Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An evaluation of partially stratified charge ignition in a direct injection natural gas engine Gorby, David 2007

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2007-0103a.pdf [ 8.45MB ]
Metadata
JSON: 831-1.0080782.json
JSON-LD: 831-1.0080782-ld.json
RDF/XML (Pretty): 831-1.0080782-rdf.xml
RDF/JSON: 831-1.0080782-rdf.json
Turtle: 831-1.0080782-turtle.txt
N-Triples: 831-1.0080782-rdf-ntriples.txt
Original Record: 831-1.0080782-source.json
Full Text
831-1.0080782-fulltext.txt
Citation
831-1.0080782.ris

Full Text

AN EVALUATION OF PARTIALLY STRATIFIED CHARGE IGNITION IN A DIRECT INJECTION NATURAL GAS ENGINE by DAVID GORBY B.A.Sc, the University of British Columbia, 2000 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES (Mechanical Engineering) THE UNIVERSITY OF BRITISH COLUMBIA April 2007 © David Gorby, 2007 ABSTRACT The challenge of reducing tailpipe emissions, while retaining performance, continues to motivate new engine technologies. To this end, natural gas shows promise as a clean burning alternative fuel. However, an efficiency penalty of conventional spark ignited engines persists when these are fuelled by natural gas. This penalty is due to pumping losses, which arise as the intake mixture is throttled to maintain stoichiometry over the engine's operating range. Direct injection (DI), used to create a globally lean stratified charge, provides a load control method which is decoupled from throttling. Previous work has shown this concept to be viable, but plagued by poor fuel usage and high emissions of unburned hydrocarbons. Partially Stratified Charge (PSC) technology is a novel high-energy ignition method. It involves a standard spark plug, modified such that a small portion of natural gas (<5% of the total fuel charge) is injected directly into the vicinity of the electrodes. The intention is to surround these with an easily ignitable mixture, thus ensuring a strong flame kernel that propagates through a marginal bulk charge. PSC has already been shown to improve combustion stability in ultra-lean, homogeneous mixtures. This study was conducted to determine if PSC can be combined with direct injection to improve combustion stability in a stratified charge engine. A single cylinder research engine was modified to include a natural gas direct injector, as well as PSC ignition. Experiments were conducted over a wide range of fuel injection timings, as well as different angles between the fuel jet and ignition source. Engine performance was poor at stratified charge operating conditions and results indicate that PSC did not reliably improve combustion. Experimental evidence suggests that the PSC system was not igniting the pilot charge as expected. Very high concentration gradients at the boundaries of the PSC and DI ii plumes may be the cause o f the negative result. These make it diff icul t to re l iably ensure that the spark p lug discharges into the narrow and transient regions o f combust ible mixture. A secondary invest igat ion was conducted i n w h i c h a very dilute p remixed fuel component was added to the intake air. Th i s assured reliable igni t ion o f the P S C pi lot charge; and in this configuration, stratified charge combust ion at late direct injection t imings produced stable engine operation. i i i TABLE OF CONTENTS Abstract it Table o f Contents . . . ; iv L i s t o f Tables v i L i s t o f Figures v i i Nomenclature i x Acknowledgements • • x i i 1. INTRODUCTION 1 1.1 B A C K G R O U N D 1 1.2 R E S E A R C H O B J E C T I V E S 6 2. EXPERIMENTAL APPARATUS 9 2.1 R E S E A R C H E N G I N E 9 2.2 D I R E C T INJECTION S Y S T E M 10 2.2.1 Direct Injector 10 2.2.2 Cylinder Head Modifications 11. 2.2.3 Jet Development. 12 2.2.4 Injector Characterization 13 2.3 P S C S Y S T E M 14 2.3.1 PSC Timing Characterization 15 2.3.2 PSC verification 2 0 2.4 INSTRUMENTATION A N D D A T A ACQUISITION 2 3 2 .5 E N G I N E C O N T R O L S Y S T E M 2 7 2.6 D A T A P R O C E S S I N G 2 8 2 .7 STATISTICAL PILOT S T U D Y 2 9 2.8 UNCERTAINTY A N A L Y S I S 31 2 .9 C L O S I N G R E M A R K S 3 3 3. METHODOLOGY 34 3.1 DEFINITIONS: E A R L Y A N D L A T E D I R E C T INJECTION 34 3.2 Dl TIMING S W E E P S 3 5 3.2.1 PSC Settings .36 iv 3.2.2 Air Fuel Ratio 36 3.2.3 Direct Injector Jet Angle 37 3.3 PILOT C H A R G E STUDY 37 4. R E S U L T S A N D D I S C U S S I O N 39 4.1 DI TIMING SWEEPS 39 4.1.1 Results 39 4.1.2 Discussion 41 4.2 PILOT C H A R G E STUDY 5 0 4.2.1 Results 50 4.2.2 Discussion 51 4.3 SUMMARY 57 5. C O N C L U S I O N S A N D R E C O M M E N D A T I O N S 59 5.1 CONCLUSIONS 5 9 5.2 RECOMMENDATIONS FOR FUTURE WORK 6 0 R E F E R E N C E S 63 A P P E N D I X A : D E S I G N D O C U M E N T A T I O N 66 A P P E N D I X B : S A M P L E C A L C U L A T I O N S 76 A P P E N D I X C : E N G I N E O P E R A T I N G P R O C E D U R E S 82 A P P E N D I X D : F U E L C O M P O S I T I O N 87 v LIST OF TABLES Table 1-1 Hydrogen to Carbon Ratios 1 Table 2-1 Ricardo Hydra SI Configuration Geometry 10 Table 2-2 PSC Specifications 15 Table 2-3 PSC Verification Operating Parameters 21 Table 2-4 Ricardo Hydra Instrumentation 24 Table 2-5 Statistical Pilot Study Operating Condition 29 Table 2-6 Statistical Pilot Study Sample Sizes 29 Table 2-7 Measurement Uncertainty 31 Table 3-1 Fuelling Contributions : 38 Table 4-1 Partial Combustion: Weak Mixture with PSC 55 Table 4-2 Average BSFC and Ignition Delay —-57 Table D-l BC Natural Gas Composition 87 Table D-2 Mixture Properties • 87 vi LIST OF FIGURES Figure 1.1 Es t imated F u e l Consumpt ion Penalty for Natura l Gas vs. D iese l 3 Figure 1.2 F u e l E f f i c i ency o f L o a d Con t ro l Strategies 8 Figure 2.1 SI C y l i n d e r H e a d M o d i f i e d for J43 Direc t Injector 11 Figure 2.2 D I Jet A n g l e Adjustment Range 12 Figure 2.3 J43 Pulse W i d t h Character izat ion 13 Figure 2.4 P S C Spark P l u g 14 Figure 2.5 P S C T i m i n g Character izat ion Process F l o w Diag ram 16 Figure 2.6 P S C T i m i n g Character izat ion - 8 bar Supply Pressure 17 Figure 2.7 P S C T i m i n g Character izat ion - 15 bar Supply Pressure 18 Figure 2.8 P S C T i m i n g Character izat ion - 20 bar Supply Pressure 18 Figure 2.9 P S C Character izat ion, 8 bar, Coarse Timescale 19 Figure 2.10 Character izat ion, 15 bar, Coarse Timescale 20 Figure 2.11 Character izat ion, 20 bar, Coarse Timescale 20 Figure 2.12 His togram o f Integrated Heat Release 22 Figure 2.13 Dynoc l i en t Interface W i n d o w 25 Figure 2.14 Pressureclient Interface W i n d o w 26 Figure 2.15 T i m i n g Cont ro l le r Interface W i n d o w 27 Figure 2.16 Torque Hal f - length o f CI95 at V a r y i n g Sample Sizes 30 Figure 2.17 Speed Hal f - length o f CI95 at V a r y i n g Sample Sizes 30 Figure 2.18 F u e l F l o w Hal f - length o f CI95 at V a r y i n g Sample Sizes 30 Figure 3.1 Hypothe t ica l D I F u e l Ef f i c i ency wi th T i m i n g 35 Figure 4.1 D I T i m i n g Sweep B S F C Resul t (0° Injector A n g l e ) 40 Figure 4.2 D I T i m i n g Sweep B S F C Results (8° Injector A n g l e ) 40 v i i Figure 4.3 Histogram of Net Heat Release (0° Injector Angle) 42 Figure 4.4 Histogram of Net Heat Release (8° Injector Angle) 43 Figure 4.5 Misfire with DI Timing (0° Injector Angle) 44 Figure 4.6 Misfire with DI Timing (0° Injector Angle) 44 Figure 4.7 BSFC with Misfire (0° Injector Angle) 45 Figure 4.8 BSFC with Misfire (8° Injector Angle) 46 Figure 4.9 tHC with Misfire (0° Injector Angle) 47 Figure 4.10 tHC with Misfire (8° Injector Angle) 47 Figure 4.11 BSFC with tHC (0° Injector Angle) 48 Figure 4.12 BSFC with tHC (8° Injector Angle)... 49 Figure 4.13 Pilot Charge Study Fuel Efficiency Results : 50 Figure 4.14 Histogram of Net Heat Release for all pilot charge study conditions 51 Figure 4.15 In-Cylinder Pressure Trace Comparison: Motored and with PSC 52 Figure 4.16 Net Heat Release Rate Comparison: Motored and with PSC 52 Figure 4.17 In-Cylinder Pressure Trace Comparison: With and Without PSC 54 Figure 4.18 Net Heat Release Rate Comparison: Weak Mixture With and Without PSC 54 Figure 4.19 BSFC Comparison: Pilot Charge Study With Timing Sweeps 56 Figure A. l Aluminium SI Cylinder Head Solid Model 66 Figure A.2 GEA Analysis of Stress Due to Combustion Seal Pre-Load 67 Figure A.3 DI Head Modification Fabrication Drawings r 74 Figure A.4 Geometry: Piston # 476P by Federal Mogul for Ford Fiesta (1978-80) 75 viii N O M E N C L A T U R E SYMBOLS A Re la t ive air fuel ratio Sample M e a n S Sample Standard Dev ia t i on n N u m b e r o f Sample Points w Uncer ta inty B F i x e d Uncer ta inty due to Instrument Er ror P Statistical Uncer ta inty (95% Confidence Interval) x Measu red Parameter m M a s s F l o w Rate P B r a k e P o w e r b xb B r a k e Torque N Eng ine Speed / F u e l ABBREVIATIONS A B D C Af te r B o t t o m D e a d Centre B M E P Brake M e a n Effect ive Pressure B S F C B r a k e Spec i f ic F u e l Consumpt ion B T D C Before T o p D e a d Centre C A D Crank A n g l e Degrees C I Conf idence Interval C L Conf idence L i m i t C O C a r b o n M o n o x i d e C 0 2 C a r b o n D i o x i d e C O V Coeff ic ient o f Var i a t i on D C Di rec t Current D I Di rec t Injection D I S C Di rec t Injection Stratified Charge D O E Department o f Energy E G R Exhaus t Gas Rec i rcu la t ion E x V l v . Exhaus t V a l v e G C Gas Chromatograph G D I Gaso l ine Di rec t Injection H / C H y d r o g e n to Ca rbon Rat io H R R Heat Release Rate I C Internal C o m b u s t i o n I H R Integrated Heat Release I M E P Indicated M e a n Effect ive Pressure Int V l v . Intake V a l v e L M L L e a n M i s f i r e L i m i t M B T M i n i m u m A d v a n c e for Best Torque N O x Ox ides o f N i t rogen P M Particulate Mat te r P S C Par t ia l ly Stratified Charge P W Pulse W i d t h (R+M)/2 Average o f Research and M o t o r e d Octane Number s R A F R Rela t ive A i r F u e l Ra t io S A E Society o f Au tomot ive Engineers rpm Revolu t ions per M i n u t e S C Stratified Charge SI Spark Igni t ion S O I Start o f Injection t H C To ta l Hydrocarbons U B C Unive r s i ty o f Br i t i sh C o l u m b i a U S U n i t e d States x i ACKNOWLEDGEMENTS I w o u l d l ike to thank m y supervisor, D r . Robert Evans , for p rov id ing the opportunity for me to come back to the Wes t Coast and study at U B C . I w o u l d also l ike to thank C o n o r Reynolds , whose mentorship and guidance were invaluable assets dur ing this project. I am grateful to my lab partner, D a v i d W i l l i a m s , and m y other lab-mates M a l c o l m Sh ie ld , E d Chan , James Saunders and A n d r e w M e z o for the good discussions and for keeping me laughing throughout. I am very grateful to Westport Innovations Inc, for p rov id ing the hardware and support that made this thesis possible. I w o u l d part icularly l ike to thank M i k e Hebbs for lending his experience and guidance towards getting this project started. I w o u l d l ike to thank the M e c h a n i c a l Engineer ing faculty and staff, par t icular ly: M a r t i n D a v y , B o b Parry, G l e n n J o l l y , M a r k u s Fengler and G o r d W r i g h t for their assistance, advice and explanations o f things I knew nothing about. F ina l l y , I w o u l d l ike to thank m y wife , E r i n , w h o went through this first, then encouraged me to fo l low. x i i 1. I N T R O D U C T I O N LI BACKGROUND The reciprocating internal combustion (IC) engine remains the dominant power-plant of the world's ground based transportation system. As it continuously evolves, the dynamic standard it sets in terms of cost effective performance and reliability has seen challengers relegated to the fringes of market viability. Given the ubiquity of IC engines, the twin challenges of increasing fuel efficiency and reducing tailpipe emissions remain relevant today, as they will into the future. To this end, natural gas has both economic and environmental benefits as an alternative fuel for IC engines. As a gaseous fuel, long chain hydrocarbons and high molecular weight compounds are absent, thus the atomic ratio of hydrogen to carbon (H/C) is favourable compared to liquid hydrocarbons fuels (Table 1-1). Table 1-1 Hydrogen to Carbon Ratios Fuel Molecular H/C Ratio Gasoline 1.6-2.1 Diesel -1.8 Natural Gas 3.7-4 Source: US DOE 2006a These attributes are beneficial for the combustion of natural gas: they contribute to minimizing the production of soot and other particulate matter (PM) (Heywood 1988); and the favourable H/C ratio results in an approximate 25% reduction in C 0 2 per unit of heat energy released, when compared to gasoline or diesel (BP 2005). Also inherent in natural gas engines is an enhanced potential for reducing the formation of oxides of nitrogen (NOx). NOx emissions are often mitigated through use of exhaust gas recirculation (EGR); however, at high EGR ratios the NOx 1 reduction benefits become offset by increasing level o f P M . The l o w propensity for natural gas to form P M makes it a good candidate for m a x i m i z i n g the N O x reducing potential o f E G R (Goudie , D u n n et a l , 2004). F o r the convent ional spark ignited engine, op t imized for natural gas rather than gasoline, there is a potential benefit to thermal eff iciency. T h i s is due to the potential for increased compress ion ratios, w h i c h are supported by the higher octane characteristic o f natural gas (gasoline: 86-94, natural gas: 120+, ( R + M ) / 2 , U S D O E 2006a). The economic attractiveness o f natural gas results from