Downloaded from http://iranpaper.

ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

SAE TECHNICAL
PAPER SERIES 962058

Exploring the Use of Multiple Injectors

and Split Injection to Reduce DI
Diesel Engine Emissions

A. Uludogan, J. Xin, and R. D. Reitz

University of Wisconsin-Madison

Reprinted from: Modeling and Diagnostics in Diesel Engines

(SP-1205)

The Engineering Society International Fall Fuels &

For Advancing Mobility Lubricants Meeting & Exposition
Land Sea Air and Space® San Antonio, Texas
INTERNATIONAL October 14-17, 1996

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (412)776-4841 Fax:(412)776-5760

Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

The appearance of the ISSN code at the bottom of this page indicates SAE's consent
that 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 article
copy fee through the Copyright Clearance Center, Inc. Operations Center, 222
Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sections
107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of
copying such as copying for general distribution, for advertising or promotional
purposes, for creating new collective works, or for resale.

SAE routinely stocks printed papers for a period of three years following date of
publication. Direct your orders to SAE Customer Sales and Satisfaction Department.

Quantity reprint rates can be obtained from the Customer Sales and Satisfaction
Department.

To request permission to reprint a technical paper or permission to use copyrighted

SAE publications in other works, contact the SAE Publications Group.

GLOBAL MOBILITY DATABASE

All SAE papers, standards, and selected
books are abstracted and indexed in the
Global Mobility Database.

No part of this publication may by reproduced in any form, in an electronic retrieval

system or otherwise, without the prior written permission of the publisher.

ISSN 0148-7191
Copyright 1996 Society of Automotive Engineers, Inc.

Positions and opinions advanced in this paper are those of the author(s) and not
necessarily 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 is
published 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 publication

through SAE should send the manuscript or a 300 word abstract of a proposed
manuscript to: Secretary, Engineering Meeting Board, SAE.

