Mixture Formution Process and Approackes
`eS
`
`patnarneter ts the {ight distance to impact, with a Seis
`ondary Gietor bemthe approaching angle, ‘This is made
`evident by simply rotating the impact plane around the
`pout oY intersection of the Sprayaxts, thus increasing
`she unpact distance on one side and decreasing the dis-
`race onthe other, Forall G-DI fucl sprays the inerease
`in wetted footprint area on the closer side is cvident.
`kiight distances maybe. for example, 35 mm on the
`gpriy ANis. ith 28 mmfor droplets on the near edge of
`the spray and 50 mmfor drops on the far edge. As may
`pe seen in Pig. 5.4-12. this change is more than suffi-
`cient to drasticallyalter the wetting pattern, even though
`4 haher velocity crossflowair does not. On the far side
`ofthe swirl spraythere is no longer any wetting, whoreas
`the near side experiences a much wider footprint.
`It
`should be noted that such footprints contain important
`information as to where wall wetting occurs or does nor
`occur. and as to the exact pattern, but do not provide
`quantitative formation asto the total fuel massorits
`distribution within the footprint.
`If 2% to 12% ofthe
`total injected fuel mass for a light-load operating point
`wers the impact surface and forms a film, then the mass
`affuelin the film will range from about0,25 to 1.3 mg.
`The total weight can be measured byutilizing a very
`low volatility fuel and weighing the detection paper
`using af analytical balance, but this can be done practi-
`callyonlyfor a bench room test, not for tests in an opti-
`cal spray chamberat elevated temperature and pressure,
`Laser interferometric measurements of the instantaneous
`film thickness would most likely have to be utilized in
`conjunction with the total footprint area to obtain an
`indication of the fuel mass that wets the impact surface
`ior a ¥iven combustion chamber configuration.
`Substantial] changes in the wetting pattern are seen
`for all but slit-type injectors when elevated opcrating
`iemperatures and back pressures arc used. The effect of
`‘he Operating environment on spray development was
`Uiscussed in detail in Chapter 4, but with regard to wall
`Welling it may be stated that the wetted footprints are
`vbsurved to change markedly with transitions of the
`erat characteristics of the spray. Thus the observed
`vollapse ofthe sprays for a multihole injector from six
`indy idual plumes 10 one is tracked by a change in the
`“elted foorprint from six individual spots to one central
`lanwey Spot, This may seem obvious with hindsight, but
`Certaintycould have been theorized that the small
`Subset of drops that wet the surface might have been
`ndcpendent ofthe collapse process. When the spray is
`vbserved tg collapse to. a new. narrower geometry, the
`
`wetted footprint is also found to coalesce into a morc
`central position. There is an cffect of the angle of
`approaching that should also be noted,
`If the impact
`distance is maintained constant at any value such as
`35 mm, and the angle ofapproach relative to the impact
`surface is varied.
`then some variation in wall wetting
`docs occur for G-DI sprays. However,
`it docs not
`appearto be correlated bya simple trigonometric con-
`version that adjusts the droplet velocityto its value nor-
`malto the impact plane. A very small localtarget placed
`at an angle within the G-DI spray will be wetted regard-
`less of the angle of inclination, whereasif it is part of a
`larger impact plane at the same location and angle it
`maynot be wetted,
`The intersection of a cone with an interposed plane
`is, Of course, a classical conic section. Because the
`majority of G-DI sprays are relatively axisymmetric and
`relatively conical, and because the majority of intended
`impact surfaces such as piston bowlfloors are relatively
`flat, the interpretation of spray footprints is often facili-
`tated byevaluating them as conic sections. This simple
`consideration can help to explam and correlate a num-
`ber of ostensibly complex patterns, even without requir-
`ing flow modeling. For most low to moderate engine
`loads and speeds the FPW isbrief (under 1.8 ms) and
`the piston motion during the impact event is slight
`(under 3 mm), thus the cone/plane geometryis almost
`Static, Cven in a running engine. Hence there is good
`agreement between footprint data from a pressure cham-
`ber test in the laboratoryand spray impact outlines on
`pistons taken from engines that have run on the dvna-
`mometer for many hours. [f the ambient density. injec-
`tor operating temperature, injector mounting angle and
`impact distance are emulated in a laboratory pressure
`chamber, the footprints so obtained will correspond to
`the outlines in the carbon on a removed piston. This
`gives confidence that some portion of G-D] spray: tar-
`geting studies related to bowl design. spraycone anvle
`selection, spray-bowl overshoot and smoke reduction
`can be conducted in a spray chamber, which is gener-
`ally quicker than dynamometer testing. The value of
`the wetted footprint is that it shows the precise location
`of the fuel that must undergo a transition from a liquid
`wall film toa vapor while confined to thal spot, whereas
`the remainder ofthe fuel in the sprayis notin a film and
`is not stauionary. Finaltests would still have to be per-
`formed on a running engine to verifythe system. but
`time muy be saved by screening out combinations
`that have obvious problems, such as excessive. broad
`
`FORD Ex. 1010, page 159
`IPR2021-00340
`
`FORD Ex. 1010, page 159
` IPR2021-00340
`
`

