { ) VAR Pl,P2,Vl V2,SuiDeltaP,DeltaP,FinalPress: real; BEGIN SutDeltaP:=0.0; FOR i:=startcomb TO endcoib DO BE6IN Vl:=vol(i); V2:=vol(i+l); Pl:=ax [il; P2:=ax ti+ll; DeltaP:=P2 - PUexp(CotbGaMaCiltln(Vl/V2)); f  A  A  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(; 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 ay [ih=0.0; FOR i:=startcoib TO endcotb DO ay [i3:=HassBurned[i]; FOR i:=endcoib+l TO 360 DO ay [ih=1.0; SaveBinaryData('y'); delay(2000); END; A  A  A  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:  ');  Save6anas;  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 Ethane Propane Butane Nitrogen Carbon Dioxide Lower Heating Value  94.00 3.30 1.00 0.30 1.00 0.30 48,558 kJ/kg  :  P a r i s o n of the Engine Performance f o r D i f f e r e n t P i s t o n Geometries at WOT , 3000 RPM , RAFR =1.01 , MBT  C o m  Enrjl ne_sneed_=_3000_RPM RAFR = 1.01  BSFC Eff. (g/kWhr ) (%)  Torque Ign.Ad\ A l r f l . (deg) (Nm) (g/s)  N.G.fl (g/s)  RAFR  Piston No.  Speed (r/s)  Power (IcW)  BMEP (bar)  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 o f the Engine Performance f o r D i f f e r e n t P i s t o n Geometries, at WOT , 3000 RPM , RAFR = 1.26 , MBT  Engine_speed_=_3000_RPM RAFR  =1.26  BSFC Eff. (g/kWhr ) %  Torque Ign.Ad\ A i r f l . N . G . f l (Nm) (deg.) (g/s) (g/s)  AFR  Piston No.  Speed (r/s)  Power (kW)  BMEP (bar)  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 o f the Engine Performance f o r D i f f e r e n t P i s t o n Geometries, at WOT , 2100 RPM , RAFR = 1.02 , MBT  Engine speed = 2100 RPM RAFR = 1.02  Piston No.  Speed (r/s)  Power (kw)  BMEP (bar)  Eff BSFC g/kWhr! (%)  Torque Ign .Adv A1r f l . N . G . f l . (deg.) (g/s) (Nm) (g/s)  AFR  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 o f the Engine Performance f o r D i f f e r e n t P i s t o n Geometries , at WOT , 2100 RPM , RAFR = 1.27 , MBT  Engine speed : 2100 RPM RAFR =  1.27  Speed (r/s)  Power (MO  BMEP (bar)  BSFC Eff. (g/kWhr ) (%)  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  Piston No  Torque Ign.Adv A1r f 1 . N . G . f l (deg. (Nm) (g/s) (g/s)  AFR  Table_8 :  Comparison o f the Engine Performance f o r D i f f e r e n t P i s t o n Ceometries , at WOT , 1200 RPM , RAFR = 1.03 , MBT  Engine speed = 1200 RPM RAFR = 1 .03  BSFC  Speed (r/s)  Power  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  (kw)  BMEP (bar)  Eff.  (rj/kwhr ) U )  Torque (Nm)  Ign.Adv A1r.fl (deg) (g/s)  Piston No.  N.G.fl .  AFR  (g/s)  Table 9 :  Comparison of the Engine Performance f o r D i f f e r e n t P i s t o n Ceometries , at WOT , 1200 RPM , RAFR = 1.28 , MBT  Engine speed_=_1200_RPM RAFR = 1.28  Piston No.  