its current status as the least expensive o f a l l fuels, when evaluated on a cost per unit energy basis ( U S D O E 2006b). Despite the potential benefits, there are barriers to the wide acceptance o f natural gas by the transportation market. These are p r imar i ly due to the relat ively l o w volumetr ic energy density inherent i n gaseous fuels, w h i c h subsequently require more complex fue l l ing and on-board storage systems than are necessary for l i qu id fuels. The complex i ty o f these systems varies wi th the strategy used (normal ly compress ion or l iquefaction), but they can add signif icant ly to a vehic le ' s weight and expense. Progress is needed i n terms o f cost effective, l ight weight storage options. In addi t ion to the ini t ia l capital expenditure, the added weight inherent in current gaseous fuel storage decreases vehic le fuel economy, eroding the potential for reductions i n cost and emissions. Current ly , fleet vehicles are i n the best posi t ion for a transition to natural gas. The potential cost savings, leveraged over their h igh fuel use, results in a substantial margin for offsetting the increase in vehic le complex i ty . A d d i t i o n a l l y , fleet vehicles are often serviced by central refuell ing depots and are subsequently less affected by the l imi ted avai labi l i ty o f publ ic natural gas infrastructure. A m o n g fleet vehicles , natural gas is a part icularly attractive alternative in med ium duty diesel applicat ions; and indeed as o f 2000, natural gas had penetrated the U S transit bus market more effectively than any other vehic le market (F ra i ly , C l a r k et a l , 2000). The 2 already substantial size and expense o f these vehicles further reduces the marginal cost and mileage penalty o f adding natural gas storage. Addi t ional ly , the three-way catalyst found on gasoline engines is currently an effective pollutant countermeasure. T h e absence o f such after-treatment on diesel engines, combined with their high P M output, provides the greatest opportunity for m a x i m i z i n g the emission reducing potential o f natural gas. T h o u g h beneficial in terms o f air quality, there is an efficiency penalty when diesel is replaced by natural gas, using a conventional spark ignition (SI) cycle. Th i s penalty is greatest under part load, where conventional SI engines experience throttling losses in order to maintain a stoichiometric air-fuel ratio. T h e problem is further compounded by the stoichiometric mixture's inability to exploit the thermodynamic benefits o f excess air, such as higher specific heat ratios, and reduced heat transfer to the combustion chamber walls (Ferguson and Kirkpatr ick , 2001). Figure 1.1 shows qualitatively the fuel consumption penalty o f a typical natural gas engine compared to an equivalent diesel. Reduction or E l iminat ion o f this penalty would enhance market acceptance o f natural gas, further realizing its potential for reducing emissions and operating costs. Torque curves Speed Figure 1.1 Estimated Fue l Consumpt ion Penalty for Natural Gas vs. Diese l (Ricardo, F r o m K u b e s h 2002) 3 One approach to mitigating this fuel consumption penalty is through a homogenous lean burn strategy, in which engine output is controlled by varying the fuel flow, without throttling the airflow. Germane et al (1983) provides a complete review of the lean burn concept; while others (Gupta and Bell, 1994) have contributed results directly related to lean burn natural gas engines. As a means of load control, this approach is effective near the upper end of an engine's load range, however combustion quality eventually degrades as the relative air fuel ratio (RAFR, denoted: A ) approaches the lean misfire limit (LML 1 - for natural gas A L M L ~1.6, Gupta and Bell 1994). Assuming good combustion until very near L M L , this represents throttleless operation over the engine's upper -35% load range. In order to achieve throttleless operation over the entire load range, one potential strategy is to operate at global relative air fuel ratios well in excess of the homogeneous lean misfire limit. To achieve this, the fuel/air charge must be stratified such that an ignitable mixture contacts the ignition source at the correct time, though the balance of the working gas is composed of inert dilutent (EGR or Air). The combusting gasses may be separated physically or aerodynamically from the working gas, which may be air, recirculated exhaust gas or a combination thereof. Abata (1986) provides a detailed historical review of the stratified charge concept, while Kubesh (2002) provides more current examples of the concept applied in natural gas engines. Currently, there is renewed interest in the direct injection (DI) category of stratified charge (SC) engines. In this variation (acronym: DISC), fuel is injected directly into the combustion chamber, where it interacts with surrounding air to form a combustible mixture. Zhao, Harrington et al L M L is reached when COVIMEP>10% (Heywood, 1998). Though leaner combustion is possible, this threshold defines the point at which drivability. is said to be adversely affected. 4 (2002) review this concept in its most common application of gasoline direct injection (GDI). Early examples of DI engines were limited by their mechanically controlled fuelling systems. Freed of these limitations by modern advances in electromagnetic actuation and microprocessor based engine management, researchers are now motivated to re-examine the benefits of DISC engines. The allure of the stratified charge engine is its potential to hybridize spark ignited and compression ignition engine models, combining the advantages of each. However, applications of this concept remain beset by poor fuel usage and high emissions of unburned hydrocarbons, particularly at reduced load. Managing the interface between the combustible mixture and working gas remains problematic. Abata (1986) describes this region as highly dynamic, influenced by inlet aerodynamics, piston movement and combustion induced fluid motion. More specifically to DISC, Lahbabi et al (1993) describe the free surface of a transient gaseous jet as highly complex, with turbulent structures producing undulations on the order of the jet's half width. These factors can contribute to separation between the combustible mixture and the ignition source, resulting in misfire. In addition, local mixing beyond the lean combustion limit leads to flame extinction. Indeed many researchers continue to report difficulty in managing this highly complex phenomenon (Goto and Sato, 2001, Turner, Pearson and Kenchington, 2004, Huang et al 2003) Local charge stratification has emerged as a sub-set of the stratified charge concept. In this variant, a small portion of the overall fuel charge is inducted separately to create a localized pilot region with a distinct composition: an easily ignitable mixture in the vicinity of an ignition source. When applied as an enhanced ignition system, this has been shown to improve combustion stability in ultra-lean mixtures, which require greater ignition energy than a stoichiometric charge (Green and Zavier, 1992, Arcoumanis et al., 1997). In their review of 5 alternative igni t ion methods, D a l e et al (1997) describe enhanced ign i t ion systems as those w h i c h increase igni t ion energy and improve its dispersion throughout a combust ible mixture. In the case o f loca l charge stratification, the spark is augmented by the added fuel ' s chemica l energy, w h i c h is then dispersed by turbulent mot ion . 1.2 RESEARCH OBJECTIVES The objective o f the research presented in this thesis was to determine i f a technique o f local charge stratification, referred to as part ial ly stratified charge ( P S C ) , can improve combust ion stability and fuel usage in a natural gas powered D I S C engine. P S C is a method o f loca l charge stratification developed at The Un ive r s i t y o f B r i t i s h C o l u m b i a (Evans 2000). It involves injecting a small port ion o f fuel direct ly into the v i c in i t y o f a modif ied spark p lug ' s electrodes. Reyno lds (2001) demonstrated an extension to the lean misfire l imi t in a homogeneous charge natural gas engine using P S C . F o l l o w i n g these results, B r o w n (2003) reported s imi lar f indings i n a dual fuel regime, w i t h natural gas as the P S C fuel and port injected gasoline as the ma in charge. B r o w n also evaluated a mono-fuel strategy, in w h i c h a gasoline direct injector p rov ided the pi lot charge, and the bulk charge was, again, port injected. The results o f the mono-fuel study were not as posit ive, o w i n g to diff icul t ies in re l iably igni t ing the pi lot charge. This thesis describes a study i n w h i c h the concept o f local charge stratification was removed from its famil iar context o f homogenous charge engines, and combined w i t h a fu l ly stratified approach. The goal was to determine i f mixture enrichment loca l to the spark p lug w o u l d a id the igni t ion and combust ion o f a stratified D I fuel charge. The specific research objectives were to: 6 1. A d d a natural gas direct injection system to a single cy l inder research engine. 2. Determine i f je t guided stratification, created by late cyc l e direct injection, w i l l facilitate g loba l ly lean operation. 3. Compare the direct ly injected engine's performance when operating w i t h conventional spark igni t ion , to combust ion init iated by P S C . It was hypothesized that P S C w o u l d improve fuel usage i n the D I S C engine through two modes. First , it was thought that the enflamed pi lot charge w o u l d be m u c h larger in vo lume than the ion ized region produced from a standard spark discharge. Th i s larger vo lume o f in i t i a l ly active chemistry w o u l d increase the probabil i ty o f successfully ign i t ing the D I bu lk p lume 's combustible region. Secondly , it was thought that the extra ign i t ion energy o f P S C w o u l d promote combust ion through regions o f the D I plume where m i x i n g w i t h the w o r k i n g gas has progressed to an otherwise over ly lean composi t ion . I f real ized, these improvements w o u l d reduce unburned hydrocarbons levels and increase efficiency through improved combust ion and reduced misfire. A s outl ined qual i ta t ively i n F igure 1.2, this hypothesis supports an increased range o f un-throttled load control for SI engines. Brake specific fuel consumption ( B S F C ) is a measure o f engine efficiency that describes the amount o f fuel necessary to produce a quantity o f shaft work ; thus lower values are preferred. B rake mean effective pressure ( B M E P ) represents the range o f loads an engine is expected to operate over. I f successful, the D I S C strategy w o u l d extend the range o f un-throttled load control beyond that o f a homogenous lean burn engine. 7 Baseline: Throttled Load Control BMEP Figure 1.2 Fuel Efficiency of Load Control Strategies (Lean Burn and Lean Burn with PSC Approximated from Reynolds 2004) 2. EXPERIMENTAL APPARATUS The experimental invest igat ion described in this thesis was performed on a R ica rdo H y d r a , four-stroke, single cy l inder research engine. The H y d r a is a f lexible engine test-bed, w h i c h is easily configured to operate w i t h a variety o f fuels and igni t ion methods. F o r the purposes o f this study, it was modi f ied to support a gaseous direct injector, i n addi t ion to a P S C spark p lug . A s further described i n this section, the H y d r a is thoroughly instrumented to record a wide variety o f performance indicators and emissions data. 2.1 RESEARCH ENGINE The R ica rdo H y d r a is produced by R ica rdo Consu l t ing Engineers o f the U n i t e d K i n g d o m (Ricardo 2006). It is designed specif ical ly for research, w i t h h igh ly modular components that enable it to accept a wide range o f pistons, cyl inder heads and va lve trains. Th i s capabil i ty a l lows the H y d r a to emulate a wide variety o f diesel , gasoline, and alternative fuel engines. Fo r this study, the H y d r a was fitted w i t h a cast a lumin ium, spark ign i t ion cy l inder head, wi th two vert ical valves actuated by a single overhead cam shaft. The combust ion chamber consisted o f a flat fire-deck and bowl- in -p i s ton arrangement. The relevant geometry is detailed in T a b l e 2-1 below, and a d rawing o f the pis ton c r o w n is available in A p p e n d i x A . 9 Table 2-1 Ricardo Hydra Spark Ignition Configuration Geometry Bore 81.5 mm Stroke 88.9 mm Clearance Volume 54.7 cc Displaced Volume 463.3 cc Compression Ratio 9.47:1 Inlet Valve Diameter 36.0 mm Inlet Valve Open 12° B T D C Inlet Valve Close 56° A B D C Exhaust Valve Diameter 33.5 mm Exhaust Valve Open 56° B B D C Exhaust Valve Close 12° A T D C U B C ' s Hydra is coupled to a D C dynamometer which provides regenerative load absorption, as well as motoring power. The dynamometer is trunnion mounted, with a swing-arm that is connected to a load cell for torque measurement. 