SAE TECHNICAL PAPER SERIESInternational Fall Fuels &Lubricants Meeting & ExpositionSan Antonio, TexasOctober 14-17, 1996
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`The appearance of the ISSN code at the bottom of this page indicates SAE's consentthat copies of the paper may be made for personal or internal use of specific clients.This consent is given on the condition however, that the copier pay a $7.00 per articlecopy fee through the Copyright Clearance Center, Inc. Operations Center, 222Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sections107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds ofcopying such as copying for general distribution, for advertising or promotionalpurposes, for creating new collective works, or for resale.SAE routinely stocks printed papers for a period of three years following date ofpublication. Direct your orders to SAE Customer Sales and Satisfaction Department.Quantity reprint rates can be obtained from the Customer Sales and SatisfactionDepartment.To request permission to reprint a technical paper or permission to use copyrightedSAE publications in other works, contact the SAE Publications Group.GLOBAL MOBILITY DATABASE All SAE papers, standards, and selectedbooks are abstracted and indexed in theGlobal Mobility Database.No part of this publication may by reproduced in any form, in an electronic retrievalsystem or otherwise, without the prior written permission of the publisher.ISSN0148-7191Copyright 1996 Society of Automotive Engineers, Inc.Positions and opinions advanced in this paper are those of the author(s) and notnecessarily those of SAE. The author is solely responsible for the content of the paper.A process is available by which discussions will be printed with the paper if it ispublished in SAE Transactions. For permission to publish this paper in full or in part,contact the SAE Publications Group.Persons wishing to submit papers to be considered for presentation or publicationthrough SAE should send the manuscript or a 300 word abstract of a proposedmanuscript to: Secretary, Engineering Meetings Board, SAE.Printed in USA 96-0049
`
`FORD Ex. 1051, page 2
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`1962018Understanding the Thermodynamics ofDirect Injection Spark Ignition (DISI)Combustion Systems: An Analyticaland Experimental InvestigationW. Anderson, J. Yang, D. D. Brehob, J. K. Vallance,and R. M. WhiteakerFord Motor Co.ABSTRACTDirect-injection spark-ignition (DISI) engines havebeen investigated for many years but only recentlyhave shown promise as a next generation gasolineengine technology. Much of this new enthusiasm isdue to advances in the fuel injection system, which isnow capable of producing a well-controlled spray withsmall droplets. A physical understanding of newcombustion systems utilizing this technology is justbeginning to occur. This analytical and experimentalinvestigation with a research single-cylindercombustion system shows the benefits of in-cylindergasoline injection versus injection of fuel into theintake port. Charge cooling with direct injection isshown to improve volumetric efficiency and reduce themixture temperature at the time of ignition allowingoperation with a higher compression ratio whichimproves the thermodynamic cycle efficiency.Volumetric efficiency improved by 2-3%,wide-open-throttle (WOT) output improved by 5-10%and part-load fuel consumption improved by 4-5%versus a production PI engine with stoichiometriccalibration. A further gain of 6-7% was recorded withlean, homogeneous air-fuel operation.INTRODUCTIONFour-stroke direct-injection spark-ignition (DISI)combustion systems have been pursued for severaldecades as a fuel efficient alternative to premixedcharge engines [1, 2]. To maximize fuel efficiency,most of these engines were designed for unthrottledoperation. As such, load control was achieved byvarying the fuel quantity. At light loads, localstratification of the fuelair mixture was necessary inorder for the combustible mixture to remain withinflammability limits. A homogeneous mixture wasdesirable at high loads in order to maximize airutilization. A high compression ratio, for greaterthermal efficiency, was present in most of the designs.Knocking tendency was suppressed with high flowmotion, late injection and a compact chamber.Two general types of stratified-charge engineswere investigated: those employing low pressureinjection systems (~2 MPa) with early compressionstroke injection, such as the Ford PROCO [3-5], andhigh pressure, diesel-type systems such as theTexaco Controlled Combustion System (TCCS) [6, 7].Although the fuel consumption of these engines wasquite good, hydrocarbon emission control wasextremely difficult at light-load operation. Because oflean