Printed in USA 96-0049

 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
Exploring the Use of Multiple Injectors and Split
Injection to Reduce DI Diesel Engine Emissions
Copyright 1996 Society of Automotive Engineers, Inc.
ABSTRACT
This research uses computational modeling to explore
methods to increase diesel engine power density while
maintaining low pollutant emission levels. Previous
experimental studies have shown that injection-rate
profiles and injector configurations play important roles on
the performance and emissions of particulate and NO x in
DI diesel engines. However, there is a lack of systematic
studies and fundamental understanding of the
mechanisms of spray atomization, mixture formation and
distribution, and subsequently, the combustion processes
in spray/spray and spray/swirl interaction and flow
configurations. In this study, the effects of split injections
and multiple injector configurations on diesel engine
emissions are investigated numerically using a multi-
dimensional computer code. In order to be able to explore
the effects of enhanced fuel-air mixing, the use of multiple
injectors with different injector locations, spray
orientations and impingement angles was studied. The
interaction of the spray with the geometry of the
combustion chamber was also systematically studied. The
potential for the use of multiple injectors to increase
engine power density and to significantly reduce
particulate and NOx emissions in DI diesel engines is
revealed. This work demonstrates that multidimensional
modeling can now be used to gain insight into the
combustion process and to provide direction for exploring
new engine concepts.
INTRODUCTION
Diesel engines are the primary heavy-duty power plant
for both transportation and stationary use. Recently, the
application of diesel engines has also been expanded to
include automotive power sources which used to be
dominated by spark ignition (SI) engines. Due to
the overall lean combustion associated with the
non-homogeneous combustion in diesel engines, diesel
119
962058
A. Uludogan, J. Xin, and R. D. Reitz
University of Wisconsin-Madison
engines have superior energy efficiency and relatively low
carbon dioxide and carbon monoxide emission. However,
the nitric oxides (NOx) and particulate (mainly consisting
of soot at high engine loads) emitted from diesel engines
are pollutants of major concern.
In order to meet increasingly stringent emission
standards, extensive research has been carried out to
explore various ways to reduce NOx and particulate
emissions from diesel engines. The measures can be
categorized as exhaust after-treatment and in-cylinder
treatment. Unlike conditions in the three-way catalytic
converters used in SI engines, the conversion rate of NOx
to nitrogen is low due to the presence of excess oxygen in
diesel engine exhaust stream. Particulate traps are still in
the laboratory development stage, and there are presently
no efficient after-treatment methods to eliminate NOx and
particulate.
Various methods are currently being used in the
industry to reduce NOx and/or particulate formation within
the combustion chamber, such as the use of ultra high
injection pressures, injection timing retard, exhaust gas
recirculation (EGR), etc. However, these traditional
strategies usually reduce one pollutant at the price of
increasing the other pollutant. Recently, multiple injection
strategies have been shown to be able to realize the
simultaneous reduction of both NOx and particulate
emissions, and experimental and numerical results have
demonstrated that the soot-NOx trade-off curves can be
shifted closer to the origin [1,2,3]*. When combined with
EGR, multiple injection has shown great potential to
reduce both NOx and particulate without negative effects
on fuel consumption [2]. Computer modeling has shown
that the reason for the reduced particulate emission is the
lower soot formation rates at the tip of the injected sprays
that are associated with the use of multiple-injections [3].
* Numbers in brackets designate references listed at the
end of the paper.
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
In the above mentioned in-cylinder pollutant reduction
measures, attention has been mainly focused on relatively
minor design modifications to existing diesel engine
concepts. However, it is not clear how engine power
density and NOx and soot formation mechanisms would
be influenced by radical changes to the geometric shape
of combustion chambers, or to parameters such as the
number of injectors, the nozzle locations, and the spacing
and orientations of injectors if multiple injectors are used.
It is of great practical interest to explore methods to
investigate the effects of new design factors on diesel
engine combustion and emissions. Experimentally
achieving this goal would be very time-consuming and
costly; computational models offer a more cost-effective
alternative.
After many improved submodels were implemented in
the KIVA-II code, the code has been verified against
numerous experimental investigations over a range of
engine operating conditions [3,4]. Successes suggest that
it may now be reasonable to use CFD models for
exploring new design concepts and for understanding the
physical and chemical phenomena within diesel engine
combustion chambers. Therefore, the first objective of the
present work was to apply the models to explore the
performance of new design concepts.
As an example, a radically new concept here called a
Multiple Injector Combustion System (MICS) was chosen
for investigation since this concept represents a means of
significantly enhancing fuel-air-mixing, and it is of interest
to assess the potential for simultaneous NOx and
particulate reduction as a limiting (albeit difficult to
implement in practice) concept. Special interest is placed
on the effects of interaction between sprays and the gas
phase swirling motion, and spray/spray impingement on
diesel engine combustion and emissions. The second
objective of the present work was to explore the effect of
injection rate history on diesel emissions by using split
injections to allow comparison with the multiple injector
concept. This comparison is of interest since prior studies
have already demonstrated the very beneficial effects of
split injections on diesel engine performance [1,2,3].
It should be noted that the purpose of this present
study is to explore the potential emissions reduction
capabilities of various engine concepts. Issues related to
the feasibility or practicality of design concepts are more
appropriately discussed once the underlying physical
processes are understood and promising engine design
concepts have been identified.
DESCRIPTION OF THE MODELED
COMBUSTION SYSTEM
It is well known that the fuel spray and interaction of
the spray and the combustion chamber walls are key
factors in controlling fuel distribution, fuel-air mixing, and
subsequent combustion. A multihole, single injector
combustion system is a traditional design concept, and it
is applied in almost all heavy-duty-sized diesel engines. In
120
single injector DI diesel engines, fuel distribution and
mixing are restricted by the spray penetration and the
strength of the in-cylinder gas swirl motions. By adopting a
multiple injector design, there is more flexibility in
controlling the fuel-air mixing process, and the swirling air
motion can also be enhanced by the momentum
exchange between the sprays and the gas.
Other mechanisms which might play important roles in
diesel engine emission reduction also should be noted.
With the MICS concept, a fuel spray can be injected into
the core region of another spray resulting spray/spray
impingement which might enhance atomization processes
in the core region [5]. In addition, theoretical and
experimental investigations have shown that properly
injecting fuel into combustion-generated high temperature
zones suppresses soot formation [2,3]. With MICS, it is
relatively easy to realize control of spray/spray interaction.
This makes the concept useful in a basic study of the
factors that influence engine performance and emissions.
To isolate other factors and focus on combustion
chamber geometry, spray/swirl interaction and spray/
spray impingement, four combustion system geometries
are considered in this investigation. All have the same
compression ratio (15.1:1) and the same displacement
volume (2.4L). The combustion chambers are variants
based on the Caterpillar 3406 heavy-duty truck diesel
engine geometry (see Table 1). The combustion chamber
of a single cylinder of the Caterpillar 3401 DI diesel
engine, shown in Fig. 1a, has a Mexican hat piston bowl
with a cusp and a six-hole injector located at the center of
the bowl. The second combustion chamber design has
the in-piston bowl without cusp, and a six-hole injector at
the center, as shown in Fig. 1b. The multiple injector
combustion system (MICS), shown in Fig. 1c, has six
single hole injectors equally spaced around the edge of
piston bowl. Fuel is injected tangentially or with an angle θ
in order to enhance the swirling motion of gas phase, as
shown in Fig. 2. Two cases were considered; first where θ
was the same for all six injectors (Fig. 2b) and second
where the spray were staggered and oriented as shown in
Fig 2c. The last configuration, shown in Fig. 1d, has a
bowl which is identical to the Caterpillar 3401 but with the
same injection configuration used in MICS. Figure 2
shows the top-view schematic of the considered
combustion chambers. Swirl is defined to be positive if air
rotates clockwise, negative if counterclockwise. The angle
θ represents the rotation angle of the injector in the x-y
plane, measured counterclockwise from the straight line
OA. The other engine specifications used in the
computations are given in Table 1.
COMPUTATIONAL TECHNIQUE AND MODEL
DESCRIPTION
The program used in the current study is the KIVA-II
code [6] originally developed by the Los Alamos National
Laboratory. This multi-dimensional code has as its primary
application the investigation of complex
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
Figure 1. Schematic of combustion chamber (side
view). MICS is Multiple Injector Combustion System.
combustion processes in internal combustion engines.
After many improved submodels were implemented [4],
the KIVA-II code has been used successfully to predict DI
diesel engine performance under various engine
operating conditions [3,7,8].
The KIVA-II computer code solves the compressible,
turbulent, three dimensional transient conservation
equations for reacting multicomponent gas mixtures with
the flow dynamics of an evaporating liquid spray. In order
to eliminate the uncertainty of specifying the droplet size
distribution at the nozzle exit, the wave breakup model [9]
was adopted in KIVA. In the wave breakup model the
initial droplet diameters are set equal to the nozzle hole
diameter. The atomization process of the relatively large
injected droplets is modeled using a stability analysis for
liquid jets.
Diesel spray combustion was simulated by using a
characteristic-time combustion model [10] which was
initially developed for spark ignition engines. Similar to the
SI engine combustion model, the conversion rate of the
partial density of species is given by
121
Figure 2. Schematic of combustion chambers
(top-view)
∂ρ m ρm – ρm *
----------- = ------------------------
∂t τc
where ρm* is the local and instantaneous thermodynamic
equilibrium value of the species m partial density, and τc is
the characteristic time to achieve such equilibrium. By
including carbon monoxide and hydrogen among the
combustion products, the characteristic-time combustion
model is able to predict heat release while keeping its
computational efficiency [11].
To model diesel engine ignition delay, a multistep
kinetics model (Shell model) was implemented in KIVA-II.
In the Shell model [12], eight generic species are used to
represent fuel, intermediate species, and products. The
premise of the Shell model is that degenerate branching
plays an important role in determining the cool flame and
two-stage ignition phenomena that are observed during
the autoignition of hydrocarbon fuels. A chain propagation
cycle is formulated to describe the history of the branching
agent together with one initiation and two termination
reactions [7,12].
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

