Applied Thermal Engineering 37 (2012) 165e175

Contents lists available at SciVerse ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Effects of spray impingement, injection parameters, and EGR on the combustion

and emission characteristics of a PCCI diesel engine
Robert Kiplimo a, *, Eiji Tomita a,1, Nobuyuki Kawahara a, 2, Sumito Yokobe b
a
Dept. of Mechanical Engineering, Okayama University, Tsushima-Naka 3-1-1, Kita-Ku, Okayama 700-8530, Japan
b
Diesel Group, Technology Support & Development Dept., Machinery Factory, Machines & Systems Headquarter, Mitsui Engineering & Shipbuilding Co., Ltd, Tamano Works, Japan

a r t i c l e i n f o a b s t r a c t

Article history: The effects of spray impingement, injection parameters, and exhaust gas recirculation (EGR) on the
Received 8 August 2011 combustion characteristics and exhaust emissions of a premixed charge compression ignition (PCCI) diesel
Accepted 5 November 2011 engine were investigated using a single-cylinder test engine and an optically accessible engine. Tests were
Available online 15 November 2011
carried out under constant speed with variable injection pressures and EGR rates. Exhaust emissions and
in-cylinder pressures were measured under all experimental conditions. Analyses were conducted based
Keywords:
on diesel spray evolution and combustion process visualisation coupled with performance and exhaust
Internal combustion engine
emissions. Higher injection pressures led to lower smoke, hydrocarbons (HC), and nitrogen oxide (NOx)
Spray
Airefuel mixture
emissions but had roughly the same CO emissions compared with lower injection pressures. Higher EGR
PCCI rates led to the simultaneous reduction in NOx and soot emissions due to lower combustion temperatures
Flame luminosity compared to conventional diesel combustion. However, HC and CO emissions increased due to fuel
Exhaust emissions impingement, bulk quenching, and over-mixing, leading to an airefuel mixture that was too lean to burn.
An optimum spray targeting spot was identified, leading to lower emissions of soot, CO, and HC but higher
NOx emissions without EGR. The simultaneous reduction in NOx and soot was achieved using the optimum
spray targeting spot by introducing EGR, which was accompanied by homogenous combustion and a low
luminosity flame attributed to fuel impingement on the piston bowl wall.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction for engines, new combustion strategies, such as homogeneous

With growing concerns about fossil-fuel depletion, environ-
mental protection, and global warming, production engines with
higher fuel conversion efficiency and lower engine emissions are
urgently needed. Additionally, many countries are implementing
stricter emission regulations. Diesel engines are more attractive for
powering light duty and heavy commercial vehicles because of their
superior fuel economy, durability, reliability, and high specific
power output compared to spark-ignition (SI) engines [1]. However,
conventional diesel combustion suffers from increased particulate
matter (PM) and nitrogen oxide (NOx) emissions. This implies that
newly developed diesel engines must have lower PM and NOx
emissions while maintaining or attaining higher fuel conversion
efficiency. To meet the strict performance and emission targets set
* Corresponding author. Tel.: þ81 90 6430 1830; fax: þ81 86 251 8266.
E-mail addresses: dns421215@s.okayama-u.ac.jp, robertkiplimo@yahoo.com
(R. Kiplimo), tomita@mech.okayama-u.ac.jp (E. Tomita), kawahara@mech.
okayama-u.ac.jp (N. Kawahara), yokobesu@mes.co.jp (S. Yokobe).
1
Tel.: þ81 86 251 8049; fax: þ81 86 251 8266.
2
Tel.: þ81 86 251 8235; fax: þ81 86 251 8266.
charge compression ignition (HCCI), have been widely investigated.
HCCI eliminates locally rich-fueleair mixtures and reduces the
combustion temperature, thus simultaneously achieving low PM
and NOx emissions with relatively good engine performance.
However, HCCI faces many challenges, including a lack of combus-
tion phase control under different operating conditions, fuels with
different properties, high pressure-rise rates under high loads, and
limitations in creating homogenous fueleair mixtures [2].
Different combustion concepts have been developed to control
the in-cylinder combustion process, in addition to the use of
expensive exhaust after-treatment devices. The concept of
employing in-cylinder control parameters, which is considered the
most practical, least expensive, and most effective control measure,
has been widely investigated. These parameters include but are not
limited to injection pressure, number of injections, shape and
timing of the injection, boost pressure, EGR, and swirl ratio [3,4].
One of the most promising combustion concepts is an HCCI
combustion strategy implemented in modern direct injection
diesel engines through partially premixed charge compression
ignition (PCCI), where fuel and air are not fully homogenous, but
the combustion event can be controlled more readily.

