-
ABSTRACTNumerical simulations were performed in a four stroke
pent-roof SI engine under cold flow conditions at three
differentspeeds (1000 rpm, 2000 rpm and 3000 rpm) to investigate
theflow characteristics like swirl, tumble, turbulence,
massinducted and the formation of liquid film during cold
startcondition. The suction and compression strokes weresimulated
at the three speeds mentioned. Results show thatswirl and tumble
are increased from 1000 rpm to 2000 rpmwhereas the increase is very
low between the speeds 2000rpm and 3000 rpm. The computational
model was validatedagainst the experimental data available in the
literature. Thereis a slight increase in the mass of air inducted
as the speedgoes up. Further investigation with liquid film
modellingreveal, presence of liquid film mass on the cylinder
andintake region. Around 34% of fuel quantity converted to filmmass
in the port wall region of intake while using port fuelinjection
(PFI) at the entry of Siamese port. The piston regionhas around 8%
of fuel as liquid film at compression TDC.Later this may lead to
soot and Hydrocarbon emission. Fromthese results CFD code can be
confidently used to furtherinvestigation like port orientation,
manifold, spray, film andmixture formation and combustion.
KEYWORDSSwirl, Tumble, GDI, Turbulence, Mass Inducted, Liquid
FilmINTRODUCTIONAs the new Euro 6 regulations are to be introduced
in themarket, the pressure is ever more increasing on the
enginemanufacturers to produce cleaner engines with enhanced
performance. As the modern PFI engines are already refineda
further fuel consumption reduction is possible by newengine design
like combination of good features of Dieselengine with SI engine
resulting in GDI engine. The conceptof GDI though very old has not
been popular due to thedifficulty in controlling the injection of
fuel for varying loads.Due to the recent advances in electronics
and control systemthe attention has been focused again on GDI
engines. Theintake generated air motion is important in the SI
engine. It isespecially of paramount importance in the case of
GDIengines, where different flow field generation mechanismsare
utilized in the engine[1,2]. The cyclic variation of swirlmotion is
one of the factors affecting the combustion qualityin GDI
engines[2]. This cyclic variation persists even thoughthe velocity
fluctuations become homogeneous during thelate stages of
combustion[3]. The tumble motion especiallyplays a greater role in
GDI engines. They help in mixturepreparation and transportation.
However too much tumble atthe time of combustion can destroy
mixture distribution andthe burning rate will be decreased[9]. The
cone angle of thefuel injector also affects the spray penetration
inside theengine. If the spray angle is reduced the rate of
evaporation isreduced[5]. It is due to the denser travel of fuel
particles andlesser chance of fuel coming in contact with air. In
order toimprove upon this phenomenon as the cone angle of
anyinjector is more or less fixed, multiple injection
strategieswere tried. The in-cylinder gas density also affects the
spraypenetration, spray structure and hence evaporation.[4,6]. It
isalso reported in literatures that the timing of fuel
injectiontiming and spark timing strongly influences the
particulateemission from GDI engines[7, 8]. There are different
mixture
Investigation of Flow Field Pattern in a GDIEngine at Different
Speeds using NumericalTechniques
2013-01-2791Published
11/27/2013
Ramesh P and James Gunasekaran EAnnamalai University
Copyright 2013 SAE Internationaldoi:10.4271/2013-01-2791
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
preparation strategies for the direct injection engines,
likespray guided, air guided and spray guided[9, 10, 11, 12,
13].High pressure multiple injections were also investigated forthe
GDI engines and they help in smoother combustion andlesser
particulate emission compared to single pulseinjection[14, 15]. The
formation mechanism of liquid filmwere studied by various authors
and they reported that thefilm formation on the wall is due to
stick and spread[16, 17].An attempt has been made in this study to
investigate theflow field pattern of a commercially available SI
engine forits potential for GDI engine.
GEOMETRICAL DETAILS ANDDESIGNFor the computational analysis and
physical validation thedimensions of a commercial SI engine whose
geometricaldetails are available in the literature [Kim et al, 16]
is takenand a model is created in a commercial geometrical
modelerCATIA. The created model is exported to STARCD for
thecreation of mesh and application of physical parameters
andboundary conditions. Table 1 shows the geometrical detailsof the
engine and details of fuel injection.