Cylinder bore x 137.6 x 165.1

stroke (mm)
Connecting rod 261.62
length (mm)
Displacement volume (L) 2.44
Compression ratio 15.1
Nozzle orifice diameter 0.259
(mm)
Combustion chamber Quiescent
Intake surge tank pressure 184 kPa Figure 3 Computational mesh for the Caterpillar
intake surge tank air 310 K engine.
temperature
Intake valve closure -147 degrees ATDC RESULTS AND DISCUSSION
Swirl ratio (nominal), S0, 1.0
Figures 4a and 4b show computed and measured
Engine speed 1600 rev/min cylinder pressures and heat release rates for the single
Fuel Amoco Premier #2 cylinder version of the Caterpillar DI diesel engine for a
Injection system Common Rail baseline single injection case [3]. The “measured” heat
Injection pressure 90 MPa release rates are derived from the measured cylinder
Fuel injected 0.1622, 0.2433 g/cycle pressure data using a first law analysis. The operating
Overall equivalence ratio 0.48, 0.72 conditions are 1600 rev/min and 75% of the rated load. A
single injection was used with a start of injection timing of
Injection duration 21.5, 32.25 degrees
8.5 BTDC. As can be seen, there is a good agreement
injection timing 8.5 degrees BTDC between the measured data and the model prediction.
Table 1. Caterpillar 3401 engine specifications and The computed soot and NOx emissions for various
operating conditions injection timings are shown in Fig. 5. It is shown that the
KIVA predicted emissions have fairly good agreement with
The turbulent flows within the combustion chamber are the measured engine-out particulate and NOx emissions,
simulated using the RNG k- ε turbulence model, modified although the agreement deteriorates at very retarded
for variable-density engine flows [13]. Within combustion timings where other factors such as significant SOF levels
chambers, the gas density varies greatly, not only in the particulate may be important [8]. The good
temporarily but also spatially. To account for this, a new agreement between the measured and computed results
wall temperature function was used in computing the wall (seen in Figs. 4 and 5) for the baseline engine gives
heat flux [14]. NOx is modeled with the extended confidence in the model predictions, and suggests that the
Zel’dovich mechanism [15] and soot emissions are model may be used to explore new engine concepts.
modeled with the Hiroyasu formation model [16] and the
Nagle and Strickland-Constable oxidation model [17] as
described by Patterson et al. [18].
As shown in Fig. 1, the combustion chamber geometry
and spray configurations considered herein are
periodically symmetric; for computational efficiency the
domain considered is only a 60° sector of the combustion
chamber. (For the MICS-3 concept a 120 degree sector
was employed that contained two of the six sprays.)
Figure 3 shows the computational domain for the
Caterpillar 3401 diesel engine (case 1a). There are 20
cells in the radial direction, 30 cells in the azimuthal
direction and 18 cells in the axial direction with 5 cells in
the squish region at top dead center. This mesh resolution
has been shown to give adequately grid-independent
results [13]. Similar grid resolution was used for the other
engine geometries in Fig. 1, and additional grid sensitivity Figure 4a. Comparisons of measured and computed
studies showed that those results were also adequately cylinder pressure for baseline engine case 3401-1 at
grid-independent. The computations used tetradecane 1600 rev/min and 75% load.
(C14H30) as the fuel due to its similar C/H ratio to diesel
fuel.

122
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

represents case number of the corresponding combustion

system. The letter “O” after the first dash means that the
initial swirl flow direction is counterclockwise with the
same magnitude as other the cases, i.e., the swirl ratio is,
-1.0. Unless otherwise specified, the engine operating
conditions of all cases correspond to the Caterpillar 3401
baseline case, i.e., 75% load and 1600 rev/min engine
speed.
From a preliminary optimization study, it was found that
MICS with Lz = 0.7cm, Lr = 4cm and θ = 100 had superior
emission performances. Thus, this case (MICS-1) was
chosen as the MICS baseline case. Interestingly, from the
numerical simulation results seen in Table 2 for cases
3401-1 and flat-1, it was found that the piston bowl shape
does not have an important impact on emissions at the
operating condition considered.
Figure 4b. Comparisons of measured and computed
The fuel/air mixing process is significantly influenced
heat release rate for baseline engine case 3401-1 at by the engine and injector orientation geometries. Figure
1600 rev/min and 75% load. 6 presents the total fuel mass fraction within certain
equivalence ratio ranges as a function of normalized burnt
fuel mass for the case 3401-1 and the case MICS-1,
which correspond to a lean mixture (between 0.05 and
0.5), a “flammable” mixture (between 0.5 and 2.0) and a
rich mixture (above 2.0), respectively. It is found that the
masses of lean and flammable mixture for MICS are
significantly increased at the moment the mixture is auto-
ignited (approximately 2 BTDC). In the traditional centrally
mounted injector combustion system, the fuel sprays are
clustered together near the injector. However, in MICS,
liquid fuel is injected into the space where excess air is
available. In addition, since the injectors are located
farther away from each other, the adjacent sprays do not
compete for air in the vicinity of the nozzles. As a result,
the spray atomization and dispersion is enhanced, which,
Figure 5. Comparison of measured engine-out in turn, increases vaporization rates, and more fuel vapor
is mixed with the surrounding air during the ignition delay
particulate and NOx emissions and predicted soot
period.
and NOx results at 0, -3, -6, -8.5, and -10 degrees
ATDC fuel injection timings for the baseline 3401-1 case Lr Lz θ NOx soot
case. (cm) (cm) g/kgf) g/kgf)
3401-1 0.09 0.381 0 37.5 0.63
MICS COMBUSTION - As described in Fig. 1, it is
expected that the nozzle location (Lr) and the spray 3401-1-flat 0.09 0.381 0 40.8 0.62
orientation (θ) are two important parameters in the MICS 3401-2 0.09 0.381 0 44.4 0.76
concept. Therefore, before making comparisons among (%50 more
combustion systems, a parametric study was performed fuel
to investigate the effects of the parameters Lz, Lr and θ of injected)
MICS (see Figs. 1 and 2) on combustion and emissions MICS-1 4 0.7 100 32.1 0.10
(Lz is the distance between the nozzle exit and the head.)
MICS-1-0 4 0.7 100 33.8 0.09
The half spray angle ψ for centrally mounted injector
cases (see Figs. 1a and 1b) is 62.5°, and the investigated MICS-1- 4 0.7 100 39.6 0.09
spray angle, θ, from the injector centerline for MICS (see CUSP
Fig. 1c and Fig. 1d) is 100°. The considered combustion MICS-2 4 0.7 100 32.4 1.41
systems and parameters are listed in Table 2. The first (%50 more
column in Table 2 gives the case name, the letters or fuel
numbers prior to the first dash stand for combustion injected)
system, as shown in Fig. 1. For instance, “3401”
represents Caterpillar 3401 engine, “flat” means centrally MICS-3 (% 4 0.7 100 42.7 0.63
mounted injector with flat piston bowl, “MICS” stands for 50 more and
Multiple Injector Combustion System with the flat piston fuel 220
bowl, and “MICS-1-CUSP” stands for multiple injector injected
combustion system with Caterpillar 3401 piston bowl Table 2. Spray locations and orientations considered, and
geometry (see Fig. 1d). The number after the first dash
emissions for injection at 8.5 BTDC