1359-4311/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.applthermaleng.2011.11.011
166 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175

In the PCCI combustion strategy, fuel can be introduced into the Table 1
combustion chamber through port fuel injection, early direct Engine specification and operating conditions.

injection, or late direct injection. Port fuel injection and early direct Engine type 4-stroke, single-cylinder, water cooled
injection often suffer from incomplete fuel vapourisation and fuel Bore  Stroke 96  108 mm
spray impingement on the cylinder walls, which causes high levels Swept volume 781.7 cm3
Compression ratio 13
of HC and CO as well as fuel/oil dilution [5,6]. Strategies to reduce Combustion system PCCI, direct injection
fuel-wall impingement explored in the past include the use of Combustion chamber Derby hat
narrow spray cone angle injectors [7e9] and uncooled EGR [10]. Late Engine speed 1000 rpm
direct injection avoids fuel-wall impingement and provides good Intake pressure 101 kPa
Injection system Common rail system
control over combustion phasing. A standard injector with an
Fuel injection pressure 80 MPa, 140 MPa
included cone angle of 156 can be used, thus allowing for transition Fuel injection quantity mf 12.2 mg/cycle
between PCCI at low and medium loads to conventional diesel Straight-hole injector F 0.1 mm  4holes
combustion at high loads [11e13]. Under the PCCI combustion Nozzle (VCO type)
strategy, the desired ignition delay is achieved through a lower Included cone angle 140
Injection timing sweep 2e40 BTDC
compression ratio, enhanced charge motion, higher injection pres- EGR rate 0e40%
sures, and relatively large amounts of cooled external EGR. Unlike Intake temperature 40  C
HCCI, PCCI is not fully homogeneous, but it uses injection timing and Coolant temperature 80  C
EGR to greatly increase the controllability of combustion phasing Lube oil temperature 80  C
and the rate of combustion. Much research has been conducted to
expand the high load limits and reduce the engine-out emissions of
capable of developing an injection pressure of 180 MPa was used. A
PCCI using fuel properties and additives [14e16], multiple injections
valve-covered orifice (VCO) injector with four 0.1 mm-diameter
[17e20], variable valve timing [21], and fueleair mixing enhance-
holes placed symmetrically in the nozzle tip was used. The top dead
ment [13,22]. PCCI combustion strategies employing moderately
centre (TDC) signals and every half-degree crank angle (CA) were
early injection, qinj z 20 before top dead centre (BTDC), have been
detected by photo interrupters and coupled with a controller to
widely investigated because they are advantageous in preventing
control the injection timing and injection duration. In-cylinder
lubricant dilution. Under such cases, high levels of EGR coupled with
pressure was measured with a piezoelectric pressure transducer
low compression ratios are applied to ensure that there is sufficient
(6052C, Kistler) coupled with a charge amplifier (5011B, Kistler).
airefuel mixing time, suppressing NOx formation and improving
The pressure history was analysed to obtain the heat release rate to
combustion phasing. Hence, to simultaneously achieve low soot and
investigate the combustion characteristics. Exhaust emissions were
NOx under moderately early injection timings, there is a need to
captured using a NOxeCO analyser (Horiba, PG-240), HC analyser
optimise the injection pressure, EGR rate, compression ratio, and
(Horiba, MEXA-1170HFID), and smoke meter (Horiba, MEXA-600s).
injection timing [23e27]. Like HCCI, PCCI is prone to high HC and CO
The data for each engine condition were captured when the engine
emissions and high pressure-rise rates, resulting in high combustion
was in equilibrium, during which there was almost no change in
noise or knocking; hence, it becomes difficult to implement the
the emission parameters and exhaust temperatures.
concept under high engine loads. Previous research indicates that
the simultaneous reduction in engine NOx and soot emissions can be
achieved by initiating combustion at an equivalence ratio below 2 2.2. Optical engine
and flame temperatures below 1800 K [11].
To clearly understand the combustion mechanism inside the A single-cylinder optically accessible engine was used for the
combustion chamber during PCCI combustion, there is a need to experiment. The optical engine was warmed up by circulating
study, in detail, the effects of injection timing, injection pressure, heated coolant maintained at 80  C, similar to the test engine. The
and EGR rate in the preparation of the airefuel mixture, the heat optical engine was operated in skip-fire mode to limit the risk of
release process, and the formation of soot, NOx, and other sapphire window failure due to thermal and mechanical stresses.
combustion products. Visualisation of spray development and the Diesel spray bottom-view images utilised for the spray character-
combustion process is necessary to relate these parameters to istic analysis were obtained using a high-speed camera (NAC,
engine emissions. There are still insufficient fundamental data to Memrecam GX-1). The visualisation system is shown in Fig. 1(B). An
fully understand the relationship among the injection parameters, elongated piston was installed to enable the mounting of a p/4-
mixture formation, and combustion process during PCCI. The radian mirror beneath the piston sapphire window, which gave
present study addressed the effects of spray impingement, injec- a 62 mm-diameter visual field for the 96 mm bore. Combustion
tion pressure, injection timing, and EGR on the formation of images were obtained at a frame rate of 13,000 fps with a resolution
emissions in a PCCI engine and its overall performance. The injec- of 400  400 pixels; thus, images could be taken at CA intervals of
tion pressure and injection timing were varied with and without 0.462 . In this work, injection pressure and injection timing were
EGR to establish the best operating conditions for PCCI combustion. varied while the mass injected was kept constant (mf z 12.2 mg/
cycle, equivalent to l ¼ 4.5 at 0% EGR and l ¼ 3.0 at 40% EGR, where
2. Experimental setup and procedure l is the excess air ratio).
Natural flame luminosity was used to visualise the combustion
2.1. Test engine process. The natural flame luminosity of a diesel flame consists of