2.2 DIRECT INJECTION SYSTEM 2.2.1 Direct Injector The bulk charge direct injector was a Westport model J43 gaseous mono-fuel injector provided by Westport Innovations Inc. of Vancouver B C . The J43 is an iteration in a line of prototype injectors designed for research purposes, and is not a commercially available product. The J43's gas needle is actuated by an interaction between a magnetic field and the magnetostrictive material, T e r f e n o l - D ™ . Terfenol is a ferrous metal alloy with a highly repeatable response to magnetic fields (Etrema Products Inc 2006). As coils within the injector become energized, magnetic domains within the Terfenol become realigned, causing the material to expand. This movement opens the injector's gas needle, allowing fuel to flow. In this study, fuel was metered by varying the commanded pulse width (PW), which determines the duration of 10 needle lift. Further rate shaping is possible by vary ing the J43's needle lift; however, this method did not suit the desired operating range, so m a x i m u m lift was commanded in all cases. 2.2.2 Cylinder Head Modifications T h o u g h the U B C H y d r a ' s complement o f cyl inder heads support a wide variety o f fuelling methods, none available were intended for direct injection combined with spark ignition. W i t h guidance from Westport, a design study was undertaken to determine the best approach for modi fy ing a spare SI head to include a direct injector. A representation o f the resulting design concept is shown in Figure 2.1 below, and further documentation from the design study can be found in A p p e n d i x A . Figure 2.1 SI C y l i n d e r H e a d M o d i f i e d for J43 Direct Injector 11 2.2.3 Jet Development Hill and Ouellette's (1999) review of gaseous jets, as well as follow-on work by Iaconis (2001) served as the main guide for locating and sizing the injector's single gas orifice. In its base position, the injection plume was aimed directly towards the ignition source so as to ignite the jet as it exits the injector. This is analogous to the Texaco TCCS engine: a historical example of a DISC engine, described by Abata (1986). As illustrated in Figure 2.2, small adjustments to the rotation angle of the injector's central axis offered a limited degree of control over the concentration of air-fuel mixture directed at the spark plug electrodes. Small variations in this angle could alternatively direct the jet's rich core or the more vigorously mixed shear layer at the jet boundary towards the electrodes. 0° (Neutral Angle) 8° Towards Intake Valve 16° Towards Intake Valve Injector Tip Spark Plug Fuel Jet \ Cylinder Bore Injector Angle Figure 2.2 DI Jet Angle Adjustment Range For the purposes of design, the quasi steady state jet and vortex ball model was assumed for the plume's shape (Turner 1959, from Hill and Ouellette 1999). The jet's central axis was given a 10° down angle relative to the fire deck in its base position. As suggested by Hill and Ouellette, 12 m i n i m a l w a l l interference cou ld be expected at this angle, provided the pressure ratio across the injector's outlet remained sufficiently h igh (>4). 2.2.4 Injector Characterization The orif ice diameter was s ized us ing the assumptions o f isentropical ly expanded ideal gas, described by Iaconis (2001). In his characterization o f a Westport J41 injector (a predecessor to the J43), Iaconis found this model to agree w e l l w i th experimental results, though he went on to evaluate several refinements that attempt to predict losses w i t h i n the injector. F o r the purposes o f this study, internal losses were approximated by a 10 percent loss i n stagnation pressure, as suggested by H i l l and Ouellette; sample calculations are avai lable in A p p e n d i x B . A n orifice diameter o f 0 .5mm was found to match the injector's linear response range to the desired f low rates. Th i s is i l lustrated i n F igure 2.3, w h i c h shows the injector 's characterization data, recorded at Westport 's test fac i l i ty , as w e l l as predicted f low rates. 3 W X o 0. CB s 35 30 25 20 15 10 5 0 • Westport J43 Pulse Width Characterization - Calculated Flow Rates • • • l~ 1 ( U B C Hydra) • X~3 (UBC Hydra) 0 1 2 3 4 5 Injection Pulse Width (ms) Figure 2.3 J43 Pulse W i d t h Character izat ion 13 2.3 PSC SYSTEM A US patent has been granted to U B C for the PSC system (Evans 2000). The patent document contains a thorough description of the PSC concept and its intended use in SI engines. The particular realization of the PSC concept used in this study is functionally identical to that described by Brown (2003), and changed very little from Reynolds (2001). The reader is referred to these documents for a detailed description of the PSC system; though a representation of a PSC spark plug is shown in Figure 2.4, and a list of the key specifications follow in Table 2-2. 14 Table 2-2 PSC Specifications Spark plug type Bosch XR4CS Heat Range 4 Reach 3/4 in Capillary Tube 1/16 in Stainless Steel Solenoid Omega SV122 Supply Regulator CCA 4922801 -01 -000 Supply Pressure 27 bar 2.3.1 PSC Timing Characterization Previous work on the PSC system had indicated a significant delay between the time an injection command is sent, and the actual start of injection (SOI). A prior understanding of this delay is important, since there is no relevant feedback to indicate the actual SOI while the engine is running. Reynolds (2001) performed a characterization study to quantify this delay; however the solenoid he and Brown used to meter PSC flow failed prior to this study. Though the replacement was of the same make and model, Reynolds' results could not be assumed accurate for the new solenoid; thus another study was performed. The characterization study was performed using the experimental set-up depicted in Figure 2.5. Data was collected and analyzed using the Hydra's data acquisition and processing system, as described further in 2.4 of this thesis. Injection commands were sent to the PSC solenoid at a frequency and duty cycle that were equivalent to engine operation at 2000rpm. Timing was measured in terms of crank angle degrees (CAD); thus one crank angle degree was equivalent to 0.083ms. Pressure data for 100 consecutive cycles was acquired at half crank angle degree increments. 15 T w o Stage Regulator Thermal Mass Flow Meter With Pressure Transducer From H i g h Pressure Storage Atmospheric Vent A l u m i n u m Pressure Jig Enclosed Volume: ~3cc P S C Spark Plug T o DAQ Piezoelectric Pressure Transducer Figure 2.5 PSC Timing Characterization Process Flow Diagram Trials were conducted at supply pressures of 8, 15 and 20 bar(gauge). The two higher pressures were chosen to represent the pressure differential across the capillary tube during moderately early and moderately late PSC injection timings. The lower pressure was selected following an analysis in which the capillary tube flow was modeled assuming Fanno, conditions (Wilcox 2003, White 1994; see Appendix B for a sample calculation). This analysis revealed that, during engine operation, pressure in the combustion chamber during the compression stroke was sufficient to prevent the PSC flow from frictionally choking within the capillary tube. However, pressure differentials representative of engine conditions would result in choking in the characterization test rig, where the back pressure was close to atmospheric. According to the Fanno model, 8 bar(gauge) was near the maximum supply pressure that would not produce frictional choking in the test rig. This lower pressure trial was conducted as a comparison in which the compressibility effects, rather than pressure differential, were more closely matched to actual conditions. 16 The commanded injection pulse widths ( P W ) were set to 8.5 C A D (0.71ms) and 8.75 C A D (0.73ms), to coinc ide w i t h typ ica l engine operation. Due to the coarse control afforded by the solenoid, f l ow rates at these s imi la r durations were substantially different. The results o f the t i m i n g study are shown i n Figures 2.6, 2.7 and 2.8. The ini t ia l detection o f increased pressure was taken to indicate the S O I ; thus a significant lag between the t iming command and the actual event was conf i rmed. Though the poor signal-to-noise ratio makes precise measurement diff icul t , the pressure signals appear to diverge from the baseline data between 35 and 45 C A D (2.92 and 3.75 ms) after the start o f injection. Th i s result appears independent o f pressure differential and injection duration. 1.02 + Background Background Smoothed 1.01 A O 8.5 C A D P W 8.5 C A D P W Smoothed X 8.75 C A D P W S3 1.00 -| 0.99 -0.98 0.97 0 20 40 60 80 100 T i m i n g - Beginning at Injection Command ( C A D - 2000 R P M Equivalent) Figure 2.6 P S C T i m i n g Characterizat ion - 8 bar Supply Pressure 17 1.02 1.01 ¥ loo e I in CO I 0.99 0.98 0.97 + Background Background Smoothed O 8.5 C A D PW 8.5 C A D PW Smoothed X 8.75 C A D PW 8.75 C A D PW Smoothed ! Timing - Beginning at Injection Command (CAD - 2000 R P M Equivalent) Figure 2.7 PSC Timing Characterization - 15 bar Supply Pressure 1.02 1.01 ¥ i-oo s—-0.99 0.98 0.97 + Background Background Smoothed O 8.5 C A D PW 8.5 C A D PW Smoothed X 8.75 C A D PW 8.75 C A D PW Smoothed < 0 10 20 30 40 50 60 70 80 90 100 Timing - Beginning at Injection Command (CAD - 2000 R P M Equivalent) Figure 2.8 PSC Timing Characterization - 20 bar Supply Pressure Larger scale graphs (Figures 2.9, 2.10 and 2.11) reveal the injection durations to be greater than an order of magnitude longer than the commanded pulse widths. It is unclear to what extent this is due to signal delays as the control driver amplifies the command pulse, inertia of mechanical components, or pressure build-up within the PSC line. However, it is unlikely that combustion pressures would allow the actual injection event to continue as long as observed in the test rig. Consequently, a characterization of the end of injection is not available from this data. It is interesting to note that the test rig pressure remains high more than 180 CAD after the start of injection. This may warrant further investigation during future PSC work, since it indicates the possibility that gas flow resumes once combustion chamber pressure drops during blow-down. 1.02 1 """""Background Smoothed 1.01 - 8.5 CAD PW Smoothed 8.75 CAD PW Smoothed 0.97 H 1 1 1 : 1 0 90 180 270 360 Timing - Beginning at Injection Command (CAD - 2000 RPM Equivalent) Figure 2.9 PSC Characterization, 8 bar, Coarse Timescale 19 1.02 l.oi i ¥ e 1 . 0 0 1 % 0.99 cu 0 .98 H 0.97 Background Smoothed 8.5 C A D P W Smoothed 8.75 C A D P W Smoothed 0 90 180 270 360 T i m i n g - Beginning at Injection Command ( C A D - 2000 R P M Equivalent) F i g u r e 2.10 Characterizat ion, 15 bar, Coarse T imesca le 1.02 1.01 ¥ e I.OO a % 0.99 cu 0.98 0.97 *—• v B a c k g r o u n d Smoothed 8.5 C A D P W Smoothed 8.75 C A D P W Smoothed 0 90 180 270 360 T i m i n g - B e g i n n i n g at Injection C o m m a n d ( C A D - 2000 R P M Equiva len t ) F i g u r e 2.11 Characterization, 20 bar, Coarse T imesca le 2.3.2 PSC verification D u r i n g the course o f this invest igat ion a concurrent study, w h i c h i n v o l v e d m o d i f y i n g the H y d r a ' s aspiration system to inc lude an E G R loop, was also i n progress. The addit ion o f this system required substantial modi f ica t ion to the intake and exhaust routes. The result ing effect on the H y d r a ' s volumetr ic eff ic iency and residual gas fraction were explored i n detail by W i l l i a m s 20 (2006); here it is sufficient to mention that engine characteristics in this regard had changed since the work of Reynolds (2001) and Brown (2003). In light of these changes, the pre-mixed, lean-burn application of PSC was revisited before evaluating new applications of the PSC concept. This was undertaken to determine if PSC would still produce the previously documented results, following the engine modifications. A single operating condition was chosen near the engine's lean misfire limit; Table 2-3 provides a summary of the operating parameters. Ignition timing was adjusted to minimum advance for best torque (MBT) according to the procedure described by Reynolds (2001). Data was collected with unaided spark ignition, and then at the same global air-fuel ratio, but with 2.6% of the fuel injected as PSC. Table 2-3 PSC Verification Operating Parameters Trial 1 2 PSC Fuelling (%) 0 2.6 Premixed Fuelling (%) 100 97.4 Engine Speed (rpm) 2000 2000 Overall Relative AFR ( A ) 1.6 1.6 PSC commanded SOI relative to Ignition (CAD) 41 41 Ignition Timing MBT MBT Throttle Position (% open) 100 100 A significant difference in BSFC was apparent between these two treatments (386 g/kWh with PSC, and 647 g/kWh without). This is likely due to the reduction in partial burn events observed in the PSC case. Figure 2.12 is a histogram showing the distribution of integrated heat release (IHR) for the two conditions. In each case, IHR is individually calculated for 100 consecutive cycles. Partial burn is interpreted as combustion events which fall below the normally distributed peaks concentrated above 800 kJ/m3. According to this definition, the PSC data show three percent partial burn, whereas the non-PSC case shows 32%. 21 Integrated Heat Release (kJ/m) Figure 2.12 Histogram of Integrated Heat Release The effects of varying PSC injection timing relative to ignition have been described by Reynolds (2001), and are not a major focus of this thesis. The timing of the current PSC system was "tuned" qualitatively to a commanded SOI lead of 41 CAD before ignition. According to the PSC timing study described above, this would produce an actual SOI lead of up to six crank angle degrees. This timing appeared optimal and was adopted for all subsequent PSC experiments. Shorter lead durations were accompanied by a reduction in torque; whereas lengthening the duration by up to 15 C A D had little effect, though a similar drop in performance was noticed beyond this. The fuel efficiency results suggested that, despite changes to the engine, PSC still produced the significant baseline improvement established in previous studies. This verification allowed a more meaningful evaluation of PSC in the current DI application. 