123
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

It should also be noted that the fraction of the fuel at

lower fuel/air equivalence ratios in MICS-1 is higher than
that in the 3401-1 case during the major diffusion
combustion period, as shown in Fig. 6.
Corresponding heat release histories are given in Fig.
7. The ignition delays for the two cases are almost
identical. However, due to improved mixing and dispersion
in MICS-1, the heat release spike resulting from the
premixed burn is much higher. The pressure histories are
presented in Fig. 8. Because more fuel is consumed
during the premixed combustion for MICS-1, the pressure
prediction shows a slightly different result.
The total in-cylinder emission predictions as a function
of time are shown in Figs. 9 and 10. The NOx emission
from MICS-1 is slightly lower compared to that from the
3401-1 baseline case with the same injection timing, and
the soot emission from MICS-1 is significantly lower-about
a 90% reduction (see also Fig. 11). The numerical results
show that MICS has great potential to reduce soot while
keeping a similar NOx emission level. The fact that the
MICS NOx is comparable to that of the central injection
case, in spite of its higher premixed burn peak, is due to
enhanced mixing during the ignition delay which leads to
reduced temperatures during the diffusion burn (i.e.,
between 10 and 25 degrees ATDC).
It is known that NOx formation is very sensitive to
temperature [17]. Beyond 2000 K, the NOx formation rate
is exponentially proportional to local temperature. Due to
the high premix burn in MICS-1 case, the average in-
cylinder temperature, seen in Fig. 12, reaches higher
values in the MICS-1 case before 10 degrees ATDC.
During this time, NOx formed by the MICS-1 shows an
increase relative to the 3401-1 case (see Fig. 9).
However, as the diffusion burn takes control of the
combustion starting around 10 degrees ATDC (see Fig.
7), the in-cylinder temperature is lower in the MICS-1 case
compared to the 3401-1 case. The relatively higher in-
cylinder temperature in the 3401-1 case, thus, increases
the NOx formation more dramatically between 10 and 30
degrees ATDC. The in-cylinder temperature history after
40 degrees ATDC does not affect the NOx formation since
the NOx has been moved to the relatively cool regions
where the NOx formation is very low. The relatively higher
diffusion burn rate increases the engine-out NOx
formation in the 3401-1 case, as seen in Fig. 9 by 80
degrees ATDC. Figure 6. Variations of In-cylinder fuel mass fraction in
Soot is likely to be formed in rich mixture zones. the specified equivalence ratio ranges for the baseline
Comparing the central injection and MICS-1 cases in Fig. 3401-1 central injection and MICS-1 combustion
10, it is found that the soot formation curves are very systems.
similar for the baseline central injection and the MICS
presents the cylinder average temperature variation with
combustion system. However, in MICS, the soot is
crank angle. In fact, the average in-cylinder gas
oxidized very quickly, which is clearly shown in Fig. 10.
temperature in MICS is higher than for the 3401-1 case by
There are two factors that control the soot oxidation:
about 100 K from 30 ATDC onwards. Both, the higher in-
temperature and the availability of oxidizer (air). Since
cylinder temperature after 30 degrees ATDC and the well
MICS enhances the premixed combustion, and more
mixed air/fuel mixture seen in Fig. 6, contributes the
chemical energy is released near top dead center (TDC),
enhanced soot oxidation in the MICS-1 case.
it is expected that the average in-cylinder gas temperature
is higher, as is confirmed in Fig. 12 which

124
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
Figure 7. Comparison of heat release histories for
baseline central injection and MICS-1 combustion
systems.
Figure 8. Comparison of pressure histories for
baseline central injection and MICS-1 combustion
systems.
Figure 9. Comparison of NOx histories for the
baseline 3401-1 central injection and MICS-1
combustion systems.
125
Figure 10. Comparison of in-cylinder soot formation,
soot oxidation, and net soot production histories for
baseline 3401-1 central injection and MICS-1
combustion systems.
EFFECT OF SWIRL - In addition, the swirling motion
of the gas phase plays an important role in the mixing
processes between the fuel and air, and between partially
oxidized products (soot, CO, etc.) and air.
Swirl effects on spray atomization and dispersion can
be described from two points of view. In centrally mounted
injector cases, the swirling motion drags fuel vapor and
small droplets away from the spray centerline to the
downstream volume between two adjacent sprays. Thus,
fuel is distributed non-homogeneously within the
combustion chamber. Another effect is centrifugal-force-
induced-stratification. When there exists a high swirling
motion, liquid fuel and fuel vapor (with a higher density
than the ambient gases) in the core region of the sprays
tend to move from the inner region to the outer region
where fresh air is available. In contrast, the hot (low
density) combustion products move towards the inner
region. Thus, it is expected that the induced secondary
mixing due to the swirl motion might enhance soot
oxidation.
Figure 13 reports the effect of sprays on the in-cylinder
gas swirl number. For all cases, the initial swirl number at
intake valve closure is assumed to be 1.0. The in-cylinder
swirl number is calculated by summing mass-weighted
angular velocities over all computational cells, i.e.,
where mi is the gas mass contained in cell i, ωi, is the
angular velocity of the fluid in cell i, mtot is the total gas
mass in the cylinder, and ωengine is the angular velocity of
the engine crankshaft. For complicated swirl flows such as
those of MICS, the swirl flow changes its direction
spatially and temporally, and the above definition cannot
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