A four-stroke, single-cylinder, direct-injected supercharged Table 2

diesel engine with a displacement of 781.7 cm3 was used for the Test fuel properties (JIS#2 diesel fuel).

investigation. Table 1 shows the engine specifications and the Cetane number 57e60
operating conditions. Table 2 shows the specifications of the low- Density (@15  C g/cm3) 0.8217
sulphur JIS #2 diesel fuel used in the experiment, which is Lower heating value (kJ/kg) 42.9
Sulphur mass (ppm) <10
commonly available in Japan. A schematic diagram of the test Flash point ( C) 64
engine is shown in Fig. 1(A). A common rail injection system
 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175 167

Fig. 1. Engine experimental layout: (A) test engine and (B) optical engine.

both chemiluminescence and soot luminosity [28]. Chem-
iluminescence is produced by the excitation of combustion radicals
due to exothermic chemical reactions occurring during the auto-
ignition process; soot luminosity is caused by the emission of hot
soot particles and is characterised by a continuous broadband
spectrum, similar to the Planck black body emission curve. Soot
luminosity is generally comparable to or higher than the chem-
iluminescence after ignition [29].
2.3. Engine operating conditions
of injection pressure on the combustion and exhaust emissions,
with a slightly narrow included cone angle of 140 to avoid cylinder
wall wetting. Engine speed was maintained at 1000 rpm. The
operating conditions are tabulated in Table 1. An injection timing
sweep from 2 to 40 BTDC was investigated.
In this study, simulated EGR based on N2 gas dilution was used,
as indicated in the formula below:
N2
EGR rate ¼ (1)
Air þ N2

In this investigation, late and moderately early injections were Because it is difficult to predict the exhaust gas composition for
considered. The optical and test engines were operated at two real EGR conditions, the effect of nitrogen dilution studied earlier
injection pressures, Pinj ¼ 80 MPa and 140 MPa, to isolate the effect was considered [30].