Speed (r/s)  Power (kW)  BMEP (bar)  BSFC Eff. (g/kWhr ) (%)  Torque (Nm)  Ign .Ad' A i r . f l . N . G . f l (deg) (g/s) (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 Part load  X^AFR  BMEP = 2.5  bar  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  P i s t o n noSv  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 Std.Dev. Peak Press o f c y c l . (kPa) Peak(kPa)  %  Ensenbled Std.Dev. Peak P r e s s , o f c y c l . (kPa) Peak(kPa)  RAFR = 1.00  (V /c  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 Std.Dev. Peak P r e s s , o f c y c l . (kPa) Peak(kPa)  %  RAFR = 1.00  Ensembied Std.Dev. Peak P r e s s . of c y c l . (kPa) Peak(kPa)  %  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  Piston Mo.  1200 RPM , WOT , MBT  Ensembied Std.Dev. Peak P r e s s . o f c y c l . (kPa) Peak(kPa  %  RAFR = 1.00  Ensembied Std.Dev. Peak P r e s s . o f c y c l . (kPa) Peak(kPa) 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 Piston No.  Ensembied IMEP (kPa)  Std.Dev.  o f cycl.  %  IMEP(kPa)  Ensembled IMEP (kPa)  RAFR = 1.00  Std.Dev.  of cycl .  %  IMEPCkPa!  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 Piston Mo.  Ensembied IMEP (kPa)  Std.Dev  o f cycl.  %  IMEP(kPa;  RAFR = 1.00  Ensembi ed IMEP (kPa)  Std.Dev.  of cycl.  0/ ,}  IMEP(kPa)  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)  %  RAFR = 1.00  Ensembied IMEP (kPa)  Std.Dev. of cycl . IMEP(kPa)  %  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  F i g . 1.2  O b s t r u c t i o n s on the v a l v e s G e n e r a t i n g S w i r l a. Shrouded V a l v e b. V o r t e x V a l v e  Squish Combustion Chamber  Fig. 2.1  Squish-Jet  Design  Original steel c y l i n d e r U n e r and p i s t o n o f R.C.M.  Electronic counter High Speed camera  Fig.  2.2  Flow V i s u a l i z a t i o n  Experimental  Ret-up  co cn  F i g . 2.3  Photograph of the P l e x i g l a s Model  F i g . 2.4  Photographs of the J e t Development  F i g . 2.5  Photographs of the Jet Development  90.  F i g . 3.1  Cross S e c t i o n of the Redesigned R i c a r d o Engine  F i g . 3.2  Photograph of the Cast and Machined New Cylinder Head  92.  F i g . 3.3  Photograph of the Ricardo Engine and New Pistons  93.  Fig. 3 . 4  Photograph of the New Hot Wire Probe  94.  Fig.  3.6  Photograph of the Connecting Rod with the Linkage Mechanism and the Inside View of the Piston  F i g . 3.7  Photograph of the Probe P o s i t i o n in the Piston and the Connection between the connecting rod and the Piston  Exhaust  Press .Trans Amp. HI.' Anemometer & Filter  14: D.A.S.  DATA TRANSLATION  I sr! PC  VAX  3.8  11/750  D.A.S. ISAAC 2000  1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 15. 17. 18.  Temp. of exhaust P r e s s , o f exhaust Temp, o f i n t a k e Press, of intake Intake a i r f l o w r a t e Temp, o f n a t . gas Press . o f n a t . gas Mat. gas f l o w r a t e Variable ign.timing T r i g g e r . 2 c . a . & BDC Speed Torque H.W.A.signal H.W.A.signal Pressure AVL crank angle meter Diff. press, trans. Diff. press, trans.  Schematic of the Data A c q u i s i t i o n System  Piston n o . l  4 channels Piston  Piston  05/32  no.3  Piston  no.5  8 channels 0 Piston  F i g . 4.1  5/32  no.4  4 channels 0  5/32  P i s t o n no. 6  5/32  no.7  4 channels 0  4 channels 0 Piston  4 channels t> 5/32  no.2  8 channels p Piston  5/32  5/32  no.8  4 channels 0  12/64  Schematic of the Geometries f o r the Flow Experiments  99.  Fig.  4.2  Hot Wire Probe P o s i t i o n s Across  the P i s t o n  Bovl  100.  