Figure 11 Spray (black color) and soot iso-surface (9e-7 g/cm3) (gray color) predictions for the MICS-1 and 3401-1
cases at 5, 15, and 50 degrees ATDC (Spray parcel size shows the amount of fuel in each parcel).

126
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

Figure 12. In-cylinder mass averaged temperature

histories for central injection 3401-1 and MICS-1
combustion system designs.
give a clear picture of the strength of the instantaneous in-
cylinder swirl motion. Therefore, an absolute swirl number
is also defined as

As shown in Fig. 13, at the time of injection, the swirl

number in the charge has been increased to -1.5, since,
as the piston approaches top-center, the air in the squish
region is forced into the compact bowl-in-piston
combustion chamber. Neglecting the effects of friction,
angular momentum is conserved and the angular velocity
therefore must increase. For the 3401-1 case, the spray
centerline is perpendicular to the swirl direction. The swirl
increases when the fuel vaporizes. Some swirl number
oscillations can be observed during the ignition. The swirl
number decreases gradually due to the dissipative effects
and the gas expansion after the injection period.
On the other hand, in the MICS case fuel is injected
tangentially at a small angle with respect to the swirl
motion. In modern fuel injection systems, the spray
velocity at the nozzle exit is greater than 300 m/s, and it is
expected that the momentum exchange between the
sprays and the surrounding fluid in the spray direction is
very significant. In the presented case MICS-1, the swirl
motion direction of the surrounding fluid is opposite to the
spray injection direction. Due to the momentum input from
the sprays, the swirl flow gradually changes its direction
(swirl number changes from -1.5 to 1.4), as shown in Fig.
13a, and then it stabilizes around 1 (a similar magnitude
as that of the 3401 base case but with different sign).
Figure 13. Swirl ratios for central injection and MICS
combustion system designs.
Figure 13b shows the absolute swirl number
variations. It is clear that the swirl motion is significantly
enhanced by the sprays. It was observed from the
predicted velocity fields (not shown) that the swirl flow in
the piston bowl changes its direction near TDC while the
initial swirl flow direction near the cylinder wall is
maintained until the later stages of the expansion stroke.
This indicates that the fluid far away from the sprays is
influenced less than that in the spray region. The
interaction of the spray and the swirl flow in MICS
significantly enhances the mixing of unburned and
partially burned species with the air, and most of the
formed soot is thus oxidized quickly. Similar results have
been obtained using alternative methods of enhancing
soot oxidation, such as the use of air jets to increase
mixing later in the combustion process [19].
To investigate the effect of the swirl flow direction on
emissions, an additional case (MICS-1-O) was calculated
in which the sprays and the initial swirl flow have the same
direction. Figure 14 shows the swirl number variation
versus crank angle. Compared with the case MICS-1, the
absolute swirl number is much higher for the case MICS-
1-O. However, the emission predictions for cases MICS-1
and MICS-1-O, given in Table 2, show that the direction of
swirl motion with respect to the spray direction has little
impact on

127
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

emissions since the spray induced swirl dominates the

process.
It is known that NOx is formed in the early stages of
combustion. The NOx level freezes out at a maximum
value at about 30 degrees after top dead center, as shown
in Fig. 9, since the piston expansion lowers the charge
temperature. Therefore, the high temperature zones
during the early combustion period have important effects
on the NOx formation rate. Figure 15 shows the mass
fraction of the cylinder gas with local temperatures greater
than 2600 K. It is found that the mass fraction for case
3401-1 is lower compared to the MICS-1 case before 10
degrees ATDC. In MICS, though the amount of premixed
burn is higher, the improved mixing depresses the
formation of stoichiometric mixture zones (i.e., the Figure 15. Percent of the cylinder gas mass over 2600
combustion is leaner). Furthermore, the strong swirl flows K for baseline central injection and MICS cases.
enhance the mixing of the hot combustion products with
the fresh charge which has a relatively low temperature.
As a result, the high swirl velocities may lead to short
residence times of combustion product gases in the high
temperature zones. As an integrated result of the above
factors, the NOx emission level from MICS-1 does not
differ much from the 3401-1 case.