`

`Automotive Gasoline Direct-Injection Enyintes
`
`wetting or the overspraying of a target. Figure 5.4-13
`shows a comparison of LIF-mcasured fuel film and the
`carbon deposit pattern on the piston crown with a cen-
`trally: mounted 30°-cone-angle fuel injector 474), Car-
`bon deposits are observed on the piston top following
`late injection experiments. The deposit pattern shown
`in Fig. 5.4-13b bears a close rclationship to the shape
`and location offuel films as illustrated in Fig, 5.4-13a.
`The deposits accumulate rapidly and start to be notice-
`able onlyafter a fewhundred fired cycles. and are more
`pronounced for later injection timings. During initial
`stages the deposit appears to forman outline of the fuel
`film. as shownin Fig. 5.4-13b. With more fired cycles,
`the deposits fill
`in gradually and form a continuous
`laver that is similar to the pressure-chamber fuel film
`in shape and location. An image of fully developed
`
`carbon deposits on a piston crown is shownin Fig. § 4.
`{3¢. Clearly these carbonaceous deposits are accump.
`lated around the regions where substantial fuel films
`are formed due to spray impingement.
`The caveat regarding the matching of ambien,
`density and injector operating temperature is an impor.
`tant onc, Benchtesting at room conditions does indeed
`provide footprints, but, strictly speaking,these footprints
`only occur in the engine for cold crank and start. ang
`possiblynot even then because ofthe significant chanpe
`in injection timing that may be invoked.
`Injection ming
`changes can significantly alter the impactdistance, which
`ig a key parameter. Spray geometry and wetted footprint
`changesignificantly with operating conditionsforall but
`slit-type injectors, thus these conditions must be emu-
`lated if the footprint is to be representative.
`
`fuel film
`
`piston deposit
`
`(3}
`
`{b)
`
`
`
`Kigure 54-13 Comparison ofan LIF-measuredfuelfilm and the carbon deposit pattern on the
`piston crown with latefuel mjection timing: centrally mounted, swirl-typefuel injector. fa)fueifilm
`recorded at 65 crank angle degreesafier the SO] of-90 crank angle degrees on compression
`stroke; (bh) typical piston deposit pattern for late injection operation after approximately
`[00fired cycles, (€) piston-tap carhon depastts after several thousandfireel cvcles with
`fate injection (SOL
`-90 crank angle degrees)pe
`
`142
`
`FORD Ex. 1010, page 160
`IPR2021-00340
`
`FORD Ex. 1010, page 160
` IPR2021-00340
`
`