operation, nitrogen oxide emissions werecontrolled with large amounts of EGR (40-50%). TheNOx emission level was still too high for today’sstandards, however, which would necessitate a leanexhaust after-treatment system. The output of thelate-injection diesel-type systems was also sootlimited. Further studies of stratified-charge engines [8,9] showed that high swirl was desirable for goodfuel-air mixing and that large, high-energy, ignitionsources helped to improve multifuel capability. In spiteof measures to reduce knock and maximize airutilization, these concept engines suffered from lowoutput compared to contemporary SI engines.Recently there has been a resurgence of effort ondirect-injection (DI) engines [10-15] which has beendriven by worldwide energy usage and emissionsconcerns. Results to date have been moreencouraging as a consequence of improvements infuel systems and in tools for the understanding ofcombustion systems. While most all of the systemsare designed for stratifiedcharge operation at lightload, fuel system improvements have made itmore possible to obtain good air utilization atwide-open-throttle (WOT) conditions. In fact, animprovement of 5-10% has been reported [14, 15] forWOT output versus an equivalent port injected (PI)engine. Such claims are believed to be partially due tosuperior fuel-air mixing [16] of the fuel systems.Emission capability is still a concern, however.While one design is scheduled for production [15],capability to meet ULEV and Stage III standards of theUS and Europe has yet to be demonstrated. Thereremain difficulties with controlling hydrocarbonemissions under
`
`FORD Ex. 1051, page 3
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`2light-load conditions with stratified-charge designs.Nitrogen oxide emissions require substantial amountsof ECR (~40%) along with exhaust after treatment tohave promise of meeting these standards.Nevertheless, much progress has been made;research is continuing [17] and new understandingswill continue to be made in advancing the future ofthese engines.In this paper, a design approach, analyses andresults are presented for a research combustionsystem that was designed to meet future customerperformance and fuel economy and emissionsrequirements. The analyses help to explain the fueleconomy and performance benefits of the system andin particular, the effect of charge Cooling which resultsfrom the new fuel systems. The data show promise forsignificant cold-start hydrocarbon emission reductionwithout substantive increases in either hydrocarbonsor nitrogen oxide emissions during steady, warmed-upoperation.COMBUSTION SYSTEM DESIGNThe design of the research combustion systemwas first approached by developing some systemrequirements which are shown in Table 1. It wasdesired to have comparable power and torque to thebaseline system, which in this case was a 4-valveengine of comparable displacement per cylinder. Peakspeed capability, which was an issue with earlierstratified-charge approaches, was desired to becompetitive with the base system. The remainingrequirements for emissions and fuel consumptionwere considered essential for future markets for whichthe system might be considered.After deciding on requirements, a number ofdesirable system attributes for meeting therequirements were developed. These are shown in theright column of Table 1. The challenge was to,incorporate as many of these attributes as possibleinto the actual design while taking consideration ofavailable fuel systems.Some of the system attributes are complementarywhile others are not. For example, the desirability toachieve high air utilization implies sufficient fuel-airmixing to obtain near homogeneous operation. On theother hand, achieving lowest fuel consumptionsuggests one might want the maximum amount ofstratification possible at light loads to potentiallyeliminate pumping losses. This has been achievedwith a recent design [15] for a different emissionrequirement. The emission requirement for thisapproach, however, places a much greater burden onsystem performance. As a consequence, it wasdecided to pursue a homogeneous-charge approachwhich could be run lean if adequate aftertreatmenttechnology became available. By taking thisapproach, other system attributes such as light-weightpiston, low S/V chamber and improved knocktolerance could be more easily incorporated in thedesign.The research combustion system was designedwith a CADDS-4X software package. This allowednumerical machining of flow boxes for portdevelopment. Their development took placeconcurrently with the cylinder head. Base designspecifications for the single-cylinder research engineare shown in Table 2.To obtain good air utilization, it was desirable toposition the fuel injector near the