Figure 16. Comparison of soot-NOx trade-off curves

Figure 14. Effect of initial swirl direction on swirl ratio
for MICS combustion system.
While engine-out NOx formation is about the same, in
the case of MICS-1 the soot oxidation occurs at a higher
rate and this decreases the net soot formed to a much
lower level (see Fig. 10). One reason behind the much
lower soot emission is the location of the soot in the
combustion chamber. Injecting the fuel from the center
generates an air motion that causes the soot to move from
the bowl toward the low temperature head [3]. MICS on
the other hand, uses the fuel momentum in the radial
direction to keep the soot in the bowl and in the squish
regions where the temperatures are much higher
compared to those in the rest of the engine. Having the
soot away from the walls increases the soot oxidation
rate.
The superior emission performance of MICS for a
specific injection timing (8.5 BTDC) has been
demonstrated above. It is of interest to examine MICS
performance over a range of injection timings. Figure 16
for the baseline central injection and MICS
combustion systems for various start of injection
timing.
shows the predicted NOx-soot trade-off curve. For
reference, the NOx and soot emission predictions for the
3401-1 baseline case with various injection timings are
also presented. It is seen that the MICS soot emission is
significantly lower than the central injection baseline
cases within the entire injection timing range considered.
The NOx-soot trade-off curve of MICS is shifted closer to
the origin.
INCREASED POWER DENSITY - Since the amount of
soot emission is decreased substantially with MICS, it is of
interest to explore whether more fuel can be injected each
cycle to obtain a higher power density without undesirable
soot and NOx emission increases. Fig. 17 shows
predictions of the cylinder gas pressure, average
temperature, heat release rate, air/fuel mass fraction with
equivalence ratio greater than 2.5, and emissions with 50
percent more fuel injected in each cycle, with start of
injection timing of -14 degrees ATDC in each case. The
overall equivalence ratio is increased from 0.48 to 0.72.
The fuel injection duration is also increased to 32.25
degrees to keep the spray velocity at the nozzle exit the
same as for the 0.48 equivalence ratio cases. Description
of the overall results for 3401-2, and MICS-2 (with flat
bowl) are also given in Table 2. (As noted before the 3401
flat bowl results are seen in Table 2 to be similar to the
3401 cusp bowl results.)

128
Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

Figure 17. Predicted (a) cylinder pressure, (b) heat

release rate, (c) cylinder average temperature, (d)
mass fraction with equivalence ratio greater than 2.5,
and (e) soot/NOx predictions for 50 percent more fuel
injected per cycle for the baseline central injection
and MICS systems.
As in the lower equivalence ratio case, well premixed
fuel/air in the MICS-2 concept generates a higher pre-
mixed burn heat release rate. The diffusion burn initially
shows a lower heat release rate. This lowers the engine
peak temperature and, therefore, the NOx production (see
Fig. 17e). Soot production, on the other hand, is higher in
the MICS-2 case due to the lower combustion rate during
the early stage of combustion (see Fig. 17b) which
increases the extent of the fuel rich regions and therefore
the soot formation rate, as shown in Figs. 17d and 17e.
Thus, the engine-out soot is increased because of the
higher soot formation rate in the MICS-2.
Figure 18 shows the engine-out soot and NOx
predictions for various fuel injection timings for the MICS-
2 and 3401-2. The NOx predictions are lower in the MICS-
2 case. The soot, on the other hand, is lower in the 3401-2
case. Compared to the MICS-1 and 3401-1 cases, shown
in Fig. 16, the NOx predictions are lower despite the 50
percent more fuel injected in the MICS-2 and 3401-2
cases. Both the MICS-2 and 3401-2, however, produce
significantly more soot compared to the MICS-1 and
3401-1.

Figure 18. Comparison of soot-NOx trade-off curves

for the 3401-2 and MICS-2 cases.

129
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

The injectors in the MICS-2 case are equally spaced

around the edge of piston bowl. The injected spray
directions are the same for all injectors (see Fig. 2 and
Table 2). The centrifugal force produced by the strong
spray angular momentum forces the spray and air to the
edge of the piston bowl in MICS-2. Due to the higher
density of the spray compared to the oxygen gas, the
spray is forced further toward the edge of the piston bowl
relative to the less dense oxygen gas. This enhances air/
fuel mixing in the MICS-1 case due to the large amount of
available fresh air at the edge of the piston bowl, as seen
in Fig. 6. However, as the amount of fuel injected is
increased to increase the power density, more fuel now
resides at the periphery of the bowl and the local fuel/air
ratio increases. Therefore, the soot increases as more
fuel is injected with the current injector configuration in
MICS-2. One possible solution to reduce the extent of the
fuel rich regions and therefore the net soot formed is to
enhance the air/fuel mixing more by reoptimizing the
injector orientation for each injector in MICS for the higher
power density case of MICS-2. Thus, a new fuel injector
configuration was chosen to optimize the air/fuel mixing.
In order to spread the fuel to a larger volume, 3 of the 6
injectors were pointed toward the center of the engine
cylinder while the remaining 3 injectors were pointed to
the edge of the engine bowl as shown in Fig. 2c. This
configuration of the injectors can reduce the amount of
fuel that resides at the periphery of the engine bowl
resulting in a lower overall equivalence ratio that can
reduce the soot and NOx. Figure 19 shows the in-cylinder
average pressure, heat release rate, average
temperature, soot, and NOx predictions for the MICS-3
and 3401-2 cases.
The main difference between the MICS-2 and MICS-3
cases (compare Figs. 17 and 19) is the heat release rate
during the diffusion burn after 10 degrees ATDC. The fuel
injected toward the center of the engine as well as that at
the edge of the engine bowl improves the mixing of the
fuel and air causing higher diffusion burn rates in the
MICS-3 case. Thus, the in-cylinder temperature is
increased before 10 degrees ATDC and again after 40
degrees ATDC. While the earlier relatively high in-cylinder
temperature increases the soot formation dramatically in
the MICS-3 case (see Fig. 19d), the higher soot oxidation
rate enhanced by the high in-cylinder temperature after 40
degrees ATDC reduces the net soot level to a lower value
compared to the 3401-2 case. Figure 20 has the soot and
NOx trade-off curve for the 3401-2, MICS-2, and MICS-3
cases. It is clear that the MICS-3 case provides better
emissions at the various fuel injection timings shown. The
emission, predictions of the MICS-1 and MICS-2 and 3
cases indicate that the engine emissions depend on the
engine spray geometry. The lower soot and NOx
predictions of the MICS-3 case shows that the MICS Figure 19. (a) Cylinder pressure, (b) heat release rate,
system can provide different fuel injector configurations
(c) cylinder average temperature, and (d) soot/NOx
that can be used to effectively control the emissions.
predictions for 50 percent more fuel injected per cycle
for the baseline central injection and the MICS-3 case
(fuel injection timing is -14 degrees ATDC).