Table 1. Engine Details and fuel injection
The details and features of the modeled engine are shown
inFig.1. Fig. 2 shows the discretized domain of the engine. Inorder
to understand the flow characteristics inside the engineat
different speeds simulations were conducted 1000 rpm,2000 rpm and
3000 rpm. The Siamese intake port is at the leftand the exhaust
port is at the right. The flow towards theexhaust port is assumed
to be positive and towards the intakeport is assumed to be
negative.
Fig.1. Draft of the Engine Model
(a). Isometric View of the Engine depicting the intakeand
exhaust Valves
(b). View through the sectional plane passing throughintake and
exhaust Valves
Fig. 2. (a) and (b) - Discretized domain of the Engine
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
VALIDATIONThe model is validated by comparing the calculated
CFDresults with the experimental results of Kim et.al.
Themeasurement points as done in the experiments (Kim et. al.)are
shown in Figure.1. Measurements were done using asingle component
LDV along the axial direction (V- line) andon a line which is
parallel to both the intake ports and passingthrough a point which
is 13 mm below TDC which is denotedas H-line. The spacing between
the measurement points is5mm. These locations are clearly
illustrated in Fig-1.
Mean Velocity Along the V-LineAs the intake ports are at the
left of the cylinder axis the flowof air towards the exhaust side
is assigned a positive value (u)and the one which is flowing
towards the intake port isassigned a negative (u) value. The x axis
is parallel to theports and measurements were done at 60, 120, 180,
240,300, and 360 crank angle (CA) aTDC. Fig. 3 shows thecomparison
of CFD results with that of the experimental(Kim et. al.) results.
The comparison pertains to themeasurement along the cylinder axis
as shown in Fig.1. TheValidation is between the experimental and
CFDmeasurement at 1000 rpm. The chart also shows the
velocitymeasurement along the same measurement points at
higherspeeds 2000 rpm and 3000 rpm. The CFD results arereasonably
in good agreement with the experimental data. At60 CA the tumble
centre is approximately at 5mm belowTDC. This crank angle the
rotating vortices have notdeveloped vigorously yet. This can be
understood by noticingthe low values of the velocity (towards the
intake) near thepiston crown. At higher speeds at the same crank
angle thetumbling vortices have strongly evolved at the same
crankangle with the negative velocity reaching a value as high as18
m/s near the piston crown. At 120 CA the predictedvalues along the
axis closely matches the experimental value.At 240 and 300 CA also
the predicted value closely followsthe trend measured in
experiment.
(a). 60 degree CAFig.3.
(b). 120 Degree CA
(c). 180 Degree CA
(d). 240 Degree CAFig.3. (cont.) (a) to (f) - Comparison of
Experimental
(Kim et al) and CFD Results at a speed of 1000 rpm andthe
Predicted Values at higher Speeds 2000 rpm and
3000 rpm along the V - Line
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
(e). 300 Degree CA
(f). 360 Degree CAFig.3. (cont.) (a) to (f) - Comparison of
Experimental
(Kim et al) and CFD Results at a speed of 1000 rpm andthe
Predicted Values at higher Speeds 2000 rpm and
3000 rpm along the V - Line
Mean Velocity Along the H-LineFig.4.shows the comparison of mean
velocity along the H-line and experimental value. Here also the
flow away fromthe intake port and towards exhaust port is treated
as positiveand the one flowing from the exhaust port towards the
intakeis treated as negative. Except for 60CA the other timingshows
a reasonable good agreement with the experimentalvalue. This may be
due to insufficient data available for thevalve lift profile and
port orientation. At 60 CA along the H-line the velocity shows a
negative direction at a distance of 5mm in both directions. At this
crank angle the piston is nearthe H-line and descending downwards.
This is in clearconformity with the earlier result along the
vertical line (V-line). This point to the fact that the incoming
air is organizedalong a clockwise rotating tumbling vortex. At
180CA it canbe observed that the vortex has almost stabilized and
the CFDresults follow the experimental trend.