22 2.4 INSTRUMENTATION AND DATA ACQUISITION U B C ' s R ica rdo H y d r a is equipped w i t h a ful l instrumentation package suitable for thermodynamic research. A n a l o g u e Signals from the instruments are received by condi t ioning boards, where they are filtered and ampl i f ied as needed. Af te r this in i t ia l condi t ioning , a data acquisi t ion card installed i n a Pent ium III based computer samples the signals for conversion to digi tal data. Fast response instruments, such as the in-cyl inder pressure transducer, are sampled in 0.5 crank-angle degree increments (24 k H z at 2000rpm). Da ta that is not evaluated on a cyc le to cycle basis, such as engine speed and torque, fuel and air f l o w rates, and exhaust emissions are sampled at lower rates. These parameters, referred to as "performance data", are sampled at approximately 1.3Hz. Table 2-4 provides a list o f the H y d r a ' s instrumentation. 23 T a b l e 2-4 Ricardo Hydra Instrumentation Measurement Manufacturer Model Range Intake Air Flow Rate Meriam 50MW 20-1.5 0-30 scfm Intake Air and Exhaust Temperature Omega 1/8" K-Type - 2 0 0 - 1250°C Intake Manifold Pressure Differential Sensym LX1803AZ 0-30 psia Pressure Transducer (Intake Air Flow Rate) AutoTran 600 D-014 0 - 20" H20 Exhaust Relative Air-Fuel Ratio E C M 2400G 1 =0.4- 10.0 In-Cylinder Pressure Transducer A V L QC33C 0 - 200 bar Natural Gas Flow Meter/Controller - Port Fuel MKS Instruments 1559A-100C-SV 0 - 100 slm Natural Gas Flow Meter - PSC MKS Instruments 179A-24-S3BM 0-20 slm Natural Gas Flow Meter - Direct Injection Endress + Hauser Promass 80A DN02/1/12" 0-20 kg/h Engine Crank Angle / Speed US Digital H1-360-IE 0 - 10,000 rpm Engine Torque B L H Load Beam Transducer 0 - 50 Nm Carbon Dioxide Emissions Beckman 880 0 - 20% Hydrocarbon Emissions Ratfisch RS-55 0 - 10,000 ppm Carbon Monoxide Emissions Siemens Ultramat21P 0 - 10,000 ppm Nitrogen Oxide Emissions API 200AH 0 - 4500 ppm A custom Lab View application, referred to as the Dynoserver, applies gains and offsets as appropriate to convert the data to engineering units before passing it to a remote computer located in an adjacent control room. This second computer operates two Lab View applications, referred to as the Dynoclient (Figure 2.13) and the Pressureclient (Figure 1.2). These display real-time operating parameters and the in-cylinder pressure trace respectively; they also record data for subsequent off-line analysis. 24 t> Dynoclient.vi BMB>| SPEED] A«*aged Torque (Nm)p^™ 7 " woo m4 Volumetric Disp Jjfo.459 Spark Tirrlno.1 ;j |-Z3.0 ~ « E £ l PSC Irrrectron Ttrmqi Throttle MS™ MPukeWttn l ;Jfo5o MAIN PARAMETERS 0.0 1.0 2.0 3.0 1 1 y . T_ | _> 0.5' o.'e 1.0 1.2 i.s -—y y -> i ; 0,9 1.0 1.1 1.2 <r 1—'— 0.00 0.01 0.020.03 0.0 2S.0 50.0 NG fuel Ficw Rate! 1 , kq/hour| 11.355 Intake Man. Pressure! D9837 | : baKajl Exhaust Han, Pressurel t: « ' « I bartol P5C_Ftow Retej : 1-0.004 j kg/hourl P5C Pressure} J barfgll 1-0.29 Unused Teffljyjtue ' | dead Unused Channell 0.90 0.95 1.0 1.1 Barornetrk: Pressurel j g S a r l ba,(a)| Mass tambdaj O.OO 0.50 1.00 1.50 2,00 Exhaust Lambda] DIAGNOSTICS 0 2 4 6 8 10 * T • 50 r* » ) 100 120 0 50 100 150 Oil Pressurel : I&.16 | baKq)l Oil Temperaturel 175.8 [ d e a d Water Out Temperaturel | drocl Intake Man, Tempi 0 200 -100 600 800 4 j f 1r 20 ;25 " 30 35 ENGINE AIR Engine Air Temperaturel J18.7 dead O.0: :; 1.0 2.0 3.0 .. |Z9.0 Esthaust Man. Tempi : Test Cei Temperaturel iff "\ d e a d Relative HumMltyl $7T~~ SI Air How Rate (IFOI 0,65 |^20(rjffP)l LFE Air Pressure} 11,0046 [ barjaj (Other Injection Trming and Pursewridth here) (RESERVE SPACE FOR PRESSURE TRACE) corrals 1-0.003 | r-0-27 02i 0 -13.0-• 10.0-1-0.01 'A 5.0-0.0-CO] f-io ppmj •:, CH4j j-1282 C02J Ko62 Test# I Ir*akrjC02 O.OOOB | t POINTS 3 l O O Figure 2.13 Dynoclient Interface Window 25 Intake Figure 2.14 Pressureclient Interface Window An exhaust oxygen sensor provided a rough guide for determining the Hydra's air-fuel ratio. However, since the effectiveness of this instrument is reduced by poor combustion quality, air-fuel ratios reported in this study are calculated from intake flow measurements. 26 2.5 ENGINE CONTROL SYSTEM The Hydra's speed and throttle position are set by adjusting potentiometers on a control console. Engine speed is maintained at constant set points by a feedback controller. Input from the potentiometer is compared with output from a tachometer coupled to the dynamometer and discrepancies are corrected by the thyristor drive. Input to the throttle potentiometer is mapped through a transfer function, then output to a servomotor that is directly linked to the throttle plate. Ignition and fuel injection are controlled by a custom built timing system. Commands are entered into a Lab View application (Figure 2.15), which runs on the same computer as the DynoServer. These commands are passed to a counter/timer card, where they are indexed to the crank-shaft's angular position, then output as trigger signals for the hardware drivers. The drivers generate the ignition voltage and current, as well as the lift and hold currents for the PSC injector. The timing, duration and lift commands for the natural gas direct injector are sent to a driver system supplied by Westport. • M ®| ••] Timing Control T i m i n g D u r a t i o n ( O N J l9*ion Control ! * |-36.0 -90.0 -60.0 -40.0 -20.0 0.0 20.0 31 fro 1 • ™ 1-23.00 -90.0 -60.0 -40.0 -20.0 0.0 20.0 3! |^8.00 OFF j NG Stratfied Injector ' 1 * ' ' 3J-56.00 -90.0 -60.0 -40.0 -20.0 0.0 20.0 3! 1 fl8.25 Q F F J Main NG Direct Injector O F F ) PortGasolneInjection '« Trrrr 5J194.5 0.0 1 00.0 200.0 300.0 400.0450.0 — 1 , Stop ) 3168.5 Figure 2.15 Timing Controller Interface Window Ported natural gas flow is controlled by an MKS thermal mass flow meter/controller. Set points are input from a remote interface in the control room. 27 2.6 DATA PROCESSING Data recorded during experimentation's imported into a combination of spreadsheets, which partially automate the task of data processing. Pressure data is recorded in half crank angle degree increments for 100 consecutive cycles. The spreadsheets calculate the. 100 cycle mean at each half crank angle degree, though data for each cycle is retained for individual analysis if necessary. For performance data, the spreadsheets calculate each parameter's sample mean (X), standard deviation (Sx), coefficient of variation (COVx) and 95% confidence interval (CI9s%) for the mean. The contingent expressions are (Bowker and Lieberman, 1972): X = -txi (Eq.2.1) (Eq. 2.2) (Eq. 2.3) (Eq. 2.4) Where ^ 0 0 2 5 «-i ' s m e 2.5% point of the t distribution with n-1 degrees of freedom. The 95% confidence interval for the mean is the range that, if calculated repeatedly from multiple samples, would contain the true population mean 95% of the time. 28 2.7 STATISTICAL PILOT STUDY Drift in operating parameters such as engine speed and torque is an observable characteristic of the Hydra test-bed. In order to support the assumption of steady-state conditions while logging data, it is desirable to keep the collection interval short relative to the time scales of drift. Since the logging rate is fixed by the equipment, the number of sample points determines the duration of the collection interval. A pilot study was therefore undertaken to determine a sample size appropriate for characterizing operating conditions quickly, yet with a reasonable level of statistical certainty. Due to difficulty in modifying the pressure data sampling process, this study was limited to performance data. Trials were conducted at a moderately rough operating point (COVIMEP ~ 5%) so as to support a conservative estimate of the data's spread. Details of the engine parameters are given in Table 2-5 and Table 2-6 below. T a b l e 2-5 Statistical Pilot Study Operating Condition Engine Speed 2000 rpm Load Wide Open Throttle Premixed Fuelling lkg/hr ( A K 1.3) PSC Off Direct Injection Off T a b l e 2-6 Statistical Pilot Study Sample Sizes Trial Sample Size 2.5% point t distribution 1 11 2.228 2 21 2.086 3 31 2.042 4 41 2.021 5 51 2.009 6 76 1.992 7 101 1.984 29 Figures 2.16, 2.17 and 2.18 be low show the half-length o f the 9 5 % confidence interval as a function o f sample size for several key performance indicators. The values shown are added to and subtracted f rom the sample mean to construct the confidence interval . These are calculated according to E q . 2.4, us ing each ind iv idua l sample 's standard deviat ion. m 0.0 11 21 31 41 51 76 Number of Sample Points 101 Figure 2.16 Torque. Half - length o f CI9} at V a r y i n g Sample Sizes 11 21 31 41 51 76 Number o f Sample Points 101 Figure 2.17 Speed Half - length o f CI95 at V a r y i n g Sample Sizes 0.3 ^ 0.2 -I o E 0 1 "33 3 0.0 Sli 11 21 31 41 51 76 Number o f Sample Points 101 Figure 2.18 F u e l F l o w Half - length o f CI95 at V a r y i n g Sample Sizes 30 The incentive for increasing the sample size beyond 51 data points was m i n i m a l ; therefore performance data for a l l subsequent experimentation was sampled i n sets o f 51 data points. 2.8 UNCERTAINTY ANALYSIS The uncertainties due to the instrument error associated w i t h each measured parameter are provided i n Table 2-7 be low. T a b l e 2-7 Measurement Uncertainty Measurement Uncertainty Intake A i r F l o w Rate ± 0.3 scfm Intake A i r and Exhaust Temperature ± 2 . 2 ° C Intake M a n i f o l d Pressure Different ial ± 0.6 P s i Pressure Transducer (Intake A i r F l o w Rate) ± 0.2" H 2 0 Exhaust Re la t ive A i r - F u e l Rat io ± 0.009 In -Cy l inde r Pressure Transducer ± 0.4 bar Natura l Gas F l o w Mete r - P S G ± 0.12 s lm Natura l Gas F l o w Meter /Cont ro l le r - Port F u e l ± 0.6 s lm Natura l Gas F l o w Mete r - Di rec t Injection ± 0.011 Eng ine C r a n k A n g l e / Speed ± 0 . 5 ° / ± 2 . 5 rpm Eng ine Torque ± 0.5 N m Carbon D i o x i d e Emis s ions ± 0 .2% Hydroca rbon Emiss ions ± 100 p p m Carbon M o n o x i d e Emiss ions ± 100 ppm Ni t rogen O x i d e Emis s ions ± 50 p p m The overal l uncertainty for each measured variable is reported as the instrument error, combined in quadrature w i t h the 9 5 % confidence interval, as recommended by the Journal o f F lu ids Engineer ing ( R o o d and Te lo in i s , 1991). F o r a measured parameter, the overa l l uncertainty w is g iven by: w = V#2 +P2 ( E q . 2.5) Where B is the f ixed uncertainty due to instrument error, and P is the 9 5 % confidence interval , calculated as i n sect ion 2.6 above. 31 For a parameter R, that is calculated from several measured variables R(xi,x2, ...x„), the uncertainty of each variable (wh w2, ...w„), is propagated in the following manner (Holman, 2001): ( dR 2 (dR ) 2 (dR 1 w, + w2 + . w„ > {dx2 ) (Eq. 2.6) As an example, BSFC is a fuel efficiency indicator that is particularly relevant in this study; the overall uncertainly for BSFC, at an operating condition that involves DI and PSC fuelling,, is reported in the manner described below. First, BSFC (g/kWh) is defined as: BSFC = (Eq. 2.7) Where mf is the total fuel flow (g/h), comprised of DI ( mDI) and PSC (m P S C ) contributions: mf = mDl + mPSC (Eq. 2.8) and Pb is the brake power (kW), defined as: * 60 where N is the engine speed (rpm) and rb is the brake torque (Nm). So the uncertainty in fuel flow according to Eq. 2.6 is: 32 mf V  mDI mPSC (Eq. 2.10) and the uncertainty in power is: 60 (Eq. 2.11) so the overall uncertainty for BSFC becomes: ^2 ( m w BSFC 1 p m/ + p 2 *i v ^ y (Eq. 2.12) 2.9 CLOSING REMARKS The PSC timing characterization and verification study, as well as the statistical pilot study were conducted as a collaborative effort between the Author and fellow M.A.Sc. Candidate, Mr. David Williams. The Author wishes to recognize Mr. Williams' contribution to these segments of this thesis. 33 3. METHODOLOGY As further described in this chapter, the experimental investigation was carried out in two parts. A primary study was aimed at characterizing the DI and PSC combustion system in terms of operating characteristics at varying levels of stratification. A secondary study was focused on the viability of the PSC pilot charge; and also involved re-visiting stratified combustion under conditions that guaranteed an ignitable pilot charge. 3.1 DEFINITIONS: EARLY AND LATE DIRECT INJECTION One of the goals of this study was to implement a late direct injection strategy for charge stratification. As defined by Abata (1986), this is a combustion system in which heat release is controlled by varying the injection rate. Ignition immediately follows the start of injection; then combustion propagates to engulf the plume as it issues from the injector. In this study, firing the spark plug shortly after the start of injection was the base strategy for igniting the stratified charge; however it was also anticipated that releasing the main fuel jet into the already burning PSC charge would provide more reliable ignition. Despite the original hypothesis, practical considerations kept most operation biased towards early DI timings. An early direct injection combustion system is one in which ignition occurs after injection is complete. Heat release is controlled by mixing processes, as well as the chemical kinetics of combustion. Stratification in early direct injection engines, when desired, is established through wall or air-guided approaches (Zhao 2002). This study did not involve a mechanism for maintaining the stratification of an early DI charge, so early injection timings were expected to produce a relatively homogenized charge. 