130
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

In MICS, the fuel is injected toward the other injectors

from the edge of the engine bowl (see Fig. 2). This
configuration keeps the fuel inside the engine bowl and
squish regions near the injectors where the in-cylinder
temperature is relatively high compared to that in the rest
of the engine. Thus, controlling the spray in the high
temperature region is already done by the MICS single
injection configuration. That is why the soot is much lower
due to the high soot oxidation rate compared to the
3401-1 case. Using split injection, therefore, does not
contribute any additional reduction of the soot, as seen in
Fig. 21. The fuel injected during the second pulse is
rapidly consumed near the nozzle exit due to the higher
local temperature caused by the first pulse (see Fig. 25).
Figure 20. Comparison of soot-NOx trade-off curves Therefore, the extent of the fuel rich regions and thus the
for the 3401-2, MICS-2, and MICS-3. soot formation rate are reduced in the split injection
SPLIT INJECTION - It is of interest to explore methods MICS-1 case (see Figs. 26 and 27). Despite the lower
to reduce the soot and NOx emissions even more in the soot formation rate, the engine-out soot value is higher
MICS concept. Previous studies have shown that split due to the lower soot oxidation rate in MICS. Since the
injection can reduce both soot and NOx emissions fuel injection occurs at a later time due to the dwell time
significantly [1,2,3]. It is worthwhile, therefore, to explore between the two pulses and the location of the soot is not
the use of split injection in conjunction with the MICS changed in the MICS systems, less soot is oxidized during
concept. Engine-out soot and NOx predictions for split the split injection in MICS (see Fig. 27). Also, more fresh
and single injections are presented in Fig. 21 for the air/fuel mixing occurs during the dwell time and this
MICS-1 and 3401-1 cases. The predictions in Fig. 21 were increases the fuel burning rate at the beginning of the
done by using the same fuel equivalence ratio, 0.48 for all second pulse around 10 degrees ATDC, as seen in Fig.
cases. The injection velocity was also kept the same for 25. This contributes to the formation of the NOx during the
each case. Thus, the fuel injection duration was chosen to split injection in the MICS systems.
be 29.5 degrees for the split injection case. In the first
pulse 75 percent of the fuel is injected. The rest of the fuel
was injected after an 8 degrees dwell time. The
predictions show that the split injection reduces soot
significantly for the 3401-1 case to levels that approach
those of MICS-1. This result is in agreement with the
previous studies [3]. However, the soot is seen to be
increased when split injection is used in the MICS-1 case.
Figure 22 shows that the normalized mixture volume of
the gas mixture that has equivalence ratio greater than
2.5 is lower with a split injection in the 3401-1 engine. The
fuel injected during the second pulse is rapidly consumed
near the nozzle exit due to the higher local temperature Figure 21. Soot /NOx trade-off curve using single and
caused by the first pulse (see Fig. 23). Thus, the extent of split injections with various fuel injection timings.
the fuel rich soot forming regions is reduced in the split
injection 3401-1 engine [3] (see Fig. 24). The location of
the soot formation regions is important in determining the
final engine-out soot production at the end of the fuel
injection cycle. In split injection, forcing the combustion to
take place near the nozzle also keeps the soot in a
relatively high temperature region. This increases the soot
oxidation rate therefore decreases the soot formed by the
end of the fuel injection period. However, the NOx
predictions are similar in the split and single injection
cases because of the similar in-cylinder temperature
history in the 3401-1 engine while the initial 75 percent of
the fuel is injected during the first pulse, as seen in Fig.
23. By about 30 degrees ATDC the NOx formation
freezes. Figure 22. Normalized mixture volume with
equivalence ratio greater than 2.5 for single and split
injection cases for the 3401-1 case.