(a). 60 Degree CA
(b). 120 Degree CA
(c). 180 Degree CA
(d). 240 Degree CAFig.4. (a) to (d) - Comparison of Experimental
(Kim et
al) and CFD Results at a speed of 1000 rpm and thePredicted
Values at higher Speeds 2000 rpm and 3000
rpm along the H-Line
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
Fluctuating Velocity Along the V-LineFig. 4 shows the comparison
between the experimentalfluctuating velocity and the simulated
result along the V- line.The simulated results are in good
agreement with theexperimental data at 1000 rpm. At higher speeds
thefluctuating velocity is showing a higher value. Thefluctuation
is maximum during the valve opening period andcontinuously
decreases well in to the middle of thecompression stroke. After
this the fluctuating velocity slightlyincreases indicating that the
destruction of tumbling vortices.The destruction of tumbling
vortices increases turbulenceresulting in a higher velocity as
evidenced by Fig. 5. Hencethe higher tumble motion helps in the
generation ofturbulence.
Fig. 5. Comparison of Temporal Variation ofFluctuating Velocity
along the V- Line with
Experimental Data (Kim etal)
RESULTS AND DISCUSSIONThe main advantage of present port
inducted engine is theprecise control of fuel quantity on cycle
basis. But they havecold start emission startup. The formations of
the film are themain reason for this and are influenced by flow
dynamics.The following section discusses the flow characteristics
andfilm formation.
Effect of Speed on Swirl and TumbleThe GDI engines are normally
designed to be tumbleoriented. The pent roof combustion chamber and
theorientation of the intake port geometry helps in the
generationof tumbling motion inside the engine cylinder. Fig. 6
showsthe temporal variation of swirl and tumble for the threespeeds
mentioned earlier.
The rotation of vortices about the axis Y (Try) is called
asnormal tumble about y coordinate and the tumbling vorticesabout
the axis X is denoted as cross tumble (Trx) and therotating motion
about the cylinder axis Z is called as swirl.Tumble motion inside
the engine cylinder can be divided intothree phases as generation,
stabilization and destruction. The
generation phase occurs usually during the intake whichproceeds
up to 200 CA. The stabilization and spin up phaseoccur due to the
upward moving piston. Because of themoving piston the spin up phase
enhances the tumble motionagain up to 300 degree. The tumble
destruction phase resultsin increased turbulence. This can be
confirmed with referenceto Fig 6 where there is a slight increase
in the turbulence levelstarting from 250CA and lasts up to 310CA.
The Crosstumble denoted by Trx is almost near zero level in
thisparticular geometry considered. The Swirl about the
cylinderaxis denoted by Srz increases up to 30CA and begins to
fadeout and eventually becomes zero at a crank angle of 70CA.From
this crank angle it begins to change direction androtates in
opposite direction and reaches a maximum value of0.93 around 120CA
aTDC. Regarding the effect of speed thetumble is increased as high
as two folds between the speeds1000 and 2000 rpm, whereas the
increment in tumblebetween 2000 and 3000 rpm is only marginal. This
may beattributed to the poor strength of the intake
generatedtumbling and swirling vortices which are not sustained
duringthe stabilization and spin up phase.
Fig. 6. Temporal variation of Swirl and Tumble insidethe Engine
Cylinder for Three Speeds
Effect of Speed on the Quantity of AirInductedThe speed of any
GDI engine is controlled by quality ratherthan by quantity, that
means the throttle controlled speedadjustment can be altogether
eliminated with qualitygoverning. Fig. 7 compares the effect of
speed on the quantityof air inducted. It is evident that the
accumulated masssteadily increases until 180CA and the mass
decreasesslightly until the valve closure period around 45CA
afterBDC. The accumulated mass is more at higher speed. This isdue
to the inducted air mass comes at higher velocity in to thecylinder
and this inertia of the incoming air is more than theinertia of the
air mass moving upwards near the piston crown.
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
Fig. 7. Effect of Speed on the Temporal Variation of AirInducted
into the Engine Cylinder
In Fig 8(a) at 60CA smaller clockwise rotating vorticesbegin to
form. At this stage the incoming air through theintake valves on
the exhaust side and near the cylinder linerside interact. By the
time of 120CA more air is flowingalong the cylinder liner towards
the piston. At 180CA thevortices are well organized and there is a
central region wherea low velocity region is prevailing and so is
the case at240CA where the tumble motion can be clearly seen.