34 3.2 DI TIMING SWEEPS The primary invest igat ion i n v o l v e d sweeping the direct injection t i m i n g across a range that a l lowed for v a r y i n g degrees o f fuel-air m i x i n g before igni t ion . The var iable o f interest was the injection t i m i n g relative to ign i t ion , as this determined the duration avai lable for fuel to m i x wi th the surrounding air before combust ion began. B e g i n n i n g wi th early injection t imings, a combust ion mode close to p remixed was expected to progress towards a stratified regime as the t iming was delayed later into the cyc le . The igni t ion method was var ied between an unaided spark and P S C for each operating condi t ion . In this manner, it w o u l d become apparent i f P S C ignit ion offered any improvements across the va ry ing degrees o f stratification. The hypothesized result is illustrated graphical ly in F igure 3.1. A n increase i n B S F C was expected at intermediate t imings in w h i c h the m i x i n g interval w o u l d be too short to produce a homogeneous mixture, yet too long to effectively produce a late D I stratified charge. pa •Without PSC With PSC Spark Early (Premixed) DI Timing Late (Stratified) F i g u r e 3.1 Hypothe t ica l D I Fue l Ef f i c i ency w i t h T i m i n g 35 In practice, the D I t i m i n g range was bound early i n the cyc le by inlet va lve closure, and later in the cyc le by severe combust ion degradation. Operat ing points were established by f i x i n g the D I t iming in 10 C A D increments and then adjusting igni t ion t im ing to M B T . F o r completeness, experiments were attempted at late t imings , beyond the onset o f severe performance degradation. In these cases, the fue l l ing systems were energized long enough to brief ly exhibi t results, and then deactivated so that continued motor ing cou ld dilute the h igh levels o f unburned fuel that were evident at these condi t ions. In these cases, combust ion was not stable enough to proceed safely through the duration required for data col lect ion, without p roducing an unacceptable r isk o f explos ion in the exhaust mani fo ld . 3.2.1 PSC Settings For operating condit ions at w h i c h data was recorded, the commanded P S C S O I was held constant at 41 C A D before igni t ion . Th i s t im ing was selected since it was k n o w n to be effective in the premixed, lean burn case where P S C is k n o w n to produce a significant result (see Sect ion 2.3). The P S C f l o w rate was held between 2% and 3 % o f the total fuel charge; precise control w i th in this range was not avai lable f rom the control system used. A g a i n , for completeness, the t im ing o f P S C relative to igni t ion was var ied over a wide range (approximately 41 ± 3 0 C A D ) at most operating condit ions. Qual i ta t ive observation indicated that v a r y i n g this t im ing had no effect, so the relative t i m i n g o f the p remixed study was adopted for consistency in recording data. 3.2.2 Air Fuel Ratio For the t iming sweeps, the g loba l relative air-fuel ratio was he ld constant at A= 1 . 3 ± 0 . 0 8 . Th i s moderate air-fuel ratio was concentrated enough to produce rel iable combust ion i n a homogenous mixture. K n o w i n g that the g loba l mixture was inherently combust ible , performance degradation observed as the D I t i m i n g was retarded could be attributed to the effects o f stratification; and benefits due to the use o f P S C had the potential to reduce this degradation. That a g loba l ly 36 stoichiometric mixture was not used is due to safety concerns: as a precaution against severe backfire result ing from h i g h instances o f misfire at late t imings , a lean mixture was preferred. 3.2.3 Direct Injector Jet Angle The D I t im ing sweeps were repeated at each o f three jet orientations. These were chosen to direct va ry ing concentrations o f air-fuel mixture towards the electrodes, i n the manner described in section 2.2. A s d iagrammed previously in Figure 2.2, angles o f 0 ° , 8 ° , and 16° were selected to direct r ich , moderate, and lean mixtures respectively towards the electrodes. Th i s approach to mixture control has more va l id i ty for late D I , since it depends on a transient jet structure w h i c h , i n the case o f early injection, w i l l have dissipated w e l l before ign i t ion . The angle is defined relative to a line bisect ing the centre o f the injector sac and the centre electrode o f the spark plug, as shown i n F igure 2.2. A l l testing was conducted at 2000 rpm and wide open throttle. The D I supply pressure was maintained at 3000psi , this represents a pressure ratio o f approximately 12, relative to m a x i m u m motored in-cy l inder pressure. 3.3 PILOT CHARGE STUDY F o l l o w i n g the pr imary investigation, it was hypothesised that the P S C charge was fa i l ing to ignite under the condit ions o f the t i m i n g sweep study. G i v e n the posi t ive results obtained from the applicat ion o f P S C to homogenous lean burn, it may be that the presence o f a surrounding air-fuel mixture, even very dilute, is required for the P S C charge to ignite. The pi lot charge study was a prel iminary explorat ion o f this hypothesis. Ini t ia l ly , the engine was run on P S C alone. A f l o w rate and relative t i m i n g typ ica l o f that used for the t im ing sweeps was selected, and pressure arid emissions data were evaluated against motored 37 engine data for indicat ions o f combust ion. A very lean background mixture was then inducted through the intake mani fo ld . Th i s mixture was too dilute to support combust ion throughout; however, when P S C was appl ied, some evidence o f combust ion was detected. The concentration o f the background mixture was adjusted to be as weak as possible, w h i l e s t i l l support ing a l imi ted combust ion event w h e n P S C was i n use. F i n a l l y the shortest burst avai lable from the direct injector's control system was introduced. Th i s was t imed to lag the inject ion and igni t ion o f the P S C charge. In isolat ion, each o f the three fuel sources w o u l d have produced an extremely lean mixture; however, as out l ined i n Table 3-1, the combinat ion produced a g loba l ly r i ch charge. T a b l e 3-1 F u e l l i n g Contr ibut ions F u e l Source F l o w Rate (kg/h) A (In Isolation) D I 0.56 2.14 P S C 0.029 42.3 P remixed 0.71 1.69 O v e r a l l 1.30 0.92 The injector orientation was maintained i n the eight degree posi t ion for the duration o f the pi lot charge study. 38 4. RESULTS AND DISCUSSION One of the primary research objectives was to determine if jet-guided stratification, created by late direct injection, would facilitate globally lean load control. As will be discussed in this section, a negative result was obtained: without significant time for premixing, combustion quality was poor, with high instances of misfire. Results for direct injection timing sweeps are therefore confined to relatively early injection timings in which the directly injected fuel had the opportunity to mix with the surrounding air prior to ignition. The parallel objective was to compare the directly injected engine's performance when ignited by un-aided spark ignition, to when ignited by PSC. As will be discussed further in this chapter, the results did not reveal a significant difference between these two ignition methods. 4.1 DI TIMING SWEEPS 4.1.1 Results BSFC data.obtained from the zero and eight degree injector angle timing sweeps are presented in Figures 4.1 and 4.2. In keeping with the convention of timing advance increasing in the negative direction, the injector lead is presented in CAD after ignition. For example, a relative timing of -70 CAD indicates that injection began 70 CAD before ignition. 39 900 600 A u O H 300 O B S F C - DI Without P S C O B S F C - W i t h P S C i <> <P -100 -90 -80 -70 -60 -50 -40 -30 D I SOI T iming Relative to Ignition ( C A D After Ignition) -20 F i g u r e 4.1 D I T i m i n g Sweep B S F C Result (0° Injector A n g l e ) 900 600 A CQ 300 O B S F C - D I Without P S C O B S F C - D I Wi th P S C -100 -90 -80 -70 -60 -50 -40 -30 D I SOI T iming Relative to Igntion ( C A D After Ignition) -20 F i g u r e 4.2 D I T i m i n g Sweep B S F C Results (8° Injector A n g l e ) In the zero and eight degree injector angle cases, a rapid increase i n B S F C is evident as the S O I lead reaches a m i n i m u m threshold. Th i s threshold is shorter i n the zero degree case than for eight degrees. A t these early t imings , the jet ' s in i t ia l structure w o u l d have dissipated w e l l before igni t ion. Consequent ly , the difference in m i n i m u m threshold is l i k e l y due to the influence o f bulk f lu id mot ion, intake induced or f rom the D I jet, rather than on w h i c h component o f the jet intersected the spark p lug . The 16° angle y ie lded such poor combust ion at a l l t imings that data was not col lected. The fuel eff iciency data supports a negative result for the ma in research questions o f this study. The cr i t ical decl ine in performance demonstrates that the direct injection system installed was not able to produce a re l iab ly ignitable stratified charge. Furthermore, i n the early direct injection regime, the data does not show a significant difference between trials conducted w i t h P S C igni t ion and those ignited by conventional spark igni t ion. There is inconsistency between w h i c h o f these cases display more favourable fuel eff iciency; and i n a l l cases there is significant overlap between the uncertainty ranges o f each. These results show that the current P S C system is unable to reduce fuel consumpt ion i n a direct injection stratified charge engine. 4.1.2 Discussion In order to understand the B S F C results, it is useful to begin by character izing the D I data in terms o f the heat released per engine cyc le . Since this is a pressure-based calculat ion, it is available for each o f the 100 cycles recorded per operating condi t ion . His tograms o f heat release for a l l operating condit ions v is i ted dur ing the t iming sweeps are shown i n Figures 4.3 and 4.4. These reveal a dis t inct ly b imoda l distr ibution. The majority o f cycles are d iv ided between the first peak, a sharp concentration occurr ing between zero and 80 k J / m 3 , and a second, approximately normal distr ibution, centred on an average near 1200 k J / m 3 . The first peak is interpreted as misfire, w h i c h is subsequently defined in this thesis as combust ion events that y i e ld 41 less than 80 k J / m 3 . Par t ia l burn cycles , w h i c h are characterized by values that fa l l between the two pr imary modes, are indicated in very few cases. 160 -140 -120 -§ 100 -<L> 0 160 320 480 640 800 960 1120 1280 1440 Integrated Heat Release k J / m 3 Figure 4.3 H i s togram o f Ne t Heat Release for A l l T i m i n g Sweep Condi t ions ( 0 ° Injector A n g l e ) 42 180 160 --140 --120 - = 0 160 320 480 640 800 960 1120 1280 1440 Integrated Heat Release k J / m 3 Figure 4.4 H i s togram o f N e t Heat Release for A l l T i m i n g Sweep Cond i t ions (8° Injector A n g l e ) Mis f i r e , according to the above defini t ion, is shown in Figures 4.5 and 4.6 relative to D I t iming . The trend somewhat resembles the relationship between B S F C and D I t im ing . A g a i n , the t iming sweep data shows no consistent advantage o f P S C over standard igni t ion . 43 60 50 40 I 30 20 A io H O Misf i re - D I Without P S C O Misf i re - D I Wi th P S C o o o o o o o o o 0 -100 -80 -60 -40 -20 D I SOI T iming Relative to Ignition ( C A D After Ignition) Figure 4.5 M i s f i r e wi th D I T i m i n g (0° Injector A n g l e ) 60 50 H 40 30 20 10 O Misf i re - D I Without P S C O Misf i re - D I Wi th P S C 8 o o <6 o o -100 -80 -60 -40 -20 DI SOI Timing Relative to Ignition ( C A D After Ignition) Figure 4.6 M i s f i r e w i t h D I T i m i n g (0° Injector A n g l e ) M i s f i r e must certainly contribute to the B S F C trends; however, for the majority o f t im ing condit ions, misfire and B S F C do not correlate w e l l . Figures 4.7 and 4.8 show a poor ly defined relationship between B S F C and misfire, as w e l l as a lack o f differentiation between P S C and standard igni t ion . In these figures, the majority o f data is grouped be low 3 0 % misfire; when these groups are examined separately (sol id trend lines) the correlat ion between B S F C and misfire is very poor. u C/) CQ 1200 1000 800 600 400 200 O B S F C - DI Without P S C O B S F C - D I Wi th P S C CP 4r Linear ( A l l 0 °Da ta ) y = 12x + 211, R 2 = 0.90 Linear (<30% Misfire) y = 8.6x + 264, R 2 = 0.57 10 20 30 40 Misfire (%) 50 60 Figure 4.7 B S F C w i t h M i s f i r e ( 0 ° Injector A n g l e ) 45 u oo CQ 1200 1000 800 600 400 200 1 0 O B S F C - D I Without P S C O B S F C - D I Without P S C Linear ( A l l 8° Data) y = = 7.9x + 297, R 2 = 0.72 Linear (< 30% Misfire) y = = 4.9x + 321, R 2 = 0.30 > o 0 10 20 30 40 Misfire (%) 50 60 Figure 4.8 B S F C w i t h M i s f i r e (8° Injector A n g l e ) The relationship between total hydrocarbon emissions and misfire is s imi la r to that o f B S F C and misfire. Figures 4.9 and 4.10 reveal no clear difference between P S C and standard igni t ion, and there is poor correlat ion for data that falls at misfire rates o f less than 3 0 % . 