131
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
Figure 23. Average in-cylinder temperature during the
single and split injection cases for the 3401-1 case.
Figure 24. Comparison of in-cylinder soot formation,
soot oxidation, and net soot production histories for
single and split injection cases for the 3401-1 case.
Figure 25. Average in-cylinder temperature during the
single and split injection for the MICS-1 case.
While the use of a split injection increases the engine-
out soot amount in MICS, the soot predictions are still
seen to be lower compared to those of the single injection
case in the 3401-1 system. The soot predictions are even
lower compared to the 3401-1 engine with split injection at
-14 degree fuel injection timing, as shown in Fig. 21. Split
injection effects on the soot and NOx emissions seem to
be strongly dependent on the fuel injector configuration.
132
Figure 26. Normalized mixture volume with
equivalence ratio greater than 2.5 for single and split
injection cases for the MICS-1 case.
Figure 27. Comparison of in-cylinder soot formation,
soot oxidation, and net soot production histories for
single and split injection cases for the MICS-1 case.
SUMMARY AND CONCLUSIONS
A new Multiple Injector Combustion System (MICS)
concept is explored to enhance the air usage in a DI
diesel engine. The MICS system provides a standard
against which the performance of other, possibly more
practical, engine designs can be compared. Beside the
injector location and orientation, the use of single and split
fuel injection schemes are also studied to get lower
pollutant emissions together with higher engine power
density.
The simulations indicate that locating the injectors at
the edge of the engine bowl enhances the swirl motion at
the early as well as at the later stages of each engine
cycle resulting in improved air/fuel mixing. Having the
injectors outside of the engine bowl provides the
opportunity to inject fuel any place into the combustion
chamber, and it also helps to enhance the swirl motion.
The swirl generated by the spray is relatively strong, and
this helps to mix the fuel/air mixture and causes higher
soot oxidation and less soot production in each engine
cycle. NOx production, on the other hand, does not
change much due to the similar temperature history in the
combustion chamber during the early stages of
combustion when NOx is formed.
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016
In order to increase the power generated during each
cycle the effect of increasing the amount of fuel injected
was studied. The predictions show that MICS has much
less NOx but somewhat higher soot production. The
decreased NOx is due to the lower peak temperature
resulting from the better air/fuel mixing. However, with the
MICS-2 injector configuration, the spray is directed more
toward the edge of the piston bowl. Thus, injecting more
fuel increases the extent of the fuel rich regions and
hence the soot formation is increased. In order to reduce
the fuel rich regions created by the high swirl motion, a
different injector configuration (MICS-3) was chosen and
much lower soot and NOx were predicted with this
multiple injector system. This shows that the engine spray
geometry conditions affect the emissions and the MICS
system provides different injector configurations that can
reduce soot and NOx productions significantly.
The effect of split injection on emissions was also
studied. A considerable difference in the emissions trends
is seen with the use of split injection for the 3401-1 and
MICS-1 systems. Soot is reduced considerably with split
injection in the 3401-1 engine, as also seen in previous
experimental and modeling studies. NOx is reduced by
the use of split injection in the MICS-1 system due to the
lower in-cylinder temperature caused by the well mixed
air/fuel mixture early in the combustion. Soot on the other
hand, is increased with the use of split injection in the
MICS-1 system. The reduction of the net soot formed
seen in the 3401-1 engine was not seen in the MICS-1
system because the location of the soot formation regions
(which was a main reason for the soot reduction in the
3401-1 system) is not changed in the MICS-1 system.
The potential for the use of multiple injectors to
increase engine power density and to significantly reduce
particulate and NOx emissions in DI diesel engines is
revealed. This work demonstrates that multidimensional
modeling can now be used to gain insight into the
combustion process and to provide direction for exploring
new engine concepts.
ACKNOWLEDGMENTS
This work has been supported at the University of
Wisconsin-Madison by the Army Research Office under
Grant NO. DAAH04-94-G-0328, and by Caterpillar, Inc.
and DOE/NASA-Lewis.
REFERENCES
1. Tow, T. C., Pierpont, D. A., and Reitz, R. D., “Reducing
Particulate and NOx Emissions by Using Multiple
Injections in a Heavy Duty DI Diesel Engine,” SAE Paper
940897, 1994.
2. Montgomery, D. T., and Reitz, R. D., “Six-Mode Cycle
Evaluation of the Effect of EGR and Multiple Injections on
Particulate and NOx Emissions from a DI Diesel Engine,”
SAE Paper 960316, 1996.
133
3. Han, Z., Uludogan, A., Hampson, G., and Reitz, R. D.,
“Mechanism of Soot and NOx Emission Reduction Using
Multiple-injection in a Diesel Engine,” SAE Paper 960633,
1996.
4. Reitz, R. D., and Rutland, C. J., “Development and
Testing of Diesel Engine CFD Models,” Prog. Energy
Combust. Sci. Vol. 21, pp. 173-196, 1995.
5. Sinko, K.M., Shih, S., and Chehroudi, B., “Emission
Characteristics of a Dual-injector Diesel Fuel Injection
System”, SAE Paper 960839, 1996.
6. Amsden, A. A., O’Rourke, P. J., and Butler, T. D.,
“KIVA-II -A Computer Program for Chemically Reactive
Flows with Sprays,” Los Alamos National Laboratory,,
LA-11560-MS, 1989.
7. Kong, S. C., and Reitz, R. D., “Multidimensional
Modeling of Diesel Ignition and Combustion Using a
Multistep kinetics Model,” Journal of Engineering for Gas
Turbine and Power, Vol. 115, pp. 781-789, 1993.
8. Xin, J., Montgomery, D. T., Han, Z., and Reitz, R. D.,
“Computer Modeling of the Six-Mode Emissions Test
Cycle of a DI Diesel Engine,” submitted to Journal of
Engineering for Gas Turbine and Power, 1996.
9. Reitz, R. D., “Modeling Atomization Processes in
High-Pressure Vaporizing Sprays,” Atomization and Spray
Technology, Vol. 3, pp.309, 1987.
10. Abraham, J., Bracco, F. V., and Reitz, R. D.,
“Comparisons of Computed and Measured Premixed
Charge Engine Combustion,” Combustion and Flame,
Vol. 60, pp. 309-322, 1985.
11. Reitz, R. D., and Bracco, F. V., “Global Kinetics
Models and Lack of Thermodynamics,” Combustion and
Flame, Vol. 53, pp.141-143, 1983.
12. Halstead, M. P., Kirsch, L. J., and Quinn, C. P., “The
Autoignition of Hydrocarbon Fuels at High Temperatures
and Pressures - Fitting of a Mathematical Model,”
Combustion and Flame, Vol. 30, pp.45-60, 1977.
13. Han, Z., and Reitz, R. D., “Turbulence Modeling of
Internal Combustion Engines Using RNG K-e Models,”
Combustion Science and Technology, Vol. 106, pp.
267-295., 1995.
14. Han, Z., and Reitz, R.D., “A Temperature Wall
Function Formulation for Variable-Density Turbulence
Flows with Application to Engine Convective Heat
Transfer Modeling,” accepted for publication, International
Journal of Heat and Mass Transfer, 1996.
 Downloaded from http://iranpaper.ir
Downloaded from SAE International by Shanghai Jiao Tong Univ, Monday, August 29, 2016

15. Westenberg, A. A., “Kinetics of NO and CO in Lean,

Premixed Hydrocarbon-Air Flames,” Combustion Science
and Technology, Vol. 4, pp 59-64, 1971.

16. Hiroyasu, H., and Kadota, T., “Models for Combustion

and Formation of Nitric Oxide and Soot in DI Diesel
Engines,” SAE Paper 760129, 1976.

17. Nagle, J. and Strickland-Constable, R. F., “Oxidation

of Carbon between 1000-2000 C,” Proceeding of the fifth
Carbon Conference, Vol. 1, p.154, 1962.

18. Patterson, M. A., Kong, S.-C., Hampson, G. J., and

Reitz, R. D, “Modeling the Effects of Fuel Injection
Characteristics on Diesel Engine Soot and NOx
Emissions,” SAE Paper 940523, 1994.

19. Mather, D.K., and Reitz, R.D., “Modeling the Use of

Air-injection for Emissions Reduction in a Direct-injected
Diesel Engine” SAE Paper 952359, 1995.

134