`

`At room conditions a swirl-type injector of spray
`‘one angle greater than about 45° will typicatly be a hot-
`'
`“joane SPTAY. andfor orthogonal injection will provide
`hos
`vpull’sexe pattem for the wetted footprint. This will
`ae of anouter annular nng that corresponds to the
`‘
`nannel of pintie-cxit angle. and a central spot that
`yy irl-cl
`i venerated by the sac spray.
`In fact, the fuel from every
`individual swirl-channel Is generally evident. This is
`clearly shown! in Fig. S.4-12a for a swirl injector,
`If the
`injector axis 1S inclined from being orthogonal to some
`other angle. such as 45° as shown in a perspective view
`iy Big, 34-126. the wetted footprint will be totally
`hanged. The pattem gencrally changes from a bull’s-
`ave toa crescent, with the central circular spot changing
`jo ancllipse. The ellipse is not centered, but will have
`whe myector axis at one ofthe foci. The spray-cone baund-
`qries do translate very well into elliptical boundaries on
`sninclined impactplane, as shown in the perspective view
`mfg. 34-14, Fora hollow-cone spray the wetted foot-
`print will be bounded by ellipses corresponding to the
`inner and outer cone angles of the spray. Even without
`measuring the actual spray it may be surmised that the
`spray outer cone is about 81 degrees, the inner cone
`ig about 60 degrees and the sac spray cone is about
`|? dearees. Itis helpful to consider that the wetted area
`ic q subset of the sprayintersection area of the impact
`plane, The spray must be present at a point in the plane
`i wet however, the spray may be present but not wet.
`The minor exception to this is the wetting that may occur
`‘ive to splashing from a liquid film that originates in the
`gray intersection area. Thus occurs in the footprint in
`rie 54-14.
`The very essence of G-DI spray-wall interaction
`sanplenityis illustrated by the footprints in Figs. 5.4-12
`
`
`
`ihme Std Spray ines, auter and sac cones derived
`rant the wetted fovipeint for an inelined inpection angle
`uf 45°for a YU? ywirl-type G-DE injector
`
`Mixture Formation Process and Approaches
`
`and 54-14. Even forthe verysimple casc ofa hollow-
`cone G-DI spray at room conditions, tilted from normal
`impact (90°) to 45° impact without changing the spray
`axis impact distance.
`the wetted footprint docs not
`extend continuouslyalong the space betweenthe cltipses.
`Even though the spray 1s everywhere within the annulus.
`the spray wets in someregions but reaches a point at which
`no wetting occurs. The presence of the spray maybe
`verified by passing a pulsed laserlight sheet through the
`spray at the cxact position where an impact plate would
`be. Thus, for significant areas of the spraythere 1s no
`interaction with the wall, and no wall-wetting. The drop-
`lets in the spray are obviously present at those locations
`(at least very close to the wall) but do not wet the surface
`It might be expected that the wetted footprint resulting
`from the spray from a hot injector with hot fuel would
`have a greatly reduced area, butthis is generally not the
`case. For swirl, multihole, atr-assist and slit-type G-DI
`injectors the total wetted area for hot operation is not
`muchless than is observed for the same impact geometry
`at room conditions. The biggest difference due to hot
`operationis illustrated in Fig. 5,4-12c, which shows that
`the spray collapses for someinjector types, and the spray
`footprint becomesa single centralized area. If the spray
`axis is inclined to the impact plane,
`the footprint will
`generallybe displaced toward the injector. as that side of
`the intersection ellipse will wet, while the other side may
`not wet atall.
`One ofmanynecessities ofmodeling the spray-wall
`interactions is to be able to predict the wetting patterns
`shown in Fig. 5.4-12. which is much moreinvolved than
`it may appear An accurate prediction ofboth the pattern
`and the distribution of fuel within the pattern must be
`made for any spray targeting geometry, various fuel
`properties, and ambient conditions. For 90°C operation
`the spray spatial distribution and charactenization will be
`completely different from that for 20°C operation,
`as wilt the rate of change of droplet diameter due to
`vaporization during the flight to the impact surface.
`This will alter the drag characteristics of the spray
`From Fig. 3.4-14 it mayalso be seenthat the footprint is
`broadened from the actual spray boundaries at the short-
`est impact distances. which is likely due to splashing
`from a thicker liquid film. For arbitrary points on an
`inclined impact plane, both the local fight distance and
`the compound angle ofapproach to the wall are changed.
`as are the local spraycharacteristics and the focal time of
`arrivalofthe first droplets.
`
`183
`
`FORD Ex. 1010, page 161
`IPR2021-00340
`
`FORD Ex. 1010, page 161
` IPR2021-00340
`
`