46 140 1 120 100 80 y 60 40 -20 -0 O t H C - D I Without P S C O t H C - D I Wi th P S C X 1 Linear ( A l l 0° Data) y = 1.9x + 12.4, R 2 = 0.92 Linear (<30% Misfire) y = 1.5x + 17.1, R 2 = 0.66 10 20 30 Misfire (%) 40 50 60 Figure 4.9 t H C w i t h M i s f i r e (0° Injector A n g l e ) 14U 120 100 80 u 60 40 20 0 O t H C - D I Without P S C O t H C - D I W i t h P S C Linear ( A l l 8° Data) y = 1.6x + 16.6, R 2 = 0.81 •Linear (<30% Misf i re) y = 1.5x + 17.9, R 2 = 0.56 0 0 X 10 15 20 25 Misf i re (%) 30 35 Figure 4.10 t H C w i t h M i s f i r e ( 8 ° Injector A n g l e ) However , B S F C and total hydrocarbons ( t H C ) exhibi t a strong correlat ion. Th i s is plotted in Figures 4.11 and 4.12 for each t im ing sweep. The data is plotted without differentiation between P S C and standard igni t ion , since no significant difference was found between these when B S F C and t H C where examined separately. 1200 1000 ^ 800 H 600 u fcu m 400 200 -I B S F C Linear y = 6.62x + 126 R 2 = 0.99 20 40 60 80 t H C (g/kWh) 100 120 Figure 4.11 B S F C w i t h t H C (0° Injector A n g l e ) 48 1200 200 -0 -j , , , , 1 1 0 20 40 60 80 100 120 t H C (g/kWh) Figure 4.12 B S F C w i t h t H C (8° Injector A n g l e ) The correlat ion between B S F C and t H C indicates that the D I engine 's eff iciency was strongly influenced by misfire and incomplete combust ion. The lack o f correlat ion between misfire and B S F C or t H C indicates that even when fuel was successfully ignited, the combust ion quali ty was poor. This suggests problems such as a lack o f flame front penetration into over ly lean or r i ch areas, or isolat ion o f ignitable regions from active combust ion chemistry. In this case the mixture was not ful ly stratified, but characterized by differing levels o f heterogeneity; however the combust ion problems encountered are also c o m m o n in stratified charge engines. W h e n compared to conventional spark igni t ion , P S C d id not exhibi t a significant effect on these problems. 49 4.2 PILOT CHARGE STUDY 4.2.1 Results In the presence o f a weak homogenous mixture (k ~\J, see Table 3-1 for fuel l ing details), late direct injection, combined w i t h P S C , produced reliable combust ion. B S F C is plotted at several direct injection t imings i n F igure 4.13. 5 10 15 20 25 D I SOI T iming Relative to Igntion ( C A D After Igntion) Figure 4.13 P i l o t Charge Study F u e l E f f i c i ency Results 30 Perhaps as significant as the B S F C result, is that no misfire was observed as l o n g as P S C was i n use. The histogram o f net heat release for a l l three condit ions (Figure 4.14) reveals a tightly distributed peak w i t h an average o f 1370 k J / m 3 . 50 70 3 | M 0 f 60 4-20 + 50 4-10 + 0 0 160 320 480 640 800 960 1120 1280 1440 Integrated Heat Release k J / m 3 Figure 4.14 H i s tog ram o f Ne t Heat Release for a l l pi lot charge study condit ions A s w i l l be discussed further, the background mixture was too lean to support combust ion on its own . Consequent ly, it was expected that D I injected after igni t ion w o u l d not burn. Indeed, non-P S C trials at the operating condit ions o f Figure 4.14 produced only misf i re , therefore data was not collected. 4.2.2 Discussion B y compar ison w i t h motored engine data, it was clear that the P S C system, operating i n the absence o f any other fuel source, fai led to produce evidence o f combust ion. Figure 4.15 shows that the two averaged pressure traces are identical , though the instrument is coarse at this scale. 51 17 1* 1 6 ¥ 3 OO. U fc 15 14 -I O Motored O 2 . 6 g / h P S C / o o o o o o V C ^ Representative Error Bar -O o 0 o o o -15 0 15 Crank Postion ( C A D A T D C ) Figure 4.15 I n -Cy l inde r Pressure Trace Compar i son : M o t o r e d and w i t h P S C The composite heat release profiles i n Figure 4.16 are s imi lar ; both ind ica t ing the same loss to heat transfer through the combust ion chamber wal ls . Figure 4.16 N e t Heat Release Rate Compar i son : M o t o r e d and w i t h P S C A l t h o u g h the same P S C settings produced a significant result in the lean burn case o f section 2.3, there is no evidence to suggest that the P S C pilot charge is ign i t ing when surrounded by air. G i v e n this result, it is not surpris ing that P S C failed to produce a significant result when compared to convent ional spark igni t ion during the direct injection t i m i n g sweeps. W h e n P S C is injected into pure air, it is l i ke ly that the concentration gradients are in i t i a l ly quite high, since the overa l l transit ion must span the concentration range between zero and 100% natural gas. W h e n P S C is injected into a dilute air-fuel mixture , the gradients are l i ke ly to be less severe since the overa l l concentration range is narrower. In both cases, a complex and highly transient vo lume o f combust ible chemistry w i l l result, however this region w i l l be larger in the case o f a weak ly fuelled bu lk mixture, and more l ike ly to bridge between the ign i t ion source and combustible mixture . W h e n a very weak (X, ~ 1.7), pre-mixed bulk charge was inducted, no heat was released from spark igni t ion alone. Y e t evidence o f combust ion was detected when the P S C charge was added. This is evident i n Figures 4.17 and 4.18, w h i c h show the pressure traces and heat release profiles from the two cases. 53 -15 0 15 Crank Position ( C A D A T D C ) Figure 4.17 In-Cylinder Pressure Trace Comparison: With and Without PSC O 3 Qi (U <Z) CvS <D O Pi • s X X 1.69 Without P S C / \ X.1.69 Wi th 28 g/h P S C •. . . . . 'UAA^JIHS ft.i Figure -60 -40 -20 0 20 40 60 80 100 120 Crank Position ( C A D A T D C ) 4.18 Net Heat Release Rate Comparison: Weak Mixture With and Without PSC It is not clear to what extent the heat released in the PSC case was from pilot fuel or the surrounding bulk charge; it is likely a mixture of both. It appears, however, that this initial combustion does not propagate through the ultra-lean mixture. This inference is supported by the 54 high exhaust levels o f unburned hydrocarbons and the discrepancy between the heat evo lved from combust ion and the fuel energy entering the combust ion chamber (Table 4-1). Th i s suggests the presence o f a l imi ted combust ion event that extinguishes or continues to burn without propagating through the mixture . T a b l e 4-1 Part ial Combus t ion : W e a k M i x t u r e w i t h P S C Tota l Hydrocarbons > 10,000 ppm-wet N e t Integrated Heat Release 41 (J /cycle) Heat Cont r ibu t ion o f P S C 23 (J /cycle) Heat Cont r ibu t ion o f P remixed F u e l 555 (J /cycle) Tota l Inducted Heat Energy 578 (J/cycle) The ini t ial combust ion produced when P S C is combined wi th a dilute mixture seems able to act as a pi lot charge, p rov id ing reliable igni t ion to the directly injected fuel . Inspection o f the B S F C data, shown this t ime in compar i son to the t im ing sweeps (Figure 4.19), indicates that reliable combustion was occur r ing at direct injection t imings that were far later than was possible dur ing the t iming sweeps. 55 900 600 —' u CQ 300 0 O 0° T iming Sweep O 8° T iming Sweep A Pilot Charge Study -90 -80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 D I SOI T iming Relative to Ignition ( C A After Ignition) Figure 4.19 B S F C Compar i son : P i lo t Charge Study W i t h T i m i n g Sweeps In t iming sweep cases that exhibi t s imi lar B S F C to the pi lot charge study, a m i x i n g interval o f at least 70 C A D was required before igni t ion . Th i s was fo l lowed by the beginning o f heat release approximately 19 C A D later. Table 4-2 shows average values o f ign i t ion delay for certain operating condit ions selected from Figure 4.19. A l l those from the pi lot charge study are included, as w e l l as t i m i n g sweep condit ions w i t h a B S F C o f 350 g / k W h or less. In the case o f the pi lot charge study, substantial heat release lags direct injection by a s imi la r interval as igni t ion delay i n the t im ing sweeps. In both cases, this interval is smal l compared to the timescales required for adequate m i x i n g i n early D I . The prompt arr ival o f heat release i n p i lo t charge study cases indicates that the late D I charge was burning i n a h ighly stratified manner. 56 T a b l e 4-2 Ave rage B S F C and Ignit ion De lay T i m i n g Sweeps P i lo t Charge Study B S F C ( g / k W h ) 0-10% net Heat Release Dura t ion ( C A D after D I ) 0-10%) net Heat Release Dura t ion ( C A D after Ignit ion) The stable combust ion and absence o f misfire found i n the pi lot charge study part ial ly validates the concept o f c o m b i n i n g stratified charge wi th pi lot igni t ion . Howeve r , g iven the over ly r i ch global mixture inherent i n this approach, h igh levels o f unburned hydrocarbons and relat ively poor fuel efficiency are inevitable. Therefore this data does not support conclusions on the effectiveness o f P S C ign i t ion to improve fuel usage i n a stratified charge engine. 4.3 SUMMARY The experimental evidence that P S C doesn't produce an ignitable p i lo t charge when injected into air provides an explanat ion for the lack o f combust ion at late direct inject ion t imings . Wi thout a pi lot charge, ign i t ion w o u l d depend on spark igni t ion alone. Spark discharge produces a very smal l ion ized vo lume between the electrodes, w h i c h in this case, were s l ight ly recessed into the cyl inder head. Y e t this energized region w o u l d need to intersect the fuel je t ' s narrow and h ighly transient margin o f combust ib le mixture at the correct t ime. Cons ide r ing the var iabi l i ty o f the je t ' s surface, and the size o f the affected vo lume produced by spark discharge, it seems un l ike ly that combust ion w o u l d occur re l iably from spark discharge alone. In the early D I t im ing sweeps, the P S C charge was injected into an air-fuel mixture composed o f va ry ing levels o f heterogeneity. Th i s is a substantially different condi t ion than injecting P S C into air, so it is not exper imental ly clear whether or not P S C produced a p i lo t charge under these 325 22 325 19 57 conditions; however the lack of differentiation between PSC and unaided spark ignition suggest that it did not. 58 5. CONCLUSIONS AND RECOMMENDATIONS The Par t ia l ly Stratified Charge concept has previously been shown to improve combust ion ini t iat ion and stabili ty i n a lean fuelled, homogeneous charge engine. It was hypothesised that P S C w o u l d also improve fuel usage i n a directly injected, stratified charge engine. The specific objective o f this research was to modify a research engine to support spark igni ted combust ion o f a directly injected stratified charge. It was expected that, i f successfully implemented, this combust ion system w o u l d produce h igh levels o f unburned hydrocarbons, as is characteristic o f external ign i t ion source stratified charge engines. Th i s is largely due to inconsistent igni t ion o f the fuel charge, poor combust ion propagation and ove r -mix ing at the fuel charge boundaries. It was hypothesised that the h igh igni t ion energy offered by the P S C system w o u l d result i n reduced hydrocarbon emissions. 5.1 CONCLUSIONS A single cy l inder research engine, previously equipped w i t h the P S C system, was modi f ied to accommodate a natural gas direct injector. A series o f direct inject ion t im ing sweeps were conducted to characterize the response o f fuel efficiency to the t ime interval a l lowed for fuel-air m i x i n g prior to igni t ion . F o l l o w i n g the t im ing sweeps, several trials were conducted i n w h i c h direct injection and P S C were combined in the presence o f a very dilute homogenous bulk charge. The fo l l owing conclusions are extracted from the resulting data: 1. The experimental engine configuration was not able to support stratified operation through late direct injection. The best B S F C was observed at the most advanced injection t imings , in w h i c h the interval available for the fuel charge and surrounding air to m i x was m a x i m i z e d . Ef f ic iency degraded rapidly w h e n this interval was reduced be low a m i n i m u m threshold. 59 2. The minimum mixing interval was influenced by the angle at which the fuel jet entered the combustion chamber. The onset of severe performance degradation in the eight degree case occurred at a mixing interval approximately 10 crank angle degrees longer than when the jet was aimed directly at the spark plug electrodes. A 16° case was attempted, but failed to produce reliable combustion even at the most advanced timings. 3. The PSC system exhibited no significant improvement in terms of BSFC, when compared to unaided spark ignition. Data collected during the early direct injection timing sweeps revealed that PSC ignition offered no systematic advantage in terms of BSFC, misfire rates, or the unburned hydrocarbon content of exhaust. 4. The PSC system did not produce an ignitable pilot charge when, prior to PSC injection, the engine's combustion chamber contained only air. Currently, the PSC system has been shown to produce combustion only when injected into a dilute homogenous mixture. 