center of thechamber such that the spray would distribute itselfmore evenly in the cylinderical geometry. As recentlydiscussed [14], this location also directs the spray atthe piston which, if impacted by the spray, would beleast likely to significantly impact hydrocarbonemissions. To minimize flame travel distance andthereby potentially improve knock limited output, thespark plug was also positioned in a near centrallocation as shown in Figure 1. As the schematicshows, this approach places a restriction on themaximum valve sizes that can be implemented.To overcome the potential flow loss due to smallervalve area, much effort was expended on portdevelopment. The results of the effort are shown inFigure 2 which shows flow coefficient versus non-dimensional valve lift. The capability of the basecomparator port was measurably improved with thisdesign. The in-cylinder flow of the design wasmeasured with a water analog rig. With both valvesoperational, there was no net swirl with only a smalltumble component. With one valve deactivated, thenon-dimensional swirl number was 2.4Table 1: Combustion system design goals.Combustion systemrequirement:System attribute Peak torque and power same as base PIHigh efficiency portsIncreased compression ratioHigh air utlization and combustion efficiencyImproved knock toleranceHigh speed (> 6000 RPM)operationLight-weight pistonHigh air utilization and combustion efficiencyFast fuel-air mixingLow HC/NOx emissions(ULEV/Stage III capable)Close-loop, stoichiometric operationThree-way catalystExcellent fuel management duringtransientsLow fuel consumptionLow S/V chamberHigh dilution tolerance (EGR/lean)High compression ratioSwirl/tumble for fuel-air mixingTable 2: Engine description.Engine type4-stroke, single-cylinder,DOHC 4-valveDisplacement (cc)575Bore/stroke (mm)90.2/90.0Compression ratio11.5Combustion chamber shapePentroofExhaust/intake valve area ratio0.77
`
`FORD Ex. 1051, page 4
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`m
`
`m
`
`m. Figure 3 shows the calculated single-dropletlife time as a function of air temperature and dropletsize based on Spalding’s theory [18]. Vaporization ofan 80
`m
`m droplet takes tens of milliseconds,corresponding to hundreds of crank-angle degrees atan engine speed of 1500 rpm. Vaporization of a 20
`
`3and the tumble number was 0.1. Single-cylinderengine data were collected with both flowconfigurations.EXPERIMENTThe test facility consisted of a specially-fabricated4-valve single-cylinder head mounted on a RicardoHydra research engine base. The engine was coupledto a DC dynamometer. A Concurrent computer wasused for: engine control, collection of low-speedsignals such as temperatures and emissions, andcollection of high-speed signals such as those fromthe cylinder pressure and fast flame ionizationdetector. A Zexel high-pressure fuel-injection systemconsiting of an injector, an injector driver, ahigh-pressure pump, and a pressure regulator systemwere used for DI operation. The intake manifold wasdesigned for PI, with injectors of 5.5kg/hr flowrateeach mounted in both runners. A plate blocking oneintake port could be installed to generate swirl flow.PHYSICS OF DIRECT INJECTIONGasoline fuel injection systems have advancedsignificantly in the past ten years. Using a swirl injectorwith high fuel pressure (50-100 bar), the Sauter meandiameter of droplets in a fuel spray can be as small as15
`mdroplet, in contrast, takes only several milliseconds,corresponding to tens of crank-angle degrees. Fastvaporization of the small droplets helps to make thedirect fuel injection concept feasible.Direct fuel injection into the cylinder producesmany advantages over PI engines, including highertorque and lower knocking tendency. Both result as aconsequence of charge cooling by the fuel. For PIengines, a fuel spray with large droplets is injectedinto the intake port. Intake port wall and intake valvewetting by the fuel commonly occurs and the liquidfuel is vaporized mainly by absorbing thermal energyfrom these surfaces. For a DI engine, in contrast, fuelis injected into the cylinder as small droplets, whichare vaporized mainly by absorbing thermal energyfrom the air. The charge temperature is, therefore,lower which results in a higher volumetric efficiencyand a lower knocking tendency.To estimate the effect of charge cooling on thevolumetric efficiency and the charge temperature, twoextreme cases are studied. In the first case, fuel isvaporized by heat transfer from intake port and valvesurfaces only,Figure 1: Schematic layout of 4-cylinder head with nearcentrally located fuel injector and spark plug.Figure 2: Improvement of flow coefficient for DISI versusbase intake ports.Figure 3: Effect of droplet size and air temperature ondrop lifetime.