`

`Antamotive Gasoline Direct-Injection Engines
`
`
`
`The properties ofthe impact surface, such as the
`temperature, presence of
`roughness (imicrofinish).
`a pre-existing liquid film
`depasits or the presence of
`amount of wetting or the
`ean have a small effect on the
`ain influences ofthese
`wetted footprint. however, the m
`after impact or after a
`8
`variables are on what occurs just
`film is formed. Mic-seattered or volume-illuminated
`imaging observations show that the bulk spray motion
`far the transicnt imteraction with an interposed surface
`is little affected bythe above parameters. Except for a
`few percent higher spray-front velocity along the
`impact plate for a hot surface (T > 120°C),
`it
`is very
`difficult to detect differences in the wall-interaction
`images between a rough and smooth surfacc, or between
`a hat and cold surface. The overall bulk motions of the
`spray plume andits spatial position at a given time are
`
`basically unaffected. However. the microsceo
`Pic details
`of the outcome of the droplet impact events;
`an be Sig.
`nificantly influenced, Whether a droplet forms a fil
`enters a film, rebounds tntact, shatters and rebounds
`splashes within a film is not only determined by a ue
`SCTIe5
`of Weber number thresholds for the drop and the angle
`of approach, but also will be influenced bythe above
`surface propertics. Whether a droplet reaches an
`impact surface will not be strongly determined by the
`surface deposits or surface temperature, but the details
`of the interaction and the subsequent inclusion of the
`wall-interaction fuel mass in the combustion event will
`be significantly influenced. The three general regimes
`that contro] the overall effect of spray-wall interactions
`on G-DI combustion are summarized in Table §.4-2
`
`Tabfe 5,4-2
`Three general regimes controlling the overall effect of spray-wall interactions
`on G-DI combustion
`
`
`Aerodynamic
`Regime
`
`Fuel mass reaching potential impact surface is determined.
`Near-tip distributions of droplet size and velocity are important.
`Flight distances to impact and ambient conditions are key
`factors.
`Angle of inclination of target surface and the in-cyiinder air flow
`fieid are of less importance.
`
`Impact
`Regime
`
`Very small fraction of the total droplets in the spray reach the
`interposed surface and interact withit.
`These croplets can wet, rebound or splash, with the actual
`fraction being dependent not only on the droplet Weber number
`and angle of approach at the time of impact, but upon the
`specific wall surface properties and stale.
`Whether the wall is wet or dry, or the degree to whichit is
`smooth or rough, can significantly alter the outcome of the
`droplet interaction.
`
`Post-Injection
`Regime
`
`Between the end of injection and the combustion event.
`The small percentage of fuel mass that participates in the wall
`interaction is affected by the physical conditions within the
`combustion chamber.
`This mainly involves the vaporization of that small mass of fuel,
`consisting af both the vaporization of dropiets rebounded or
`splashed, as well as the evaporationof the liquid film remaining
`on the wall.
`This may be a total or partial vaporization prior to the spark, and
`will depend upon many factors such as the in-cylinder flow field
`near the film, the wall temperature, the fuel distillation curve and
`the presence and structure of surface deposits at thefilm
`location.
`
`
`FORD Ex. 1010, page 162
`IPR2021-00340
`
`FORD Ex. 1010, page 162
` IPR2021-00340
`
`