Fig. 2. In-cylinder pressure history and rate of heat release, Pinj ¼ 80 and 140 MPa, 0% EGR, l ¼ 4.5.
168 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
3. Results and discussion
3.1. Effect of injection pressure on combustion characteristics and
specific emissions
Fig. 2 shows the in-cylinder pressure history and rate of heat
release for the injection timings used in this research. Each pressure
curve was obtained by averaging 84 individual pressure traces. The
higher injection pressure, Pinj ¼ 140 MPa, led to a higher in-cylinder
pressure. For retarded conditions, a single heat release peak was
observed, indicating that premixed combustion dominated the
combustion process with a slight diffusion-controlled portion. With
the moderately advanced injection timing, a two-stage heat release
pattern was observed: the first stage was associated with low-
temperature reactions, and the second stage was associated with
high-temperature reactions. The higher injection pressure resulted
in a higher heat release peak with a rapid burn rate. The higher
injection pressure also led to shorter combustion duration.
Fig. 3 shows the effect of injection pressure on indicated thermal
efficiency, indicated mean effective pressure (IMEP), and the coef-
ficient of variance of the IMEP. Indicated thermal efficiency is
defined as the ratio of the indicated work per cycle to the amount of
fuel energy supplied per cycle as shown in Equation (2).
Indicated work per cycle IMEPxVd
hi ¼ ¼ (2)
Input energy mf QLHV
where hi is the indicated thermal efficiency, vd is the engine
displacement volume, mf is the mass of fuel supplied to the engine
per cycle and QLHV is fuel energy supplied considering the low
heating value of the fuel. For conventional diesel combustion,
advancement of the injection timing led to a simultaneous reduc-
tion in the indicated thermal efficiency and IMEP, with
Pinj ¼ 140 MPa being superior to Pinj ¼ 80 MPa. This could be due to
the short injection duration and hence longer premixing time
before the onset of combustion, as well as the higher fuel exit
velocity, leading to better atomisation for the higher injection
Fig. 3. Effect of injection pressure on indicated thermal efficiency, IMEP, and coeffi-
cient of variance of the IMEP, Pinj ¼ 80 and 140 MPa, 0% EGR, l ¼ 4.5.
pressure. Stable indicated thermal efficiency values and IMEP were
realised with a phasing close to TDC because there was minimum
negative work done. There was an optimal injection timing for
which a higher indicated thermal efficiency and IMEP were ach-
ieved. This was realised between 15 and 20 BTDC and was
thought to be related to the spray targeting position and fuel
impingement on the piston bowl wall, which was discussed in the
spray analysis and combustion images in our previous work [31].
The coefficient of variance of the IMEP for the tested condition was
less than 3%, implying that combustion was stable.
Fig. 4 shows analysis conducted to understand the effect of
injection pressure on specific emissions. The in-cylinder tempera-
ture for the higher injection pressure was higher than that of the
lower injection pressure. Under all conditions, smoke emissions
were lower for the higher injection pressure. This could be attrib-
uted to the short injection duration, leading to a long premixing
time and better fuel atomisation; hence, it would be possible to
avoid a fuel-rich zone that burns with soot emissions. An injection
timing of qinj ¼ 30 BTDC and earlier resulted in higher smoke
emissions, with lower injection pressures displaying higher values.
This phenomenon was observed despite the fact that the premixing
time was longer. This could be related to the fact that under this
condition, there was fuel impingement on the piston surface, and
hence splashing to the crevices, which became a rich-fuel zone that
burned with soot emissions.
With a late injection timing of qinj ¼ 2 e15 BTDC with
Pinj ¼ 140 MPa, NOx emissions were higher compared to
Pinj ¼ 80 MPa; however, the opposite occurred in the PCCI
combustion regime with qinj ¼ 20 e40 BTDC. This implies that the
combustion transitioned to PCCI and the in-cylinder temperature
decreased, leading to lower NOx emissions for the higher injection
pressure. The higher injection pressure created better atomisation
and vaporisation, which in turn increased the premixed burned
Fig. 4. Effect of injection pressure on specific emissions, Pinj ¼ 80 and 140 MPa, 0% EGR,
l ¼ 4.5.
 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
fraction and the local combustion temperature, leading to higher
NOx emissions and a rapid pressure-rise rate due to the rapid
combustion. In a PCCI regime, the in-cylinder temperature is rela-
tively low and the premixing time is relatively long, leading to mild
combustion. The higher in-cylinder temperature led to the oxida-
tion of soot but also promoted higher NOx emissions. The injection
pressures of Pinj ¼ 140 MPa and Pinj ¼ 80 MPa had roughly the same
CO emissions under all injection timing conditions. With early
injection timing, CO emissions increased drastically. This could be
attributed to fuel impingement on the surface of the piston, leading
to incomplete fuel combustion. As the injection timing was
advanced, HC emissions for both injection pressures were identical,
with slightly higher values for the lower injection pressure of
Pinj ¼ 80 MPa. For conventional diesel combustion (qinj ¼ 5 e15
BTDC) and a moderately early injection timing/PCCI regime
(qinj ¼ 20 e25 BTDC), it is possible to achieve less than 2 g/kWih,
while at qinj ¼ 30 BTDC and earlier injection timings, the lower
injection pressure led to higher HC emissions compared to the
higher injection pressure. This could be due to fuel impingement on
top of the piston surface, leading to incomplete fuel vapourisation
and oxidation with a significant variation in the equivalence ratio,
creating either over-rich or over-lean regions.
3.2. Effect of EGR on combustion characteristics and specific
emissions
With the improvements achieved using the higher injection
pressure of Pinj ¼ 140 MPa without EGR, further advancements in
performance and exhaust emissions were sought by introducing
EGR. In this section, the effects of EGR on combustion characteris-