`
`FORD Ex. 1051, page 5
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`4presumably closer to what occurs in PI engines. In theother extreme, fuel is vaporized by absorbing thermalenergy from the air only, presumably closer to whatoccurs in DI engines. The latent enthalpy ofvaporization, L, can be calculated by [19]where LTbn is th latent enthalpy of vaporization at Tbn,the normal boiling temperature, and Ter is the criticaltemperature of the fuel. If the fuel is vaporized byabsorbing energy from the air only, the airtemperature at a constant pressure will drop byto a lower temperature of Ta When the air is mixedwith the fuel vapor, the mixture temperature becomesHence, the total volume of air and fuel vapor can be estimated byFigure 4 shows the decrease in charge volume due tocharge cooling as a function of the intake air and fueltemperatures. When the initial intake air temperatureis 100°C and the fuel temperature is 50°C, the volumeof the mixture, after vaporization, is about 5% smallerthan the volume of the intake air. Under the samecondition, however, if fuel is vaporized and heated tothe air temperature by heat transfer from the wall, themixture volume will be larger than the volume of theintake air by 2% due to the volume of fuel vapor. Thus,the total different in mixture volume of the two extremecases can be as large as 7%.The two extreme cases, however, are not realistic.Some fuel in the PI engine can be vaporized byabsorbing thermal energy from intake air, while somefuel in the DI engine can be vaporized by the pistonand the cylinder walls. Therefore, the actual differencein engine volumetric efficiency between PI and DIengines is smaller than that between the two extremecases, depending on the engine design and the fuelspray characteristics. Data from a test engine indicatethat it can be about one-third of the maximumdifference at the extremes, as shown later.At a constant pressure, the difference in thecalculated charge temperature at the two extremecases can be as large as about 30°C, depending onintake air and fuel temperatures. When these mixturesare compressed in an engine cylinder, thetemperature difference is amplified. Figure 5 showsthe temperature difference of themixtures at the two extremes after compression as afunction of the intake air and fuel temperature. Thecalcuated temperature difference is more than 5°C,consistent with simulation results using a computercode, GESIM [22]. As indicated before, however, theextreme cases are unrealistic. The actual chargetemperature difference near TC between PI and DIengines is smaller than the predicted extremes.Because of the lower charge temperature, the DIengine has lower knocking tendency. Figure 6 showsa comparison between two pressure traces measuredin a single-cylinder engine when fuel is injected in the,port and in the cylinder. When the fuel is injected inthe cylinder, the knock-limited spark timing wasadvanced byFigure 4: Effect of evaporation heat source and initial airand fuel temperatures on final mixture volume.Figure 5: Comparison of compressed mixturetemperatures for extreme cases of all fuel vaporizedin mixture and all fuel vaporized on intake andcombustion chamber surfaces.