`

`The regimes discussed here are sequential. that is,
`alte post-imection regime is very dependent on the out-
`-ome of the impact regime, Jno droplets survive the
`jerody namic regime, with all droplets being
`entrained in the injection-gencrated flowfield and none
`reaching the wall.
`then the impact regime and post-
`jnjectiony regimes become moot. Tens ofparamcters and
`(yresholds are influential in the overall process that
`wrcarporatesall three regimes, whichis the reason why
`his is a Very difficult arca to correlate and model, Thesc
`thresholds are critical
`to spray-wall sub-models, and
`much research is being conducted to ascertain the lim-
`is of applying single-droplet data to transient G-DI
`sprays. Eyen within a single regime such as impact there
`is an ongoing debate regarding the proper Weber num-
`ber thresholds for splashing, rebounding and wetting,
`and as to whether existing data for PFI sprays or single
`droplets can be appropriately applied. It is obviousthat
`the mass, thickness, shape and time-history of any fuel
`wetted area in the engine, whether the wetting is
`intentional or unintentional, are parameters that can in~
`uence combustion and emissions.
`It
`is also evident
`that additional critical research needs to be conducted
`on G-DI] spray-wall interactions to more fully
`understand these complex processes and enhance the
`predictive capabilitics of CFD spray-wall interaction
`sub-models.
`
`5.5 Unintended Spray-Wall impingement
`
`5.5.1 Effect of Spray-Wall Impingement on
`Combustion and Emissions
`For a wall-guided G-DI combustion system fuel impinge-
`ment on a specially designed piston cavity is utilized to
`create a stable. stratified-charge mixture at light load.
`Other than the intended spray impingement that1s asso-
`ciated with guiding the spray, any unintended fuc] wall
`wetting should be cither climinated completely or mini-
`mized to the maximum extent possible ‘7°,
`It has been
`well established that matching and optimizing the spray
`cone angle and spray-tip penetration rate to the piston
`cavity geometry is one of the most important steps in
`minimizing unintended fuel impingementfor wall-guided
`G-D1 systems. With substantial unintended fuel impinge-
`ment on the combustion chamber surfaces, improved fuel
`atomization can onlypartially enhance mixture prepara-
`tion. Pool-buming ofthe resulting wall film will occur.
`along with the associated negative cffects of increased
`heat loss and elevated HC and soot emissions. Other
`negative impacts include the formation ofexcessive cham-
`ber deposits and engine oil dilution. The equivalence
`Tatio at compression TDC becomes leaner when wetting
`of the piston crown outside of the bow! occurs
`Figure 5.5-1 shows fuel film formation as a func-
`tion of fuel injection timing for a combustion svstem
`
`
`OOO
`
`
`SOI = -320 CAD
`
`SOl = -300 CAD
`
`SOl = -270 CAD
`
`SOl = -210 CAD
`
`SOl = -180 CAD
`
`SOl = -150 CAD
`
`SOl = -120 CAD
`
`SOI = -90 CAD
`
`Figure 35-1 Iuel film formation as afunction of fuel injection timing, centrally mrotwited, 30° swirl-
`ype dtectur, tmaves recorded at —30 crank ogle degrees relative 10 the compression TDC (4-4
`
`LAS
`
`FORD Ex. 1010, page 163
`IPR2021-00340
`
`FORD Ex. 1010, page 163
` IPR2021-00340
`
`

`

`Automotive Gasoline Direct-Injection Engines
`
`
`having a centrallymounted. 30° swarl-type injector 72")
`The images were recorded at the time of the spark (-30
`crank angle degrees). The mass of fucl impacting the
`piston is significantly’ greater for both carly and late
`injection whenthe pistion is closer to the injector tip.
`Histanies ofvisible pool fire luminosityfor late injec-
`tion (SOT = —90 crank angle degrees) arc shown In
`Fig 35-224, The pool burning ofthe fuel film is first
`visible at 20 crank angle degrees after the compression
`TDC as the piston descends below the fire deck. The
`poo!fires are attached to the fuel film on the piston top
`and follow the piston as it descends, Toward the end of
`the evele. pool fire luminosity weakens and the flames
`lift off the piston surface. At 300 crank angle degrees
`after the compression TDCtheflames have extinguished,
`although some liquid film maystill remain unburned
`due to the low volatility of the remaining liquid film
`and the low oxygen conccntration,
`
`The amountof fucl-wail impingementis kn
`vary significantly with injection timing and enee to
`thus the injection timing must be optimized in oct
`avoid spray over-penetration and the dasoer. to
`unintended wetting ofchamber surfaces ©245. 25», man
`It is gencrally agreed that the timing for carb; injecti
`should be adjusted so that the spray tip “chases” a
`descending piston, but does not significantly impact :
`Forthis the injection timingis critical, as the penetration,
`characteristics ofthe sac and main sprays must be maiched
`to piston recession velocities for a range ofengine specds
`and crank angles. Figure 5.5-3 shows an example
`comparison of the phasing of spray-tip penetration
`history and piston crown position at 1000 rpm for vari-
`ous injection timings *?).
`It is evident that the injection
`timing is critical to avoiding unintended sprayimpinge-
`ment for early injection, and to achieving the desired
`spray-piston-cavity targeting for creating a stable
`
`side view
`
`Bowditch view
`
`
`
`
`
`
`60 CAD
`
`120 CAD
`180 CAD
`240 CAD 300 CAD
`
`
`
`fe
`jen
`OS
`ire
`figure 35-2 2B
`gure
`33-2 Poolfire haminasity histories for liquidfitmsy: SOF myection timing
`af YO crank angle degrees 7"),
`
`La
`
`FORD Ex. 1010, page 164
`IPR2021-00340
`
`FORD Ex. 1010, page 164
` IPR2021-00340
`
`