5. When the presence of an ignitable pilot charge was assured, the jet-guided stratified charge, created by late direct injection, ignited reliably and produced stable engine operation. 5.2 RECOMMENDATIONS FOR FUTURE WORK The reliable stratified charge combustion, produced when the presence of an ignitable pilot charge was assured, supports the conceptual validity of combining late direct injection with pilot ignition. Future work on this concept is warranted; and efforts should focus on the reliable establishment of an ignitable pilot charge, such that it can subsequently ignite a directly injected bulk charge. To this end, the following work is recommended in support of further developing the PSC concept: 60 1. The P S C fuel jet development is currently not w e l l understood. It is not k n o w n h o w vigorous ly it mixes w i t h the surrounding mixture, nor is it clear to what extent combust ion chamber bulk mot ion carries it away from the electrodes. A better understanding o f the P S C system's in-situ f lu id dynamics may expla in the pilot charge's failure to ignite when injected into air; and may produce recommendations for remedying this. 2. The effectiveness o f P S C appears to be h ighly sensitive to geometric aspects o f the fuel path through the spark p lug body. A new P S C spark p l u g was bui l t for this study. It had seemingly minor variat ion, a imed at improv ing its reach into the combust ion chamber, as compared to a p lug recently used by Reynolds (2004). Th i s new p lug was in i t ia l ly instal led dur ing the lean burn ver i f icat ion study previously discussed i n 2.3. N o n e o f the benefits that P S C had exhibi ted when previously appl ied to the lean burn case were apparent when the new p lug was i n use. It wasn' t unt i l the or ig ina l p lug was replaced, that P S C showed the posi t ive result reported in 2.3. The previous ly val idated p lug was used throughout a l l subsequent experiments reported in this thesis; the reasons for its superior performance are not understood and warrant further investigation. 3. The P S C system w o u l d benefit from a more precise means o f meter ing fuel f low. A s mentioned i n section 2.3, the current system's vaguely defined end o f injection may result in a resumption o f fuel f low as combust ion chamber pressure drops dur ing the b low-down. Th i s w o u l d cause increased levels o f unburned hydrocarbons and reduce fuel eff iciency. U s e o f a purpose-built gaseous fuel injector may offer a more clearly defined end o f injection, as w e l l as an overal l improvement in t i m i n g promptness and metering precis ion. 61 4. Shou ld efforts to improve the igni tabi l i ty o f the P S C pi lot charge prove successful, late D I combined w i t h P S C should be further explored. Th i s should include variations in the D I strategy such as differ ing injector angles, mul t ip le or if ice injection and mult iple injection pulses. 62 REFERENCES Abata D (1987). A review of the stratified charge engine concept. In R. L. Evans (Ed.) "Automotive Engine Alternatives" New York, New York: Plenum Press, pp.37-82, Arcoumanis C, Hull D R, Whitelaw J H (1994). An Approach to Charge Stratification in Lean-Burn, Spark-Ignition Engines. SAE Technical Paper Series. 941878. Bowker A H, Lieberman G J (1972). "Engineering Statistics" 2nd Ed. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. pp 296. ISBN: 0-13-279455-1 BP p.I.c. (2005). "BP Statistical Review of World Energy - June 2005" (electronic resource URL:http://www.bp.com/liveassets/bp internet/globalbp/globalbp uk english/reports and publi cations/statistical_energy_review_2006/STAGING/local_assets/downloads/pdf/statistical_review _of_world_energy_ful l_report_2005.pdf Brown G (2003). "Performance of a Partially Stratified-Charge Gasoline Engine" University of British Columbia, Canada, M.A.Sc dissertation. Dale J D, Checkel M D, Smy P R (1997). Application of High Energy Ignition Systems to Engines. Progress in Energy and Combustion Science, Vol. 23. pp. 379-398, 1997 Etrema Products Inc. (2006). "Terfenol-D Data Sheet" (electronic resource URL: http://www.etrema-usa.com/products/terfenolA) Webmaster: Global Reach Internet Productions, Accessible as of March 2007. Evans R L (2000). "Control Method for Spark-Ignition Engines" United States Patent No: 6,032,640, Issued March 7, 2000 Ferguson C R (1986). "Internal Combustion Engines - Applied Thermosciences" 2n d Ed. USA: John Wiley & Sons, Inc. ISBN: 978-0-471-35617-2 Fraily M , Norton P, Clark N N and Lyons D W (2000). An Evaluation of Natural Gas versus Diesel in Medium-Duty Buses. SAE Technical Paper Series. 200-01-2822. Germane G J, Wood C G, Hess C C (1983). Lean combustion in spark-ignited internal combustion engines - a review. SAE Technical Paper Series. 831694. Goto Y, Sato Y (2001). Combustion improvement and exhaust emissions characteristics in a direct injection natural gas engine by throttling and exhaust gas recirculation. From "Direct Injection SI Engine Technology 2001" (SP-1584) Warrendale, Pennsylvania, USA: SAE International. ISBN 0-7680-0731 -3 Goudie D, Dunn M , Munshi S R, Lyford-Pike E, Wright J, Duggal, V, Frailey M (2004). Development of a Compression Ignition Heavy Duty Pilot-Ignited Natural Gas Fuelled Engine for Low Nox Emissions. SAE Technical Paper Series. 2004-01-2954 63 Green R K, Zavier C C (1992). Charge Stratification in a Spark Ignition Engine Proc. Instn Mech Engrs, Vol. 206 Part A, pp 59-64. Gupta M , Bell S R (1994). An Investigation of Lean Combustion in a Natural Gas Fuelled Spark Ignited Engine. ASME - Natural Gas and Alternative Fuels for Engines ICE-Vol. 21 1993. Heywood J B (1988). "Internal Combustion Engine Fundamentals" New York: McGraw-Hill. ISBN 0-07-100499-8. Hill P G, Ouellette P (1999). Transient Turbulent Gaseous Fuel Jets for Diesel Engines. Journal of Fluids Engineering, Vol. 121, pp 93-101 Holman J P (2000). "Experimental Methods for Engineers" 7th Ed. New York: McGraw-Hill, Inc. ISBN 0073660558 Huang Z, Shiga S, Ueda T, Nakamura H, Ishima T, Obokata T, Tsue M , Kono, M (2002). Combustion Characteristics of Natural-Gas Direct-Injection timings, Proc. Instn Mech. Engrs, Vol. 217 Part D, pp 393-401 Iaconis J-L (2003). "An Investigation of Methane Autoignition Behaviour Under Diesel Engine-Relevant Conditions" University of British Columbia, Canada, M.A.Sc dissertation. Kubesh J T (2002). Development of a Throtteless Natural Gas Engine - Final Report. NREL/SR-540-31141, National Renewable Energy Laboratory. Lahbabi F Z, Boree J, Nuglisch H J, and Charnay G (1993). Analysis of Starting and Steady Turbulent Jets by Imaging Processing. Techniques Proceedings of the 1993 ASME Winter Annual Meeting, ASME-FED-Vol. 172, pp. 315-321. Reynolds C (2001). "Performance of a Partially Stratified-Charge Natural Gas Engine" University of British Columbia, Canada, M.A.Sc dissertation. Reynolds C, Evans R L (2004) The Low NOX Potential of Partially Stratified-Charge Combustion in a Natural Gas Engine. Combustion Institute / Canadian Section, Spring Technical Meeting, Kingston, Ontario, Canada, May 9-12, 2004 Ricardo p.l.c. (2005). Hydra web-page (electronic resource URL: http://www.ricardo.com/engineeringservices/engine.aspx?page=hydra) accessible as of March 2007 Rood E P, Demetri P T (1991). Journal of Fluids Engineering Policy on Reporting Uncertainties in Experimental Measurements & Results. Journal of Fluids Engineering, September 1991 Editorial on Experimental Uncertainty. Society of Automotive Engineers. (1999). Stoichiometric Air-Fuel Ratios of Automotive Fuels -SAE J1829 Dec97. from "SAE Handbook". Warrendale, PA. U.S.A.: Society of Automotive Engineers, Inc. 64 Turner J W G, Pearson R J, Kenchington S A (2004). Concepts for Improved Fuel Economy from Gasoline Engines. International Journal of Engine Research, IMechE 2005, vol. 6, DOI: 10.1243/146808705X7419, pp 137-157, 2005 U.S. Department of Energy (a) - Energy Efficiency and Renewable Resources (2005). "Properties of Fuels" (electronic resource URL: http://www.eere.energy.gov/afdc/pdfs/fueltable.pdf) U.S. Department of Energy (b) - Energy Efficiency and Renewable Resources (2006). "Clean Cities Alternative Fuel Price Report - October 2006" (electronic resource URL: http://www.eere.energy.gov/afdc/resources/pricereport/pdfs/afpr_oct_06.pdf), Contact: Michael D. Laughlin, New West Technologies, L L C , 4351 Garden City Drive, Suite 600, Landover, MD 20785. Accessible as of March 2007. White F M (1994). "Fluid Mechanics" 3 r d Ed . McGraw-Hill, ch 9, ISBN 0-07-113765-3. Wilcox D C (2003). "Basic Fluid Mechanics" 2n d Ed. California, USA: DCW Industries, Inc. pp 631-639, ISBN 1-928729-03-7 Williams D (2006). B.A.Sc University of Toronto, 2003. M.A.Sc Candidate, U B C 2003-2006. personal communication. Zhao F, Harrington D L , Chi M (2002). "Automotive Gasoline Direct-Injection Engines" Warrendale, Pennsylvania, U S A : S A E International. ISBN: 0-7680-0882-4 65 A P P E N D I X A : D E S I G N D O C U M E N T A T I O N DIHEAD MODIFICATION: GEOMETRIC REPRESENTATION The solid model pictured in Figure A . l below was constructed in Pro/Engineer version 2000i2 This model served as the basis for the design study in which room for the J43 fuel injector was found within the head's internal geometry. Information was gathered from the original design drawings, as well as measurements taken from the sand casting patterns. Figure A . l Aluminium SI Cylinder Head Solid Model 66 DI HEAD MODIFICA TION: STA TIC STRESS ANAL YSIS Figure A.2 below shows the output of a geometric element analysis (GEA) performed using Pro/Mechanica version 2000i2. The purpose of this analysis was to estimate the stresses in the fire-deck due to the axial pre-load necessary to seal the injector penetration against combustion pressure. Pictured below is the copper combustion seal (see Figure A.3, sheets 3 and 6) transferring load to the aluminium fire-deck. The pre-load (see Appendix B) is applied normally to the upward facing surface of the combustion seal. The model includes the combustion seal and fire deck assembled as separate pieces. The clipping shown below is for presentation purposes and was not present in the geometry analysed. The fire deck material, 365 aluminium, has a yield stress of approximately 165 MPa. The results were deemed acceptable, as the majority of stresses fall below 130MPa. The apparent high stress "hot-spots", if not due to the analysis technique, could be expected to yield slightly until a new lower stress equilibrium was reached. Figure A.2 GEA Analysis of Stress Due to Combustion Seal Pre-Load 67 DI HEAD MODIFICA TION: FABRICA TION PROCESS DRA WINGS (Begin next page) NOTES RICARDO RESEARCH ENGINE ALUMINUM HEAD DIRECT INJECTION MODIFICATIONS GENERAL ASSEMBLY . FOR QUESTIONS AND ADDITIONAL INFORMATION: PtEA.SE CONTACT DAVID SOHBT AT UBC ISO*) B72-0I9I, dg.rbyftneek.ubf.ca . FABRICATION INCLUDES TRIAL RUN ON SPARE HEAD (TO BE INSPECTED WHEN COMPLETE). AS * £ L L AS FINAL MODIFICATION OF OPERATIONAL READ (ALUMINIUM HEADS PROVIOED BY UBC) C<!§X MECHANICAL ENG)NEERIHG DEPARTMENT THE UNIVERSITY OF BRITISH COLUMBIA HEAD MODIFICATION - PROCESS DRAWING GENERAL ASSEKBLT NOTES I. VERIFT AS-BUILT DIMENSION FOR USE WITH ADDED MATERIAL PART (SEE SHEET 2) • I ; I JCIIIIDEI IftAD IMKWIMB Bl OTHCK31i J l»|"JJ i ITCH : an. < Kictificn mrtmi I B ILL Of MATER IAL | HECHAHICAL ENGINEERING DEPARTMENT I THC UNIVERSITY OF BRITISH COLUMBIA HEAD BREAK-IN CUT DETAIL ADDED MATERIAL ALUMINUM 606 I I REQUIRED «EV ISIOMS NOT E S INJECTOR MOUNTING FLANGE 1018 STEEL I REQUIRED REVISION-I NOTFS: COMBUSTION SEAL HEIGHT REDUCED 0IT.I •25 \ 0 ! .2S ' J- 4 -r^ TbTtM'TL'TK I—^ V T T t S t / COMBUSTION SEAL CI 1000 COPPER I REQUIRED I iCONBUSTIOt S(»l i CWPER I i tNJECTDt (MUMTING FUKE STECL I fADOEG NATtlUL PART - ALUMINUM IT. • OEKIIHIW lUIElUL BILL OF MATERIAL MECHANICAL ENGINEERING DEPARTMENT THE UNIVERSITY OF BRITISH COLUMBIA HEVISIONS WELD SHALL FORM A WATER-TIGHT SEAL AROUND ADDED MATERIAL COMPONENT B I L L OF MATERIAL HECHANICAL ENGINEERING DEPARTMErli THE UNIVERSITY OF B R I T I S H COLUMBIA HEAD MODi AT'O HEAD/ADDED MATERIAL WELD DETAIL wev isiONS N O T E S i . b a s i c d i m e n s i o n s a r e f o r p r a c t i c e head o n l t . f o r f i n a l f a b r i c a t i o n , v e r i f y d i m e n s i o n s t o match e x i s t i n g h o l e p a t t e r n W I R E I N S E R T I S NOT N E C E S S A R Y FOR P R A C T I C E H E A D ; D R I L L 08 THRU O N L Y FOR F I N A L F A B R I C A T I O N . WIRE I N S E R T W I L L R E Q U I R E S H O R T E N I N G F R O M 8 TO - 5 R E V I S I Q N -1 n o t e s : A M B I G U O U S A N G U L A R D I M E N S I O N S R E M O V E D F R O M I N J E C T O R P E N E T R A T I O N A X I S . U S E B A S I C D I M E N S I O N S O F T R U E - L E N G T H V I E W ( S H E E T 6 ) , A N D P O I N T - 0 A S R E F E R E N C E . ,0* ^ j i I i inan.H MM-COIL till ustn ;M • n i l MrtmtL• j B I L L OF MATER I AL MECHAN JCAL £NGIN[[RING OEPARFMENT | ! T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A IHEAD MODIFICAT I ON - PROCESS DRAW I N G iPENETRATJON DET At L. - ORT HC^RjAPH IC V JEWS | H£*D.BI1("K.I« I " * **" "" 04/U/It PISTON CROWN DETAILS 3,188-£,£93 r 0.407 Dimensions in Inches Figure A.4 Geometry: Piston # 476P by Federal Mogul for Ford Fiesta (1978-80) APPENDIX B: SAMPLE CALCULATIONS COMBUSTION SEAL PRE-LOAD The following MathCad worksheet calculates the pre-load required to seal the direct injector's fire-deck penetration against combustion pressures. The pre-load is calculated at standard temperature conditions and accounts for expected thermal expansion as the engine is brought to operating temperature. The tensile side of the load path includes a substantial aluminium component, whereas all other components are steel; consequently the "cold" condition represents the maximum stress, due to the de-tensioning effect that aluminium's greater thermal expansion plays in the load path when the engine is warm. Pmax := 6000000Pa Dia := 9.2mm Est := 200- 109-Pa A r e a : = — D i a . £ £ A O m - 5 2 4 Area = 6.648 x 10 m Faxial := Pmax-Area Angle from Datum A: 9 := 63deg Faxial = 398.857 N Thermal elongation of head/bolt combination M6xlbolts a s t . = 1 1 7 ' 1 0 AT := 100-K L c M 6 := 93mm K 8thM6 := as t -ATLcM6 8thM6= 1.088 x 10" 4 m a A l := 2 4 1 0 — L c A l := 102-mm K 8thAl := a A l A T L c A l 8thAl= 2.448 x 1 0 _ 4 m Continued... 