Piston  no.2  8 channels CR = 9.0:1  Piston  no.5  8 channels CR = 9 . 0 : 1  Piston  CR =  Fig.  4.3  no.4  9.0:1  Schematic Tests  Piston  no.3  4 channels CR = 9.0:1  Piston  no.7  8 channels CR = 9.0:1  Piston  CR =  of the Geometries  no.41  9.1:1  f o r the Combustion  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  200  2 20  24 0  260  280  300  CRANK ANGLES, 180=TDC PISTON 6 , 3000 RPM F i g . 5.2  Comparison o f C y c l i c V e l o c i t y and 6 a t 3000 RPM  Profiles  for Piston 5  320  103.  *5l 40  F i g . 5.3  H  Comparison of 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 P i s t o n s 1 , 2 and 7 , at 3000 RPM  for  104.  Fig. 5.4  Comparison of Ensembled Jet velocity Profiles for Pistons 1 , 2 and 7 , at 3000  RPM  105.  30  Fig.  5.5  Comparison of Ensembled J e t V e l o c i t y P r o f i l e s f o r P i s t o n s 3 , 4 and 5 , a t 3000 RPM  Legend PISTON 5 PISTON 6  - i  60  1  1  1  i  1  1  1  1  1  1  r  i  i  80  100  120  140  160  180  200  220  240  260  280  300  320  CRANK ANGLES, 180=TDC F i g . 5.6  Comparison o f Ensembled J e t V e l o c i t y P r o f i l e s  f o r P i s t o n s 5 and 6 a t 2100 RPM  45-1  OH  40 F i g . 5.7  1  1  1  1  1  I  60  80  100  120  140  160  Comparison of Two Cycles  1— •• —i  180  200  1  1  1  220  240  260  CRANK ANGLES 180=TDC  r — T  280  '  300  320  of the J e t V e l o c i t y Measured i n P i s t o n 5 a t 3000 RPM o  108.  S3 A  g  2o-  > 15-  CRANK ANGLES. 180=TDC  1200 F i g . 5.8  120  RPM  Comparison 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 Measured i n P i s t o n 5 a t Speeds: 3000 , 2]00 and 1200 RPM  45-1  F i g . 5.9  Comparison of C y c l i c J e t V e l o c i t y 1200 RPM  Profiles  f o r P i s t o n 5 at Speeds  3000 , 2100 and o  25-1  Oi  40 Fig.  5.]0  1  i  i  i  1  60  80  100  120  140  i  160  1  i  1  1  i  i  1  180  200  220  240  260  280  300  CRANK ANGLES, 180=TDC  Comparison of Ensembled J e t V e l o c i t y P r o f i l e s Measured i n P i s t o n 5 at Speeds 3000 , 2]00 and ]200 RPM  320  oH 40  1  1  60  80  1  1  1  1  1  1  1  1  1  100  120  140  160  180  200  220  240  260  1  1  280  300  CRANK A N G L E S , 180=TDC Fig.  5.11  Comparison of C y c l i c and 8 a t 3000 RPM  Jet  Velocity  Profiles  Measured  in  Piston  5  1 320  5-  0  I  40  Fig.  i  60 5.12  1  1  1  1  I  80  100  120  140  160  —I 18 0 2 0 0  I  1  C R A N K ANGLES, 180=TDC  1  2 20  1  —I  !  240  Comparison of Ensembled Mean V e l o c i t y P r o f i l e s f o r Four P i s t o n 3000 RPM , Top Probe P o s i t i o n  26 0  1  280  Geometries a t  3 0 0 320  12-1  u 1  40  1  \  1  1  1  1  1  60  80  100  120  140  160  18 0  1  200  1  2 20  1  240  1  260  1  I  280  CRANK ANGLES, 180=TDC F i g . 5.13  Comparison 3000 RPM  of Ensembled  Turbulent F l u c t u a t i o n s  , Top Probe P o s i t i o n  f o r Four P i s t o n Geometries at  [  300  320  CRANK ANGLES, 180=TDC F i g . 5.14  Comparison o f Ensembled Mean V e l o c i t y P r o f i l e s at 3000 , 2100 and 1200 RPM  for Piston  1 , Top Probe P o s i t i o n ,  35  5-  0-| 40  1  1  1  60  80  100  r  120  1  1  i  1  1  1  140  160  180  200  220  240  1 260 1  i  280  1  300  CRANK ANGLES, 180=TDC F i g . 