tics and exhaust emissions will be explored in detail. Fig. 5 shows
the in-cylinder pressure history and heat release rate at
Pinj ¼ 140 MPa with and without EGR. The peak in-cylinder pressure
Fig. 5. In-cylinder pressure history and rate of heat release, Pinj ¼ 140 MPa, 0% EGR, l ¼ 4.5; and Pinj ¼ 140 MPa, 40% EGR, l ¼ 3.0.
169
and heat release rate decreased with the introduction of EGR. With
EGR, the ignition delay was longer, and the combustion phasing was
retarded for all conditions. This was attributed to the lower in-
cylinder temperature due to the dilution effect of EGR. Enhanced
airefuel mixing before ignition due to a longer ignition delay led to
a milder combustion process. Late injection timing was dominated
by premixed combustion with only one peak. The injection duration
was 7.4 CA. Advancing the injection timing led to a two-stage heat
release: the first stage was attributed to low-temperature oxidation
and the second stage was attributed to high-temperature oxidation
separated by a short delay due to the negative temperature coeffi-
cient (NTC) ignition behaviour of the mixture.
Previous researchers used EGR to improve the indicated thermal
efficiency and IMEP, as well as extend the ignition delay, thereby
controlling the combustion phasing and pressure-rise rate [23,25].
In conventional diesel combustion, EGR has been used to control
NOx emissions through charge dilution and by lowering the adia-
batic flame temperature [32e35]. In this section, an in-depth
analysis of the effects of EGR on the engine performance will be
presented. Fig. 6 shows the effect of EGR on the indicated thermal
efficiency, IMEP, and coefficient of variance of the IMEP. With
a retarded injection timing of qinj ¼ 2e10 BTDC, the indicated
thermal efficiency and IMEP without EGR was superior to cases
with 40% EGR. This was thought to be related to either fuel
quenching or over-mixing, leading to the formation of a mixture
that was too lean to allow complete combustion. However, at
qinj ¼ 15 BTDC, the indicated thermal efficiency and IMEP reached
a maximum value and then gradually decreased. The highest
indicated thermal efficiency and IMEP corresponded to a phasing
after TDC with a long ignition delay. Introducing 40% EGR improved
indicated thermal efficiency and IMEP for qinj ¼ 15e20 BTDC. This
was the same injection timing that improved performance with
different injection pressures, implying that there is optimal
170 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
Fig. 6. Effect of EGR on indicated thermal efficiency, IMEP, and coefficient of variance
of the IMEP, Pinj ¼ 140 MPa, 0% EGR, l ¼ 4.5; and Pinj ¼ 140 MPa, 40% EGR, l ¼ 3.0.
injection timing at which both the higher injection pressure and
EGR achieve improved performance. Fig. 7 shows the effects of EGR
on ignition delay. EGR had a significant impact, achieving a longer
fueleair premixing time well before ignition, which translated into
better performance and lower soot emissions. This was thought to
be directly related to the dilution effect of EGR.
PCCI combustion uses EGR to lower the adiabatic flame
temperature and oxygen concentration, resulting in controllable
combustion phasing and pressure-rise rate. This also leads to
a reduction in NOx emissions but an increase in HC and CO emis-
sions. Fig. 8 shows the effects of EGR on the maximum in-cylinder
temperature (Tmax), as well as smoke, NOx, CO, and HC emissions.
For all the tested conditions without EGR, the maximum in-cylinder
temperature was higher than with EGR. A retarded injection timing
had a lower maximum in-cylinder temperature, while advancing
the injection timing led to a gradual rise in the in-cylinder
temperature until qinj ¼ 15 BTDC, which drastically reduced in-
Fig. 7. Effect of EGR on ignition delay, Pinj ¼ 140 MPa, 0% EGR, l ¼ 4.5; and
Pinj ¼ 140 MPa, 40% EGR, l ¼ 3.0.
Fig. 8. Effect of EGR on specific emissions, Pinj ¼ 140 MPa, 0% EGR, l ¼ 4.5; and
Pinj ¼ 140 MPa, 40% EGR, l ¼ 3.0.
cylinder temperature for PCCI combustion. The in-cylinder
maximum temperature, both with and without EGR, was less than
1500 K. Excessive smoke emissions were produced for qinj ¼ 10 and
30 BTDC with 40% EGR and for an early injection timing of qinj ¼ 35
BTDC and above without EGR. This could be attributed to the low in-
cylinder temperature that did not allow complete oxidation of the
soot formed from the rich-fuel impinging on the surface of the
piston. Smoke emissions of less than 2% were achieved for
qinj ¼ 2e30 BTDC with 40% EGR and without EGR. As the injection
timing was advanced, the in-cylinder temperature decreased and
fuel penetration was enhanced, leading to fuel impingement on the
surface of the piston, providing a source of fuel-rich zones that
burned with smoke emissions. The smoke emissions of conven-
tional diesel combustion were lower than those in the PCCI regime.
PCCI combustion is not meant to completely eliminate smoke
emissions but rather reduce them to acceptable levels. Many
researchers have noted that it is possible to obtain a simultaneous
reduction in NOx and soot by incorporating large amounts of EGR
[11,35], but this comes at the expense of high fuel consumption and
increased CO and HC emissions. It was possible to achieve less than
0.5% smoke emissions with EGR at an injection timing between
qinj ¼ 15 and 25 BTDC. This was thought to arise from the fact that
the spray struck the piston bowl wall and the fueleair was premixed
well without creating fuel-rich pockets so that better combustion
was achieved. This agreed with the results obtained by Fang et al.
[36] for narrow-angle wall-guided spray combustion.
For conventional diesel combustion without EGR, NOx emissions
drastically increased as the injection timing was advanced up to
a maximum at qinj ¼ 15 BTDC, after which they decreased drasti-
cally in the PCCI combustion regime. This could be attributed to the
high in-cylinder temperature, which promotes NOx emissions, and
a short duration for fueleair mixing prior to combustion. EGR
effectively reduced NOx emissions. This was attributed to the
reduced oxygen concentration and the decrease in flame
 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175 171