`

`Mixture Formation Process and Approaches
`
`Injection Start
`
`120
`
`6YBTOC Oo
`
`60 OVEeeNePRO
`
` a
`
`20
`40
`60
`60
`Ratio of Fuel Well Impingement to Injected Fuel
`
`(%)
`
`Figure 5.5-4 Effect ofinjection timing on fuel wail
`wettingfor a combustion sysiem with an intake-
`side-mounted injector (467),
`
`ofthe piston crownis diminished significantly. whereas
`the wetting of the cylinder wall
`is nearly doubled as
`compared with that obtained for an injection timing of
`350°BTDC,
`It is worth noting that HC emissions are
`very sensitive to the specific fuel wetting location
`inside the combustion chamber Figure 5.5-5 shows a
`comparison of transient HC concentrations that are
`obtained for the same amount of liquid fuel wetting
`different locations inside the chamber. The lowest HC
`levels are obtained using LPG (liquefied petroleum gas).
`which vields no wall wetting. Figure 5,5-6 shows the
`effect of the engine operating (coolant) temperature on
`HC emissions for different in-cylinder fuel wetting
`
`14000
`
`12000
`
`1am)
`
` ¥4000
`
`ap
`
`000
`
`«on
`
`700
`
`Oo
`
`0
`
`mw
`
`240 x —_
`
`a
`
`™
`
`Crank Angle
`Figure 33-5) Portatrns in transient HC concentra-
`Hon with the location ofthe wetted fuel film: constumt
`Jilm mass; part load operation (1300 rpm, 0.262 MPa
`BMP); engine coolant temperature of 90°C
`(LPG: liquefied petroletan vas (74
`
`FORD Ex. 1010, page 165
`IPR2021-00340
`
`®Qc 30 5
`
`BDCG
`
`Piston
`Movemen
`0-20 80010
`BH 0
`Time ms
`Time me
`(a) Late Injection
`(b) Early Injection
`Fraure 5.5-3 Spray-tip penetrationsrelative to piston
`position al an engine speed of1000 rpm (late mjec-
`pur timing 18 based on compression TDC: early
`mjection Uiming is based on intake FDC) 25%,
`
`stratified-charge mixtute for late injection. Due to the
`time required for the physica! processes offuel evapora-
`tion and fuel-air mixing, each combustion system will
`exhibit a minimum time interval between the EOI and
`the occurrence of the spark that is required to avoid an
`oversrich mixture in the vicinity of the spark gap. By
`definition. the engine rotation rate in crank degrees per
`millisecond increases linearly with engine speed: as a
`result.
`the injection timing in terms of crank angle
`degrees must be advanced as the engine speed increases
`"Por purposes of monitoring the avoidance of spray
`Lnpingerment, SOT timing is the most meaningful, whereas
`“oT monitoring the mixture preparation interval, EOI tim-
`jie is the most applicable injection timing parameter. Both
`“Huld be recorded during G-DI engine developmentpro-
`tams. and both have been utilized in engine control
`“luonithms, The wamings given in Section 5.4.2 onthe
`ase differences between indicated and actual SO] and
`Ol should be kept in mind.
`“detailed investigation of the effect of injection
`“tum on fuel wall wetting for a prototype DI com-
`‘nistion system shows a number of key points {8 '83),
`‘igure 3 5-4 illustrates the results for an engine speed
`of 1400 rpm. Wheninjection occurscarlyin the intake
`tke. namely 350°BTDC on the compressionstroke,
`“‘=nicant fuel wetting ofthe piston crown occurs, with
`Its: or ne fuel wetting of the cylinder wall and head,
`Por dater injection timings the total extent of fuel wall
`‘stliny deereases markedly, althoughthere is a trade-
`OMbetween cylinder wall welling and piston crown wet-
`Ne. For an injection timing of 270°BTDC,the wetting
`
`FORD Ex. 1010, page 165
` IPR2021-00340
`
`