5.15  Comparison of Ensembled Mean V e l o c i t y P r o f i l e s f o r P i s t o n at 3000 and 2100 RPM  A , Top Probe P o s i t i o n ,  320  12  Fig.  5.16  Comparison of Ensembled T u r b u l en t F l u c t u s t i o n s at 3000 and 2100 RPM  f o r P i s ton 4 , Top Probe Pos i t i o n >  35-i  F i g . 5.17  Comparison of Ensembled Mean V e l o c i t y f o r P i s t o n 4 Measured A c r o s s the P i s t o n at 3000 RPM  Bowl  12-1  140  160  180  200  320  220  CRANK ANGLES, 180=TDC F i g . 5.18  Comparison of Ensembled Turbulent F l u c t u a t i o n s Bowl at 3000 RPM  f o r P i s t o n 4 Measured Across  the P i s t o n  co  35-1  u  n  40  !  60  (  80  !  !  j  ,  ,  100  120  140  160  18 0  ,  ,  200  j  2 20  j  24 0  (  260  (  280  CRANK A N G L E S , 180=TDC Fig.  5.19  C o m p a r i s o n o f e n s e m b l e d mean v e l o c i t i e s measured a c r o s s the p i s t o n bowl  for piston  1 a t 3000  RPM,  f  300  320  12  Fig.  5.20  Comparison of Ensembled Turbulent F l u c t u a t i o n s Bowl at 3000 RPM  for Piston  1 Measured Across  the P i s t o n  ro o  F i g . 5.21  Comparison of C y c l i c V e l o c i t y at 3000 RPM  P r o f i l e s for Piston  1 and 4 a t the Bottom Probe  Position  12  140  160  180  200  220  CRANK ANGLES, 180=TDC Fig.  5.22  Comparison of ensembled f l u c t u a t i o n s measured a t the bottom of the p i s t o n  for piston bowl  240  1 and  260  4 at  280  300  3000  RPM,  320  ro ro  35  u  n  40  1  1  i  1  1  1  1  1  60  80  100  120  140  160  18 0  200  1  2 20  1  24 0  1  260  1  280  1  i  300  320  CRANK ANGLES, 180=TDC F i g . 5.23  Comparison of the Ensembled V e l o c i t y P r o f i l e s at 3000 RPM  for Pistons  1 and 4 Middle Probe P o s i t i o n  124.  5000  0-J 120  1  1  1  140  160  180  — T  —  200  1  220  |  240  f  260  CRANK ANGLES  F i g . 5.24  Comparison o f the Ensembled P r e s s u r e Traces f o r F i v e P i s t o n Geometries a t WOT , 3000 RPM , RAFR = 1.00 , MBT  125.  5000  0-| 120  , 140  , 160  -, 180  T= 200  , 220  , 240  1 260  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  126.  5000  OH  120  F i g . 5.26  1  1  140  160  1—  r-  1  180 200 220 CRANK ANGLES  1—  240  Comparison o f the Ensembled P r e s s u r e Traces f o r P i s t o n s 2 , 5 , 7 at WOT , 3000 RPM , RAFR = 1.00 , MBT  T  260  127.  5000-r  Legend  450CH  PISTON 2 PISTON 3 PISTON 4  4000H  3500-1  ^  3000-1  LxJ  o: ZD Ul Ul  2500-|  UJ  cn QL  2000-1  1500-1  1000  500  o-r-  120  140  160  180  200  220  240  CRANK ANGLES  F i g . 5.27  Comparison o f the Ensembled P r e s s u r e Traces f o r F i v e P i s t o n s a t 3000 RPM , RAFR = 1.25 , WOT , MBT  260  128.  5000  Legend  4500  PISTON 4 PISTON 2 PISTON 5  4000  PISTON 5 PISTON 7  3500H  3000  p LxJ  cn ZD  2500  CO  oo UJ  Q_  2000  1500  1000H  500H  120  140  160  T 180  200  220  240  CRANK ANGLES F i g . 5.28  Comparison o f the Ensembled P r e s s u r e Traces f o r F i v e P i s t o n s a t 2100 RPM , WOT , RAFR = 1.00 , MBT  260  129.  