Fig. 9. Sample raw and processed images.

temperature in the combustible mixture. A moderately early injec-
tion of qinj ¼ 30 BTDC and above led to low NOx emissions without
EGR due to the reduced in-cylinder temperature but also led to an
increase in smoke emissions.
As the injection timing was advanced, CO emissions increased
drastically. EGR led to higher CO emissions. Under retarded
conditions of qinj ¼ 10 BTDC with 40% EGR, the CO emissions were
high. This could be attributed to the low-temperature and insuffi-
cient oxygen for complete combustion. When the injection timing
was too early, higher emissions could also be associated with piston
surface/cylinder wall wetting.
HC emissions increased as the injection timing was advanced.
The trend was the same with and without EGR. This could be
attributed to the limited amount of oxygen available for complete
combustion. In the PCCI combustion regime of qinj ¼ 10 BTDC with
40% EGR, HC emissions were relatively high compared to the case
without EGR. This could be attributed to two effects: over-mixing,
where the mixture was too lean to burn, and flame quenching

Fig. 10. Direct visualisation of combustion images, Pinj ¼ 140 MPa, qinj ¼ 20 BTDC, 0% EGR, l ¼ 4.5; and Pinj ¼ 140 MPa, qinj ¼ 20 BTDC, 40% EGR, l ¼ 3.0.
172 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175

Fig. 11. Comparison of combustion images with and without EGR conditions (images d and C in Fig. 10).