At300CA the tumble motion can still be observed but at areduced
strength. From this point the vortices are beingdestroyed and by
the time the piston reaches TDC it iscompletely destroyed. The
effect of this is increasedturbulence. This is important in SI
engines as this turbulenceenhances flame propagation. It is also
important where thisturbulence is prevailing inside the combustion
chamber.
(a). 60 Degree CAFig.9.
(b). 120 Degree CA
(c). 180 Degree CA
(d). 240 Degree CAFig.9. (cont.) (a) to (f) - Comparison of Flow
Field alonga Sectional Plane Passing through Intake and Exhaust
Valves at Different Crank Angles
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
(e). 300 Degree CA
(f). 360 Degree CAFig.9. (cont.) (a) to (f) - Comparison of Flow
Field alonga Sectional Plane Passing through Intake and Exhaust
Valves at Different Crank Angles
Liquid Film MassAlso in this investigation, the accumulation of
liquid filmmass to the different regions of cylinder (dome, liner
andpiston) mand intake system (valve face, valve stem and portwall)
with PFI as shown in fig. 1 for motorized cold startcondition at
three speeds. Fig. 10 shows the characteristics ofliquid mass
formed at different parts of the engine.
(a). DomeFig. 10.
(b). Piston
(c). Cylinder liner
(d). Intake valve faceFig. 10. (cont.) (a) to (f) -
Characteristics of Liquid Film
Mass over the region of cylinder and intake system.
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
(e). Intake port wall
(f). Intake valve stemFig. 10. (cont.) (a) to (f) -
Characteristics of Liquid Film
Mass over the region of cylinder and intake system.
Fig. 11. 3D view of Liquid film formation over the regionof port
wall
The rate of accumulation of liquid mass is more at lowerspeeds
compared to that of at higher speeds. The maximumquantity is
accumulated before TDC during compressionstroke. The 35 percentage
of fuel mass quantity is collected inthe region of port wall of
intake system as liquid filmwhereas at other regions the collected
mass is less than 8percentage of fuel mass quantity. Fig. 11 shows
the 3Drepresentation of liquid film formation over the engine
portregion. This is due to the cold spray directly impinging on
the
intake port walls and the interaction lead to the formation
ofliquid film in this region.
CONCLUSIONComputational Fluid dynamic analysis on a SI engine
wasconducted at motoring conditions at different speeds. It
isvalidated against the experimental data. Swirl and normaltumble
plays are dominant role whereas the cross tumble isvery low. These
organized vortices enhance the turbulenceinside the combustion
chamber during the late stages ofcompression stroke. The turbulence
is more at higher speeds.The port injection at cold start condition
results inaccumulation of nearly 35 percentage of injected fuel
mass(45.873 mg) in the intake port region itself. This fuel
filmpersists until the compression TDC.
Due to the accumulation of film mass in the port region,
theengine is starved of fuel vapour in the first cycle leading
tomisfire. In the subsequent cycles the film mass acts as asource
point for further accumulation of the injected fuel.Due to this
cold start emissions will be more during the firstfew cycles. Hence
suitable measures like heated port duringinitial stages may reduce
the possibility of emission for theport injection.