5000  4500  120  140  160  180  200  220  240  CRANK ANGLES  Fig.  5.29  Comparison of the Ensembled P r e s s u r e TRaces f o r F i v e P i s t o n Geometries a t 2100 RPM , WOT , RAFR =1.25 , MBT  260  130.  5000  Legend  4500  PISTON PISTON PISTON PISTON PISTON  4000  2 3 4 5 7  3500H  O CH ZD 00 00 Cr:  •_  3000H  2500  2000  1500H  1000  500  120  F i g . 5.30  140  I  160  I 180  1  200  CRANK ANGLES  I  I  220  240  Comparison of the Ensembled P r e s s u r e Traces f o r F i v e Geometries a t 1200 RPM , WOT , RAFR = 1.00 , MBT  '  260  Piston  131.  5000  Legend  4500  PISTON 2 PISTON 3 PISTON 4 PISTON 5  4000  PISTON 7  3500H  O Q_ _¥  3000H  Ld  cn  2500H  00 00  cn Q_  2000H  1500H  ioooH  500H  0  _ |  120  _  !  ,  140  160  ,  j  180  CRANK  F i g . 5.31  !  200  !  220  240  f  260  ANGLES  Comparison o f the Ensembled P r e s s u r e TRaces f o r F i v e P i s t o n Geometries a t 1200 RPM , WOT , RAFR = 1.25 , MBT  5000n  5000  4500 H  4500  4000  4000  3500  3500  O CL  o 3000H D-  PISTON 2  3000  -X  UJ CZ ZD in  UJ  ZD (/> (/> Ul CZ CL  Legend  2500H  2500 H  to ui cz  2000  1500  Q-  H  2000  1500  1000  1000  500  500H  180  200  220  24 0  260  o-t  '20  140  160  CRANK ANGLES  3000 Fig.  5.32  RPM,  RAFR  180  200  220  240  260  CRANK ANGLES  =  1.00  3000  RPM,  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 3000 RPM, WOT, MBT, f o r RAFR = 1.00 a n d RAFR = 1.25  AFR  = 1.25  Geometries  at  GO  ro  5000-r  5000  4500H  4500  4000  3500H  p  3000 H  0+120  — i  140  i  160  I  180  1—  l  200  220  CRANK ANGLES  2100  Fig.  5.33  RPM,  RAFR  =  1.00  240  2fi0  120  140  160  2100  180 200 220 CRANK ANGLES  RPM,  Comparison of Ensembled P r e s s u r e Traces f o r F i v e P i s t o n 2100 RPM, WOT, MBT, f o r RAFR = 1.00 a n d RAFR = 1.25  RAFR  =  240  260  1.25  Geometries at OJ OJ  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  200  220  240  CRANK ANGLES  Fig.  5.35  Comparison of Mass F r a c t i o n Burned f o r F i v e P i s t o n Geometries a t 3000 RPM , WOT , MBT , RAFR = 1 . 0 0  260  Fig. 5.36  Comparison of Mass Fraction Burned for Five Piston Geometries at 3000 RPM , RAFR = 1.25 , WOT , M B T  Fig.  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 M e a s u r e d v i t h HWA  138.  A  l  -  A  CH"  s u r f a c e above of the p i s t o n s u r f a c e of the p i s t o n  perimeter bowl  the i n l e t bowl  to  B  center - d i s t a n c e between and t o p o f t h e p i s t o n  CL  -  c l e a r a n c e between of t h e p i s t o n and der head  (6D - b o r e 0d  d iame t e r  - piston  bowl  L  connecting  R  crank  S  s troke  S(6)U V  top cylin  diameter rod  radius  instantaneous  - squish  - jet  8  - crank  stroke  velocity  vertical  J  length  velocity  velocity angles  Fig.  5.38  from  BDC  Geometry of Evalua t ion  the Engine  for Analytical  139.  Fig.  5.39  S c h e m a t i c s of the geometry f o r the c a l c u l a t i o n : a) c o m b u s t i o n c h a m b e r m e t r y b) m o d e l f o r calculations  jet geo-  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items