during the expansion stroke. This result agreed with the findings of
other researchers [6,37]. To mitigate the problem, it is desirable to
increase the available air mass through supercharging while
maintaining the same EGR levels.
3.3. Soot formation and oxidation
To clarify emission trends in the test engine, the combustion
process was visualised in an optically accessible engine using
bottom-view images. In our previous study, the effects of the injec-
tion pressure and EGR on spray penetration and the combustion
process were analysed for conventional diesel combustion and PCCI
combustion [31]. The four spray plumes penetrated at an angle of 90
relative to each other, but the first onset of the flame kernel occurred
between the spray plumes. In this section, the results of an in-depth
study of the combustion process, carried out under the condition for
which simultaneous low NOx and soot emissions were achieved, are
presented. Fig. 9 shows raw and processed images. The raw images

Fig. 12. Time series of the processed red component combustion images, Pinj ¼ 140 MPa, qinj ¼ 20 BTDC, 0% EGR, l ¼ 4.5; and Pinj ¼ 140 MPa, qinj ¼ 20 BTDC, 40% EGR, l ¼ 3.0.
 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
Fig. 13. Spatially integrated flame luminosity (SIFL) and rate of SIFL, Pinj ¼ 140 MPa, 0%
EGR, l ¼ 4.5.
were captured with the high-speed video camera. The image inten-
sity before fuel injection was subtracted from the image intensity at
a specific time during combustion. This eliminated background light
so that only the magnitude of the light emitted from the flame could
be plotted. The resultant colour image was then separated into its
corresponding red, green, and blue (RGB) components. The red
component was representative of the soot-forming regions in the
Fig. 14. Spatially integrated flame luminosity (SIFL) and rate of SIFL, Pinj ¼ 140 MPa, 40% EGR, l ¼ 3.0.
173
flame and plotted as a final processed image. This was achieved
through in-house programming in MATLAB.
Fig. 10 shows the time series direct visualisation combustion
images for Pinj ¼ 140 MPa and qinj ¼ 20 BTDC, with and without
EGR, together with the in-cylinder pressure history. The upper
images correspond to the case without EGR, while the lower images
represent the case with 40% EGR. Early combustion occurred 6.1
after top dead centre (ATDC) for the case without EGR, and injection
was complete at 11.05 ATDC. There was a short premixing time.
The early flame locations were more downstream of the spray and
located near the piston bowl corner. A blue flame (images bee) was
observed during early flame development; the flame quickly filled
the bowl region uniformly due to the higher injection velocity.
Some pockets of more luminous flame were observed between the
spray directions (images c and d), but these burned out quickly late
in the combustion cycle. The spray impingement location near the
piston bowl wall was observed, but these weak local flames dis-
appeared gradually when the burning rate of the film flame was
low. The regions with high luminous flames indicated some non-
homogeneities in the airefuel mixture in the piston bowl.
With the introduction of EGR, more homogenous combustion
phenomena were observed. For this case, the injection ended near
the same CA as the case without EGR, but the first sign of a flame
appeared at about 0.5 ATDC. The early flames were located further
downstream near the piston bowl wall. The combustion was evenly
distributed in the combustion chamber. The luminosity of the flame
was significantly reduced, and the flame was more homogenous.
Because the premixing time was long, the fuel that struck the
piston bowl wall tended to premix well with air and fuel, burning
with low luminosity. At 5.6 BTDC, some weak local flames due to
fuel impingement were observed near the piston bowl; these
burned longer at a lower rate.
To clarify the differences in the combustion images with and
without EGR, a detailed investigation was carried out on two typical
images for the two conditions. Fig. 11 compares these combustion
images with and without EGR. Blue and highly luminous flames
were evident for the case without EGR, while the 40% EGR case was
characterised by blue and weak luminous flames. Low engine soot
emissions were observed for these conditions (qinj ¼ 20 BTDC, 0%
and 40% EGR).
174 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
Further analysis of the raw images was carried out to isolate the
components that represent soot formation and oxidation location.
Fig. 12 shows a time series of the processed red component
combustion images for Pinj ¼ 140 MPa and qinj ¼ 20 BTDC, both
with and without EGR, coupled with the in-cylinder pressure
history. In the case without EGR, the soot formation regions were
further downstream and located between the fuel sprays, where
they impinged and deflected inwards towards each other. This was
thought to be due to the high moment of inertia of the fuel coming
out of the injector and striking the piston bowl wall, hence
enhancing the mixing of the air and fuel, coupled with the fact that
the in-cylinder temperature at this point was about 1500 K,
promoting oxidation so that the engine soot emissions at the
exhaust pipe were a minimum. With the introduction of EGR, the
soot locations were uniformly distributed in the combustion
chamber and further downstream. Due to the long ignition delay,
the fuel and air had a longer premixing time, and hence no pockets
of fuel-rich zones formed, preventing sources of soot. This case
could burn with only a small quantity of soot formed, which was
subsequently oxidised. The images indicate that at qinj ¼ 20 BTDC
with EGR, it was possible to simultaneously achieve low engine
soot and NOx emissions.
The red component of the spatially integrated flame luminosity
(SIFL) was obtained by summing the pixel values of the bottom-
view images. Previous research demonstrated that SIFL is strongly
correlated with the engine soot emissions [36]. Figs. 13 and 14 show
the SIFL and rate of SIFL with and without EGR at Pinj ¼ 140 MPa. A
high SIFL was observed with late injection timing, while moder-
ately early injection timing led to a lower SIFL. Flame luminosity is
typically dependent on the local temperature and soot concentra-
tion. Under similar temperature distributions, higher flame lumi-
nosity indicates higher soot formation [12]. The oxidation rates
correlated well with the maximum in-cylinder temperature. Higher
temperatures led to higher soot oxidation rates, as noted at
qinj ¼ 10 BTDC without EGR, and therefore less engine soot at the
exhaust pipe. As the injection timing was advanced, the in-cylinder
temperature decreased along with the corresponding SIFL peak,
leading to a low oxidation rate and slightly more engine soot
emissions. Table 3 shows the maximum SIFL and rate of SIFL for all
the conditions considered. Cases without EGR had higher values
compared to those with EGR.
To some extent, the derivative of the flame luminosity (rate of
SIFL) showed the soot formation rate and oxidation rate during the
combustion process. In general, cases without EGR showed higher
positive peaks compared with EGR cases. This indicates faster
combustion and soot formation processes. Late injection timing
showed the highest peak for the cases considered. The negative
peaks were much higher for the late injection timing without EGR,
implying a higher oxidation rate. Cases without EGR indicated
higher in-cylinder temperatures compared to those without EGR.
Engine soot emissions for the late and moderately early injection
timings without EGR (qinj ¼ 10e30 BTDC) were less than 1%, and
the temperature was above 1400 K. This confirms that a higher in-
Table 3
Maximum SIFL and rate of SIFL for 0% EGR and 40% EGR.
Injection timing 0% EGR 40% EGR
SIFL (max) Rate of SIFL SIFL (max)
(max)
10 BTDC 9.29E þ 06 1.92E þ 06 7.19E þ 03