`
`FORD Ex. 1051, page 6
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`h
`h
`c, mechanicalefficiency,
`
`m, air-fuel ratio, AF, fuel lower heatingvalue, QLHV, and the intake air condition, Pa/R Ta, aremore likely independent of the injection method.Engine volumetric efficiency,
`
`h
`h
`v, and the indicatedthermal efficiency,
`
`510.7 CAD from that with PI in the same engine.Therefore, a DI combustion system has the potentialto have a higher compression ratio to achieve a higherthermal efficiency.To show the effect of fuel injection and evaporationon the in-cylinder processes, a section of themeasured pressure traces during the compressionprocess, shown in Figure 6, are blown-up and shownin Figure 7. Before fuel injection, the slope of thepressure trace for the DI case is higher than that forthe PI case due to differences in the fluid properties.After the start of injection, however, the slope of thepressure trace for the DI case decreases. Thedecrease of slope not only results from the change offluid properties, but also results from the cooling effectof the evaporated fuel since the slope becomes evensmaller than that of the PI case. The cooling by thefuel evaporation results in lower fuel-air mixturetemperature near TDC, thereby reducing the engine’sknocking tendency.The assumption that the fuel of the DISI engine isevaporated mainly by absorbing thermal energy fromthe air has been verified by other studies ofexperimentation [20] and 3-D numerical simulation[21]. Both studies show that only a small portion of thefuel impinges on the piston and bore walls of the DISIengine for normal injection timing. There is pistonwetting, however, when the injection timing is veryearly in the induction stroke.WOT RESULTSEngine torque output or engine brake meaneffective pressure (BMEP) can be expressed asThe engine combustion efficiency,
`i, however, can be improved as aconsequence of charge cooling due to direct fuelinjection.Figure 8 shows the effect of injection timing of DIon engine volumetric efficiency compared to PI. Themeasured volumetric efficiency of the DI testsincreases by about 2.5% over that of the PI tests,which is about 1/3 of the maximum difference betweenthe two extreme cases shown in Figure 4. Thevolumetric efficiency of the DI tests depends oninjection timing. When fuel injection starts at thebeginning of the intake process, the piston is locatedat the highest position and piston wall wetting by thefuel spray occurs [20]. A portion of the fuel vaporizesat the hot wall, eliminating much of the charge-coolingbenefit. Retarding the injection timing reduces wallwetting [20], resulting in an increase in volumetricefficiency. When the injection timing is too late,however, the fuel droplets have insufficient time tovaporize before the intake valve closes. Thus, thevolumetric efficiency starts to decrease when theinjection timing is further retarded. The peak of themeasured volumetric efficiency occurs when SOI is at-240 CAD.When the operating condition is such that the sparktiming is knock limited, the engine’s thermal efficiencyrelies strongly on the spark timing. Figure 9 shows theeffect of direct injection timing on the knock-limitedspark advance (KLSA) compared to PI. Unlike thevolumetric efficiency, shown in Figure 8, the KLSAcontinues to increase as injection timing is retarded.This is because the effect of charge cooling is notrestricted by the intake valve timing. Retarding theinjection timing not only reduces wall wetting by thefuel spray, it alsoFigure 6: Comparison of measured cylinder pressuretraces from a single-cylinder engine operated with PIand DI.Figure 7: Enlargement of Figure 6 pressure traces.
`
`FORD Ex. 1051, page 7
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`6reduces charge heating by the walls, resulting in alower mixture temperature near TDC. Retarding theinjection timing, however, also reduces the availablemixing time, resulting in higher cycle-by-cyclevariation.Figure 10 shows the effect of injection timing of DItests on IMEP and a comparison of DI versus PI. Thebaseline IMEP of PI operation is lower than that ofproduction engines since the single-cylinder engine’scompression ratio of 11.5:1 causes the PI operation tobe highly knock limited. Because the IMEP of the DItests depends mainly on the product of the volumetricefficiency and the thermal efficiency, which relies onKLSA, the IMEP curve also has a shape similar to thevolumetric efficiency curve in Figure 8. The peak ofthe IMEP curve, however, occurs at a later timing thanthat of the volumetric efficiency curve due to thecontinuous increase in the KLSA with retarding SOI.At a later injection timing of -150 CAD, however, theIMEP decreases quickly due to poor mixing asmanifested by large cycle-by-cycle variation.PART-LOAD RESULTSThe results of air-fuel ratio sweeps in the DI engineare shown in Figure 11. The NOx emissions of theearly injected DI nearly match PI operation in thesame cylinder head. Both manifest the expectedpremixed NOx results, i.e., increase in NOx as leanedfrom stoichiometric. NOx levels drop as the mixture isleaned beyond 16:1 or 17:1 air-fuel ratio. In contrast,the late injected DISI NOx reveals a high degree ofunmixedness. The late injected DISI engine at 14.6:1overall air-fuel ratio produces substantially lower NOxthan either of the mostly premixed cases indicatingthat a great fraction of the fuel combusts at an air-fuelratio that is locally richer than stoichiometric.Enleanment from stoichiometric results inmonotonically increasing NOx emissions over therange of air-fuel ratio tested in the present study. Thisindicates that a substantial fraction of the fuel iscombusting at local air-fuel ratios which produce highlevels of NOx, i.e., higher than the overall air-fuel ratiowould suggest.The hydrocarbon results in Figure 11 show thatDISI suffers from a slight penalty compared to PI forall the air-fuel ratios. The late injection DISI emitssimilar hydrocarbon levels to the other two casestested at the richer ratios. However, combustionbecomes unstable at ratios leaner than about 20:1resulting in high hydrocarbons. The present DISIcombustion system was designedFigure 8: Increase in volumetric efficiency of DISIengine over PI measured in a single-cylinder engine.Results averaged over four timing scans, two ofwhich had both ports operational and the others hadone port deactivated.Figure 9: Comparison of KLSA between PI and DI atdifferent SOLFigure 10: Percent increase of IMEP of DISI engine overPI operation.