REFERENCES1. Harada, J., Tomita, T., Mizuno, H., Mashiki, Z. et
al.,Development of Direct Injection Gasoline Engine, SAETechnical
Paper 970540, 1997, doi:10.4271/970540.2. Li, Y., Zhao, H., Peng,
Z., and Ladommatos, N., Analysisof Tumble and Swirl Motions in a
Four-Valve SI Engine,SAE Technical Paper 2001-01-3555, 2001,
doi:10.4271/2001-01-3555.3. Bergin, M. and Reitz, R., Effect of
Flowfield Non-Uniformities on Emissions Predictions in HSDI
Engines,SAE Technical Paper 2011-01-0821, 2011,
doi:10.4271/2011-01-0821.4. Hutchison, A. and Wicker, R., DISI Fuel
Spray EffectsDuring Simulated Injection Timing, SAE Technical
Paper2001-01-2026, 2001, doi:10.4271/2001-01-2026.5. Nishidaa K,
Jiangping Tinab, Sumotoa Y, Longb W, SatocK, Yamakawa M An
experimental and numerical study onsprays injected from two-hole
nozzles for DISI enginesProceedings of the combustion Institute
Volume 88, Issue 9,September 2009, Pages 1634 to 16426. Karaiskos,
E., Arcoumanis, C., and Hale, T., Flow andSpray Investigation in
Direct Injection Gasoline Engines,SAE Technical Paper 2002-01-0832,
2002, doi:10.4271/2002-01-0832.7. Maricq, M., Podsiadlik, D.,
Brehob, D., and Haghgooie,M., Particulate Emissions from a
Direct-Injection Spark-Ignition (DISI) Engine, SAE Technical Paper
1999-01-1530,1999, doi:10.4271/1999-01-1530.8. Davy, M., Williams,
P., and Anderson, R., Effects ofFuel Injection Pressure in an
Optically-Accessed DISI Engine
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM
-
with Side-Mounted Fuel Injector, SAE Technical
Paper2001-01-1975, 2001, doi:10.4271/2001-01-1975.9. Kim, S., Kim,
Y., and Lee, J., Analysis of the In-CylinderFlow, Mixture Formation
and Combustion Processes in aSpray-Guided GDI Engine, SAE Technical
Paper2008-01-0142, 2008, doi:10.4271/2008-01-0142.10. Piock, W.,
Weyand, P., Wolf, E., and Heise, V., IgnitionSystems for
Spray-Guided Stratified Combustion, SAE Int.J. Engines
3(1):389-401, 2010, doi:10.4271/2010-01-0598.11. Yamashita, H.,
Seto, M., Ota, N., Murakami, Y. et al.,Spray Guided DISI Using Side
Mounted Multi-HoleInjector, SAE Technical Paper 2007-01-1413, 2007,
doi:10.4271/2007-01-1413.12. Fan, Q., Hu, Z., Deng, J., Li, L. et
al., Stratified MixtureFormation and Combustion Process for
Wall-guidedStratified-charge DISI Engines with Different Piston
Bowlsby Simulation, SAE Technical Paper 2010-01-0595,
2010,doi:10.4271/2010-01-0595.13. Suh, E. and Rutland, C.,
Numerical Study of Fuel/AirMixture Preparation in a GDI Engine, SAE
Technical Paper1999-01-3657, 1999, doi:10.4271/1999-01-3657.14.
Serras-Pereira, J., Aleiferis, P., Richardson, D., andWallace, S.,
Mixture Preparation and CombustionVariability in a Spray-Guided
DISI Engine, SAE TechnicalPaper 2007-01-4033, 2007,
doi:10.4271/2007-01-4033.15. Katashiba, H., Honda, T., Kawamoto,
M., Sumida, M. etal., Improvement of Center Injection Spray Guided
DISIPerformance, SAE Technical Paper 2006-01-1001,
2006,doi:10.4271/2006-01-1001.16. Reitz, R. and Diwakar, R., Effect
of Drop Breakup onFuel Sprays, SAE Technical Paper 860469, 1986,
doi:10.4271/860469.17. Reitz, R. and Diwakar, R., Structure of
High-PressureFuel Sprays, SAE Technical Paper 870598, 1987,
doi:10.4271/870598.
The Engineering Meetings Board has approved this paper for
publication. It hassuccessfully completed SAE's peer review process
under the supervision of the sessionorganizer. This process
requires a minimum of three (3) reviews by industry experts.All
rights reserved. No part of this publication may be reproduced,
stored in aretrieval system, or transmitted, in any form or by any
means, electronic, mechanical,photocopying, recording, or
otherwise, without the prior written permission of SAE.ISSN
0148-7191
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.SAE Customer Service:Tel:
877-606-7323 (inside USA and Canada)Tel: 724-776-4970 (outside
USA)Fax: 724-776-0790Email: [email protected] Web Address:
http://www.sae.orgPrinted in USA
Downloaded from SAE International by Ramesh P, Wednesday,
November 27, 2013 10:49:08 AM