76 Total thermal elongation of head/bolt combination: Sth := 8thM6 + SthAl 8th = 3.536 x 10" 4 m Thermal elongation of Injector LcJ43 := 183-mm 5thJ43 := ocst-ATLcJ43 8thJ43 = 2.141 x 10" 4 m Relative elongation of head/bolt combo: 8r:= 8 t h - 8thJ43 8r= 1.395 x 1 0 " 4 m Bolt Pre-load to take up relative elongation: _ Est a := or-L c M 6 a = 3 x 10 8 Pa PSC MAXIMUM BACK PRESSURE FOR CHOKE FLOW The following is a MathCad program which accepts reservoir pressure as input and calculates the maximum back pressure that will allow frictionally choked flow at the end of the PSC flow path. Reservoir Pressure (Po)Po := I5bar k := 1.32 To := 300K D := .0225-in L := 20cm f := .028 bar := 100000-Pa speed := 2000- —— min e := 0.0015mm angVel := speed-360 deg ^2 2 A:= n Rm := 518 4 s 2 K pol := Po Rm-To angVel = 209.44 Hz duration := 50-deg ung:=1.34.10- 5-^ u := m2 P°l Reo: V o D Reo= 4.487 x 10 4 u = 1.388 x 10 6 — Vo := 109 — s D 2.625 x 10 s - 3 f-L D 9.799 Given Maguess := 0.1 2 2 f-L 1 - Maguess k+ 1 (k + l)-lvlaguess s s + In • D 2 2-k 2 k-Maguess 2 + (k - 1 )• Maguess durationtime:= duration angVel durationtime = 4.167 x 10 s Mai := MinErr(Maguess) = (0.242) Given To Tl guess := 500-K = i + h L . M a i 2 Tlguess 2 Tl :~ MinErr(Tlguess) = (297.213K) al := (k-Rm-Tl) 0.5 al •= 450.802-Tl := T l Tl = 297.213 K VI := Mala ! VI = 109.131 — s Continued... 78 p i guess:= 1-—-m Given 1 7 0 1 = 1 +-.(k- l ) M a l 2 p i guess 2 1 ITT p i := MinErr(plguess) = 9.375 ^ -m pi = 9.375 S^-m" mdot:= p i - A - V I mdot = 2.624 x 1 0 ~ 4 - ^ s - 6 minj := mdot-durationtime minj = 1.094 x 10 kg speed . . avgmdot := ——minj 2 P I • - — avgmdot = 6 5 . 6 1 2 — (k-1) h r 1 + - ( k - l ) M a l 2 2 PI = 1.443 x 10 6 Pa PI Pstar := k + 1 M a i , .. . . . . ,2 2 + (k - l ) M a l Pstar = 3.259 bar P1 = 14.434 bar Po = 15 bar k 1 2 ( k - 1 } 5 Postar := Pstar- 1 + —(k - 1)-Mal Postar = 3.387 x 10 2 Prjequired := Prjequired = 1.845 2 < k" l> k+ 1 Min_outlet_i3ressure_tbr_choke Postar Prjequired Min_outlet_pressure_for_choke = 1.836 bar DIMASS FLOW AND ORIFICE SIZE CALCULATION The following is a MathCad program which accepts reservoir pressure as input and calculates the mass flux through the DI orifice for a given injection interval. This was used iteratively to solve for the orifice size. Initial •Assumptions Vs - -163.30111 kp=3 ;s:=;2GO0 3:; Vc: 56.13cm Pi:=0:97-aun ev 0 73 AFRst:^ 16.69^ To:=310K R:=8.31451< pule. •Mfu8ls=-ra-38-mokss. air st: mole mole- K. M a i r - 137.3552 kng := 1.32! faim 1 A. niojg Rmetlsine :=|5183 kg-K molar AFRsIs moles_fuel_sl := moles air st Mair moles ^ ftielLslJ mqlar_.AFR:- mnlflr_AFRst»X Mole fraction fuel moles .02 . - molr AKR .mcmJIZi* molar_AFR-3 76 total moles := moles 02 + moles N 2 4 1 xfiiel ;=: Mfuel molar tA:FRsi.=j2.:U2 moles air st = 121.5 lmol; moles fuel st = 57.537mol /total_moles: xftiel - 0.032 xfuel-Pi-Vs nfuel :=;.ev.--To-R nfucl = 4.139* 10 mol Nfuel .-nfuel;— 2' in fuel avg — NfuelMiuet mfueljwg - 0.432— hr Pcrit:. 2.5x lO^Pa minPcritQV erPo := minP.6s;6mc :='• Pcrit mihPcritbvcrPo1 minPosonic = 4.611 x 10 Pa min_sonic_PR : Pi := 5 Po: 3000090psi 1 minPcritova-Po; mill sotric PR = 1.845 From, Hill and Oulett (1999) losses in injector correspontto 1 0 % stag pressure loss Po v - 7.222/. 1 • R-T6 3 po i-y-Mfuel' po - 125.52-m 1 3 m. p e n t r-Vcritr-2-kng-RineUiaiio- To kng 4 1 Continued: 80 Iterates to find orifice size based on desired mass flux over specified injection interval Cycte:=360 Tfuel:=293-K Period:= - Period = 30x ]0" 3 s ATcrank_angle := ATcrank Jingle = 83.333x 10 6 s s Cycle lnj_duration := 13.7 fucl_delivery _t ime Jfract ion := AT_inj:= Inj_durationATcrank_angle AT_inj = 1.142x 10 s AT_inj mfuel inst := Period-2 mfuel_avg fuel_delivery_time_fractton = 0.019 mfuel inst = 6.301-^-fuel_deHveryjime_fraaion sec mfuel_injectioi. := mfuel_instAT_inj mfueHnjectibn = 7.194mg Acrit: mfueLinst-Y1 (RmethaneTfuelj >/kng-kng + 1 0 khgr-l kng-1 •Po Acrit = 0.197mm Dinj: 4-Acrit Dinj= 0.5 mm 81 APPENDIX C: ENGINE OPERATING PROCEDURES Ricardo Hydra Engine (NG) Operating Procedure Alternative Fuels Laboratory, UBC Mechanical Engineering The following is the start-up procedure for the Ricardo Hydro single-cylinder research engine in the U B C Alternative Fuels Laboratory. Please follow these steps carefully to avoid equipment damage or injury to personnel. Note: Steps preceded by the DI superscript are particular to the Natural Gas Direct Injection System. If the injector is not installed these steps are omitted. If the DI injector is installed, these steps must be followed regardless of whether or not DI will be used in current trials. Initial checks and engine warm-up: 1. Turn on ventilation in test cell (main fan - switch outside door - and fume hood). 2. Turn on emissions bench to allow sufficient warm-up (see separate procedure). 3. Check engine oil and coolant levels, and check around the engine for leaks. 4. Check that guards are on flywheel and timing belt. 5. Crank engine by hand once or twice (to ensure there has been no leak into the combustion chamber). 6. Check that there is no condensation in the exhaust (briefly open valves: engine out, muffler drain and horizontal run). Ensure all drains are closed before engine start. 7. Turn on engine cooling water (open tap fully). 8. Start pressure transducer cooling pump. 9. Turn on ignition/injection driver box (test-cell - by DAQ). 10. If the natural gas direct injector will be used, plug in the power supply first, then plug injector driver into power supply. (Not mandatory if injector will not be used) 11. Turn on thyristor drive unit in the test cell. (Main breaker on cell wall should be left on). 12. Turn on oil and water pumps/heaters. 13. Ensure that back pressure valve control box is powered and that valve is plugged in. 14. While oil and water are heating, calibrate emissions bench (see separate procedure). 82 Starting the engine software: 15. Turn on the control room.computer (password = "Ricardo"). 16. Open Dynoclient and select file to save data to (in D:\ drive). The file name should have the following format: <date>_<user>_<speed>_<throttle%>_<test description>.csv, for example: "031121_CR_2000rpm_100%_homogeneous_lean_limit_noPSC.csv". (Note: If not acquiring data, open "junkdynodata.csv".) 17. Open Pressureclient. 18. Open VNC„ (password = "Riccardo"). Run Timing Controller. 19. Open/run Ric_Emissions_Runtime for monitoring, calculated values. (Start Menu, "Programs", "Windows NT Explorer", D:\...) Starting the Engine: 20. Reset emergency stop buttons (one on post by engine, one on control panel). 21. Open the two green NG valves in test cell, and the NG valve in the control room. 22. 0 1 Turn on high-pressure supply at the test cell outer wall gas panel, and ensure that the supply pressure is between 2000 and 3000 psi - do not operate this panel without prior instruction. 23. 0 1 Ensure that in-cell high-pressure supply valve is fully open, and vent valve on test cell outer wall is fully closed. 24. Turn on ignition switch (control panel). 25. Check oil temperature is greater than 60 °C 26. Check (control panel): • Speed setting is at 3.0 (1500 rpm) • Throttle setting is at 5 0 % • Fuel setting is at 2.2 27. Make a note of the following in the Ricardo Logbook: • Date • Engine hours • Speed and Load • Operator Comment (test description) and initials 28. Press "reset" then immediately press green "start" button. (D l listen for pneumatic high-pressure supply cut-off valve to energize upon pressing "reset". Do not proceed if valve does not energize.) 83 Firing the Engine: (Engine firing is always started at 1500rpm, 50% Throttle, Lambda 1.0 and MBT Spark) 29. In. Timing Controller set spark timing to -23 deg, ("duration" should be set to 1.0 deg). Turn on ignition. 30. Check fuel control is set to "flow" (switch 1) then turn on (switch 2). Engine should fire. (Note that fuel flow output on DAQ should be 0.8 kg/h - this corresponds to lambda 1.0). 31. After one minute of firing, turn on AFRecorder (lambda sensor): • Press "sys", "6" (says 'enable'), "ENT", "1" (says 'measure'). 32. Wait until oil temperature is approx. 90 °C and coolant temperature is approx 95 °C before beginning to acquire data. Important Note: If testing for a long time, may need to recalibrate the emissions bench, as the calibration of the instruments drifts with changing temperature. Recommended Practice: Before proceeding with the test, run the engine at the above "standard" conditions for at least 20 mins, and acquire data at this point (for repeatability). The emissions bench can be calibrated during this time. Acquiring Data: 33. Set test point. Note that when running very lean, the lambda based on mass flows is more accurate than lambda from the AFRecorder, (i.e. use Ric_Emissions_Runtime by pressing "Send data to Excel" button in Dynoclient). 34. In the paper "Test Sheet", record details of test point for future reference. 35. Record Pressure data (must be done for every test point). The file name should have the following format: <date>_<tesW>_<speed>_<throttle%>_<spark>_<Lambda>_<PSCdetails>_pr.csv, for example: "031121_t05_2000rpm_50%_S23 _L1.0_noPSC_pr.csv". 36. Record Performance data: In Dynoclient, update: • Test number • Spark timing • PSC details (if running with PSC) Then click "Log data to File" and wait (approx. 2 mins.) 37. Set next test point and repeat data acquisition. 84 38. When testing is finished, return the engine to the start-up settings (1500rpm, 50% throttle, lambda 1.0). Shutting down the engine: 39. Turn off the NG supply at the flow-meter. 40. Turn off ignition (Timing Controller on computer. 41. Allow engine to motor briefly, until exhaust temperature drops below 130°C. 42. Press red "dyno" stop button. Engine will come to a complete stop. 43. Turn off ignition switch on control panel 44. Turn off oil/water heaters (but not the pumps). 45. About two minutes after firing stops, disable AFRecorder. • Press "sys", "6" (says 'enable'), "ENT", "1" (says 'measure'). • Display should now read 'Sensor Disabled' 46. Close all three NG valves on NG lines. 47. D l Turn off high-pressure supply at the test cell outer wall gas panel. 48. Turn off switch on ignition/injection box in test cell. 49. If plugged in, unplug natural gas direct injector power supply and driver box (order doesn't matter). 50. When engine has cooled (after at least fifteen minutes): • Turn off oil and water pumps. • Turn off the thyristor drive breaker. • Turn off cooling water tap and unplug pressure transducer cooling pump • Shut down emissions bench. • Stop all software and shut down control room computer. 51. When testing is complete and engine shut down, stop the Dynoclient - this closes the file you have been saving performance data to. (if you wish to continue monitoring engine conditions during the cool-down stage, start Dynoclient again and use "junkdynodata.csv"). 52. Comment on any important issues in the Ricardo Logbook, record engine hours and running time. 85 EMERGENCY SHUT DOWN: Hit red "STOP" button in Test cell or on Control Panel. Ignition and dyno are disabled and the engine will come to a complete stop. Note that fuel does not shut off automatically, nor are there gas detectors in the cell connected to the ESD. (These issues are addressed in the new CERC cells). Do not use the emergency stop button unless necessary. 86 A P P E N D I X D : F U E L C O M P O S I T I O N Table D - l BC Natural Gas Composition Compound Mole % in Fuel Moles in Fuel Molecular Mass of Component (kg/kmol) Lower Heating Value of Component (kJ/kg) Methane (CH4) 91.43 0.914326657 16.043 50030 Ethane (C2H6) 6.60 0.066023333 30.070 47511 Propane (C3H8) 0.32 0.003216667 44.097 46333 i-Butane(C4H10) 0.07 0.000703333 58.123 45560 n-Butane (C4H10) 0.05 0.00047 58.123 45719 i-Pentane (C5H12) 0.02 0.00022 72.150 45249 n-Pentane (C5H12) 0.02 0.00016 72.150 45345 neo-Pentane 0 72.150 45052 Hexane (C6H14) 0.01 0.00005 86.177 45103 Heptane (C7H16) 0.00 0 100.204 44921 Octane (C8H18) 0.00 0 114.231 44783 Carbon Dioxide (C02) 0.40 0.00397 44.010 0 Nitrogen (N2) 1.09 0.010926667 28.013 0 Based on BC Natural Gas as measured on Micro at GC Westport Innovations Inc, Oct, 2005 Table D-2 Mixture Properties Molecular Mass of Fuel (kg/kmol): 17.38 Hydrogen/Carbon Ratio (mol/mol): 3.839 Oxygen/Carbon Ratio (mol/mol): 0.007 Nitrogen/Carbon Ratio (mol/mol): 0.020 Upper Heat Value (kJ/kg): 53489 Lower Heating Value (kJ/kg): 48302 Simple Air/Fuel Ratio by mass (kg/kg): 16.50 SAE Air/Fuel Ratio by mass (kg/kg): 16.69 SAE Air/Fuel Ratio by mass H/C only (kg/kg): 17.06 SAE Air/Fuel ratios based on J1829 SAE recommended Practice 87 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0080782/manifest

Comment

Related Items