`
`FORD Ex. 1051, page 8
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`7for three-way catalyst operation, i.e., stoichiometricmixtures. The resulting open chamber facilitates rapidmixing and permits the flame to travel readilythroughout. The design was not optimized forstratified-charge operation. The intent of the presentair-fuel ratio investigation is to evaluate how thedegree of mixedness is affected by injection timing.The DISI combustion system with early injectionhas a slight fuel consumption advantage over PI in thesame cylinder head at all air-fuel ratios tested, asshown in Figure 11. The fuel economy benefit of leanoperation is 12% for early injected DI over thestoichiometric PI and about 11% over stoichiometricDI.NOx and fuel consumption data for the DISIcombustion system are compared to the baseproduction engine as a function of EGR in Figure 12;also shown are the lean air-fuel ratio data of Figure11. One can observe that the NOx is lower for the DISIengine with stoichiometric operation and that the fuelconsumption benefit versus the base productionengine with EGR is on the order of 5% at therespective minimum fuel consumption points. Leanoperation is seen to offer a further benefit asdiscussed earlier.Fuel economy and emission comparisons to thebase 4-valve PI production engine are shown inFigure 13. At stoichiometric operation, DI provides a4-5% fuel economy benefit over a production PIengine depending on whether the comparison is madeat best EGR or withoutFigure 11: Fuel consumption, hydrocarbon emissionand NOx emission for early and late direct injectionand PI in the research single-cylinder engine.Figure 12: ISFC vs ISNOx for base engine and DISIcombustion system.Figure 13: Part-load results of port and direct fuelinjection in the DISI combustion system compared tothe base, production engine.
`
`FORD Ex. 1051, page 9
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`8EGR. The DISI combustion system exhibits a penaltyin hydrocarbons as both the port and direct injectionsystems are higher than the production engine. This isdue partially to the high compression ratio whichcauses more unburned gases to be compressed intocrevices and less post oxidation in the exhaust port.The additional increase in hydrocarbons of DI over PIin the identical combustion system is a subject forfurther study. The transient fuel managementopportunities with DI are expected to more than offsetthe penalty in hydrocarbons which may exist atsteady, warmed-up operation. The NOx emissionsfrom the DISI combustion system are reducedcompared to the conventional PI engine as aconsequence of a different residual content.About 10% less fuel is consumed at idle by theDISI combustion system compared to theconventional PI engine while maintaining goodstability. The comparator engine is a multi-cylinder PIengine for which statistics for all 8 cylinders areavailable. The DI single-cylinder engine had a 35%greater COV (coefficient of variation) of IMEPcompared to the best cylinder of the comparator and60% lower than the worst cylinder.A thermodynamic engine model [22] was employedto confirm the fuel economy findings and to investigatethe sources for the DI advantage versus theproduction PI engine. The simulations were performedfor a constant level of residual gas. A 4% fueleconomy benefit and 4.5% NOx penalty are predicted,as shown in Figure 14, for the simulated compressionratio increase of 1.5. If only the compression ratiowere increased, the resulting NOx penalty is greaterthan the 4.5%.The thermodynamic engine model was used tocompare part-load fuel economy at the two extremecases of fuel vaporization in the port and fuelvaporization in the combustion gases. Identical inputconditions were employed with the only differencebeing that the fuel was considered fully vaporized atinduction in the PI case; in the DI case, the fuel wasinducted as a liquid and instantaneously vaporized atthe time of intake valve closing. A