`

`alafometn
`
`DO Idie-90"C
`
`BB idle-4o"c
`
` -88§8&&
`
`Fienre 53-6 bifect ofcoolant temperature on
`HC emissions with different in-cylinderfuel weiting
`locations foridle operation 22"
`
`1 Gasoline Direct-Injection Epyines
`Pea
`A thicker wall film exhibits a lesser depree of
`Vaponzs.
`tion before ignition occurs, and the remainin fj
`yields a diffusion flame that produces aa MC!film
`indicates that even a small amountoffuel abies Thuy
`the piston crown can produce a considerable en
`smoke if the resultant wall film is relativety fia of
`should be noted that the temperatureofthe surfac ik lt
`ing the film also has a significant effect on fyg] i ay.
`ization. A cold wall significantly diminishes the, Pr
`film vaporization rate, and generally degrades the a
`tial uniformity of the fuel-air mixture, whereas a fa
`Piston Top=LPG-only
`H
`EE
`surface temperature in the range of 90°C to 130%¢ ica.
`crally promotes fuel film vaporization °!2)_ |p Pact.the
`injection timing 1s purposelyadvancedin someprodyc.
`tion G-DIenginesin orderto obtain spray impingement
`on a hot piston to improve fuel vaporization under cor-
`tain operating conditions “*. However,the charge-
`cooling benefit will be significantly reduced bythis
`opcrating strategy.
`The general relationship betweentheinitial spray -
`tip velocity and the spray SMD is shown in Fig. 5.5-8
`for a wide range of G-DI injector types and fue! pres-
`sure levels (228).
`Jt is evident from the data that dimin-
`ishing returns occur for the spray-atomization benefit
`that can be obtained byincreasing the fuel rail pressure.
`Anyfurther increasein the initial spray-tip velocity may
`result in excessive spray penetration, but maynot sig-
`nificantly enhance atomization. Currently most fuel
`systems for production G-DI engines utilize fuel pres-
`sures in the range of 5.0 to 10.0 MPa,althoughthere are
`some recent systemsutilizing variable fuel pressures that
`have the design authority to operate in the range of
`10 to 13 MPa. The curve showsthat selected levels of
`
`locations. Wetting of the exhaust valves (E-E) byfuel
`produces the highest HC emissions, as compared to
`wetting the piston top and intake valves(J-I). This trend
`is also true for a range of coolant temperatures (7)),
`Thepresenceofa film ofiiquid fuel on a chamber
`surface during the combustion event, whether intended
`or unintended, can have a significant effect on the mea-
`sured smoke emissions. Figure 5.5-7 shows the effect
`of fuel impingement and fuel film thickness on smoke
`emissions. The smoke emissions data are obtaincd in
`engine tests conducted for a range of injection timings
`andinjected fuel quantities, with the amountoffuel wall
`wetting andwall film thickness calculated by CFD for
`each test. It is evident that the level of smoke emissions
`is more dependenton the wali film thickness andless
`related to the total amount of fuel on the piston surface.
`
`[um] o
`
`SauterMeanDiameter
`
`10
`
`20
`
`Ra
`
`“
`
`“
`aa
`L.
`ig
`scat
`20
`15
`10
`5
`0
`Spray Tip Velocity [mvs}
`Average Fue! Flim Thickness on Piston Surface=um {Calcutated)
`,
`=e
`al
`Meure 335-8 Measured relationship herveenthe eauitte
`2
`spams
`7
`xprefp penetration rate cine! mec uroplet sizefor :
`‘
`;
`‘
`saree
`wie revige of injectors anefuel rail Pressures
`
` 25
`
`
`
`(Calculated) Wins
`
` 1600 mm, WOT and 40Nm FuelAmountimpingingon mg
`
`PistonSurface
`
`20
`
`figure 53-7 Effect offuel impingement on the piston
`crows andfuelfilm thickness on smoke emixstony 13"
`
`FORD Ex. 1010, page 166
`IPR2021-00340
`
`FORD Ex. 1010, page 166
` IPR2021-00340
`
`