thermodynamicbreakdown of the distribution of the fuel energy for thetwo cases is shown in Table 3. The latent enthalpy ofvaporization absorbs 0.7% of the fuel’s chemicalenergy with DI. However, by cooling the combustiongases, less energy is transferred to combustionsurfaces and the energy exiting with the exhaustgases is decreased, a total of 1.0%. The predictedbrake thermal efficiency increases from 26.1% to26.4%, which is a 1.1% improvement. modelling of DIindicates a thermodynamic benefit associated with in-cylinder injection. This 1.1% improvement summedwith the 4% benefit due to the 1.5 unit increase incompression ratio which the model predicts comparesfavorably with the 5% fuel benefit foundexperimentally.The 1.1 % thermodynamic efficiency improvementofTable 3: Breakdown of fuel energy consumption for enginesimulations for two extreme cases: all fuel vaporized inport and all fuel vaporized in the combustion gases.DI over P1 which the model predicts for the samecombustion system has been observedexperimentally. In Figures 11 and 13, back-to-back DIand PI experiments in the identical cylinder head showabout a 1% fuel consumption benefit for DI. Such asmall fuel consumptionFigure 14: Cycle simulations of the effect ofcompression ratio on light-load (1500 rpm, 3.8 barIMEP) fuel economy for the DI engine (residual massfraction maintained constant).Categories of fuelenergy usageVaporization atport surfacesVaporization incombustion gasesBrake workMotoring frictionPumpingHeat loss to coolantHeat loss to oilExhaust gas enthalpyLatent enthalpy ofVaporization26.1%6.3%6.2%40.6%4.9%15.9%0.0%26.4%6.3%6.2%39.8%4.8%15.8%0.7%Totals:100%100%
`
`FORD Ex. 1051, page 10
` IPR2019-01402
`
`

`

`Downloaded from SAE International by Elizabeth Powell-Whyte, Tuesday, May 07, 2019
`
`9benefit is of the order of the accuracy of theexperimental measurements. However, the fact thatall part load data show a small DI fuel consumptionbenefit and the model indicates a similar level isstrong evidence that the difference exists.SUMMARY AND CONCLUSIONSAn approach for the design and development of aresearch single-cylinder direct-injection, spark-ignitioncombustion system is presented. Requirements arepresented for a system that conforms to futureperformance, fuel economy, and emissionsexpectations. The design of the 4-valve cylinder headis then presented and shown how it conceptuallymeets the desired system attributes.The physical consequences of DI are shown toresult in a lower in-cylinder charge temperature andan increase in the volumetric efficiency of thecombustion system. The reduced charge temperatureis shown to present an opportunity to increase thecompression ratio and achieve a higher thermalefficiency versus a PI engine. Data and calculationsare presented which show an improvement involumetric efficiency of 2-3%, a reduction in chargetemperature of up to 50-60°C, and an improvement inwide-open-throttle output of 5-10% for DI versus PIcombustion systems. In addition, predictions from athermodynamic model show that charge cooling ofdirect injection can lower the brake specific fuelconsumption at part load operation on the order of 1%,in addition to a decrease due to an increasedcompression ratio.The steady-state operation of the researchcombustion system is shown to result inimprovements in fuel consumption of up to 5% at partload and 10% at idle with stoichiometric operation.Best lean operation at part load is shown to offer apotential reduction of up to 11 % versus the base PIengine with EGR. Emissions are comparable to PIvalues on the same combustion system. An increaseof hydrocarbon emissions versus the baseconfiguration is mostly due to the increasedcompression ratio with DI. Nitrogen oxide emissionsare shown to be less than the base PI engine as aresult of charge cooling and the higher residualcontent of the system offsetting the effects of theincreased compression ratio.ACKNOWLEDGEMENTSThe authors recognize Mr. Paul Sleeman