`

`Mixture Formation Process and Approaches
`2SSSSe
`
`atomization for production combustion SYSLEMS are
`qssuciated with initial spray-tip yelocitics in the range
`af av to 40 ni/s, which is appronimatcly twice the
`iypical peak flow velocity of the in-eylinder flowfield.
`it should be noted that some newer types of G-DI
`injectors such as offset multihole and offsct slit-type
`yectors have initial spray-tip velocities in the range of
`onto 100 avs. Depending uponthe orientation ofthe
`spray axis relative to the in-cylinder flow field, the
`momentumof the fuel spray can be a substantial addi-
`tion to the momentum of the flowfield. Because the
`clearance height 1s generally quite small
`in G-DI
`engines, the fuel spray from G-DI injectors having high
`penetration rates could result in some degree of unin-
`jended wall impingement if targeting of the spray is not
`optimized. Such unintended impingement must be moni-
`tored carefully.
`
`5.5.2 Effect of Cylinder-Bore Fuel Wetting on
`Oil Dilution
`As discussed in the previous section, unintended fuel
`impingement on combustion chamber surfaces: cither
`on the cylinder wall or the piston crown, or both, often
`occurs in today’s G-DI engines. Non-optimalfuel injec-
`tion strategies may also lead to fuel wetting of the cylin-
`der wall, causing gasoline to be displaced by the moving
`piston ring and transported into the crankcase 120.380. 4°62,
`This can scverely degrade the oil quality, and can
`adversely impact the cylinder bore wear, Oil difution
`resulting from gasoline impingement on the cylinder wall
`is one of the factors to be investigated during the devel-
`opment of a production G-DI engine.
`Experimental results from the studyof oil dilution
`for a wide range of engine operating conditions are
`summarized in Fig. 5.5-96), As shown in Fig. 5.5-9a.
`moderate oil dilution is measured at high load with the
`
`Injection timing of DE enging
`Late injection
`Eariy injection
`(Stratified)
`
`#
`¥
`—
`oe
`(%}
`5
`=
`
`z=
`1
`gb 3
`a4
`1} Engine rev.
`: 2000rpe
`z
`S
`
`= i|Engine load : 8OT
`
`&
`=
`= 2 :|Injection period : 60T A
`
`a
`o
`2
`3 Water temp,
`: OT
`t
`Z
`
`
`an
`
`a
`
`360
`
`'
`270
`180
`Injection timing (Belore compression TOC TA)
`
`
`
`
`
`OilDilution(9%)
`
`(b)
`
`Engine rev.
`
`: 2000rpe
`
`Engine load : WOT
`
`Engine operation time
`
`(min)
`
`(Homagenesus} 100
`:
`
`a
`
`ce)
`
`5
`:i
`a
`=
`a
`
`PFI
`Engine rev.
`
`1
`
`-
`iv
`'
`‘
`i
`'
`i
`
`Water
`
`temp.
`
`(T)
`
`{c)
`53
`‘l
`cdifuti
`saxtiremertts (130
`Figure 3.35-9 Oil dilution measurements ela
`
`(d)
`
`159
`
`FORD Ex. 1010, page 167
`IPR2021-00340
`
`FORD Ex. 1010, page 167
` IPR2021-00340
`
`

`

`Automotive Gasoline Direc winjection Engines
`Invesiigations of ot! ditution in G-D
`G-DI engine, but
`is not observed in a baseline PFI
`dad PF
`engines 7") via oil sampling from the cy]
`enwine. For late injection corresponding to stratified-
`der hore
`periphery reveal
`that ot! dilution for the pF
`charge operation, a similar degree of oil dilution ts
`I enying 4,
`more pronounced on the exhaustside ofthe engi
`observed in both the PFI and G-DI engines. For this
`gil dilution occurs across the cylinder bon rae
`operadng point. the fuel fs well confined within the pis-
`G-DI engme. Both PFI and G-DI engines exhibit a theWear
`ton bow! and fuel wetting on the evlinder wall is effec-
`era ail dilution and increasing eheing
`tively avoided, For a similar reason. earlier injection
`pecific
`fucl consumption, There is evidence to inch
`timing is effective in limiting the oil dilution. as shown in
`thatoil dilution on the intake side results from fuel ais
`Fig, S.5-9b, Asillustrated in Figs. 3.5-9¢ and 5.5-9d. the
`entrainment through turbulent mixing, hema et
`coolant
`temperature and cngine operating time are
`dilution on the exhaust side results from direct impinoil
`important parameters in minimizing oil dilution. Clos-
`ment of the fucl spray on the cylinder bore sities,
`ing the flawcontrol valve on this cngine is found to
`intake-side-mounted injector.
`It is generally agreed at
`reduce oil dilution. duc to the fact that the particular air
`any operating parameters such as retarded injection ming
`flowpattem provided bythe flow control systemaffects
`and increased spray penetration that may increase fue!
`the amountofliquid fuel that impinges on the cylinder
`impingement on the cylinder bor