15 BTDC 5.23E þ 06 8.71E þ 05 1.01E þ 04
20 BTDC 1.56E þ 06 3.84E þ 05 1.11E þ 05
25 BTDC 2.28E þ 05 5.96E þ 04 6.89E þ 04
30 BTDC 1.26E þ 05 1.90E þ 04 4.62E þ 03
cylinder temperature resulted in higher soot oxidation rates. The
soot oxidation process plays an important role in determining the
engine exhaust emissions at the tailpipe. Cases with EGR resulted in
lower rates of SIFL, but the trends were similar to those without
EGR. The soot oxidation process was faster than soot formation;
hence, the engine soot emissions were also relatively low.
4. Conclusions
In this paper, the effects of spray impingement, injection
parameters, and EGR on the combustion characteristics and
exhaust emissions of a PCCI diesel engine were investigated using
a single-cylinder test engine and an optically accessible engine. Late
and moderately early injection timings were considered, leading to
the following conclusions.
Under late and early injection timings, the higher injection
pressure, Pinj ¼ 140 MPa, led to better indicated thermal efficiency
and IMEP compared to the lower injection pressure, Pinj ¼ 80 MPa.
The coefficient of variance of the IMEP was less than 3% for the
higher injection pressure. Correspondingly, smoke and HC emis-
sions were lower, while CO emissions remained relatively
unchanged. For a late injection of qinj ¼ 2e15 BTDC, NOx emissions
were higher for the lower injection pressure, but for qinj ¼ 20e40
BTDC, the higher injection pressure led to lower NOx emissions.
The use of EGR led to a simultaneous reduction in soot and NOx
emissions at the best injection timing of qinj ¼ 20 BTDC, for which
the spray struck the piston bowl wall at an optimum targeting spot,
leading to improved fueleair mixing. The indicated thermal effi-
ciency and IMEP were superior to cases without EGR but also led to
increased HC and CO emissions. EGR was effective in reducing NOx
emissions for the cases considered.
The interaction between the spray and the piston bowl geom-
etry played a key role in the airefuel mixing process. This led to low
smoke emissions of less than 1% for both the late and moderately
early injection timings without EGR. With EGR, it was possible to
achieve less than 2% smoke emissions for the same range. Both
homogenous combustion and low luminosity flames were
observed for the low soot-forming conditions. Low luminosity
flame spots were attributed to fuel spray impingement on the wall
of the piston bowl and deflection towards the space between the
sprays. qinj ¼ 20 BTDC with EGR was the optimum injection timing
that gave simultaneously low soot and NOx emissions.
References
[1] J.B. Heywood, Internal Combustion Engine Fundamentals. McGraw-Hill, Sin-
gapore, 1988.
[2] H. Zhao, HCCI and CAI Engines for the Automotive Industry. Woodhead
Publishing, Cambridge, England, 2007.
[3] B. Jayashankara, V. Ganesan, Effect of fuel injection timing and intake pressure
on the performance of a DI diesel engine - a parametric study using CFD,
Energy Conversion and Management 51 (2010) 1835e1848.
[4] S. Kook, S. Park, C. Bae, Influence of early fuel injection timings on premixing
and combustion in a diesel engine, Energy and Fuel 22 (2008) 331e337.
[5] D.M. Boot, M.C.C. Luijten, T.M.L. Somers, U. Eguz, M.T.D.D. van Erp, A. Albrecht,
G.S.R. Baert, Spray Impingement in the Early Direct Injection Premixed Charge
Compression Ignition Regime. SAE International, 2010, 2010-01-1501.
[6] M. Han, D.N. Assanis, S.V. Bohac, Sources of hydrocarbon emissions from low-
temperature premixed compression ignition combustion from a common rail
direct injection diesel engine, Combustion Science and Technology 181 (2009)
496e517.
[7] T. Fang, R.E. Coverdill, C.-f.F. Lee, R.A. White, Effects of injection angles on
combustion processes using multiple injection strategies in an HSDI diesel
Rate of SIFL
engine, Fuel 87 (2008) 3232e3239.
(max) [8] C.L. Genzale, R.D. Reitz, M.P.B. Musculus, Effects of spray targeting on mixture
2.85E þ 03 development and emissions formation in late-injection low-temperature
4.26E þ 03 heavy-duty diesel combustion, Proceedings of the Combustion Institute 32
1.54E þ 04 (2009) 2767e2774.
1.07E þ 04 [9] M. Kim, C. Lee, Effect of a narrow fuel spray angle and a dual injection
2.29E þ 03 configuration on the improvement of exhaust emissions in a HCCI diesel
engine, Fuel 86 (2007) 2871e2880.
 R. Kiplimo et al. / Applied Thermal Engineering 37 (2012) 165e175
[10] M. Boot, E. Rijk, C. Luijten, B. Somers, B. Albrecht, Uncooled EGR as a Means of
Limiting Wall Wetting under Early Direct Injection Conditions. SAE Interna-
tional, 2009, 2009-01-0665.
[11] K. Akihama, Y. Takatori, K. Inagaki, S. Sasaki, A.M. Dean, Mechanism of the
Smokeless Rich Diesel Combustion by Reducing Temperature. SAE Interna-
tional, 2001, 2001-01-0655.
[12] T. Fang, R. Coverdill, C.F. Lee, R.A. White, Influence of injection parameters on
the transition from PCCI combustion to diffusion combustion in a small-bore
HSDI diesel engine, International Journal of Automotive Technology 10 (2009)
285e295.
[13] S. Kimura, O. Aoki, Y. Kitahara, E. Aiyoshizawa, Ultra-clean combustion tech-
nology combining a low-temperature and premixed combustion concept for
meeting future emission standard, SAE Transactions - Journal of Fuels and
Lubricants 110 (2001) 2001e01-0200.
[14] K. Kawamoto, T. Araki, M. Shinzawa, S. Kimura, S. Koide, M. Shibuya,
Combination of combustion concept and fuel property for ultra-clean DI
diesel, SAE Transactions - Journal of Fuels and Lubricants 113 (2004)
2004e01-1868.
[15] H. Ogawa, N. Miyamoto, A. Sakai, K. Akao, Combustion in a Two-stage Injec-
tion PCCI Engine with Lower Distillation-temperature Fuels. SAE International
Technical Paper, 2004, 2004-01-1914.
[16] K. Lee, C. Lee, An experimental study of the extent of the operating region and
emission characteristics of stratified combustion using the controlled auto-
ignition method, Energy and Fuels 20 (2006) 1862e1869.
[17] N. Horibe, S. Harada, T. Ishiyama, M. Shioji, Improvement of premixed charge
compression ignition-based combustion by two-stage injection, International
Journal of Engine Research 10 (2009) 71e80.
[18] S. Kook, C. Bae, Combustion Control Using Two-stage Diesel Fuel Injection in
a Single Cylinder PCCI Engine. SAE Technical Paper, 2004, 2004-01-0938.
[19] S.L. Kokjohn, R.D. Reitz, Investigation of charge preparation strategies for
controlled premixed charge compression ignition combustion using a variable
pressure injection system, International Journal of Engine Research 11 (2010)
257e282.
[20] T. Li, M. Suzuki, H. Ogawa, Effect of two-stage injection on unburned hydro-
carbon and carbon monoxide emissions in smokeless low-temperature diesel
combustion with ultra-high exhaust gas recirculation, International Journal of
Engine Research 11 (2010) 345e354.
[21] Y. Murata, J. Kusaka, M. Odaka, Y. Daisho, D. Kawano, H. Suzuki, H. Ishii,
Y. Goto, Emissions suppression mechanism of premixed diesel combustion
with variable valve timing, International Journal of Engine Research 8 (2007)
415e428.
[22] T. Fang, R.E. Coverdill, C.F. Lee, R.A. White, Airefuel mixing and combustion in
a small-bore direct injection optically accessible diesel engine using
a retarded single injection strategy, Fuel 88 (2009) 2074e2082.
175
[23] Y. Aoyagi, H. Osada, M. Misawa, Y. Goto, H. Ishii, Advanced Diesel Combustion
Using of Wide Range, High Boosted and Cooled EGR System by Single Cylinder
Engine. SAE International Technical Paper, 2006, 2006-01-0077.
[24] U. Asad, M. Zheng, X. Han, G. Reader, M. Wang, Fuel injection strategies to
improve emissions and efficiency of high compression ratio diesel engines,
SAE International, Journal of Engines 1 (2008) 1220e1233 2008-01-2472.
[25] S. Kook, C. Bae, P.C. Miles, D. Choi, L.M. Pickett, The Influence of Charge
Dilution and Injection Timing on Low- Temperature Diesel Combustion and
Emissions. SAE International Technical Paper, 2005, 2005-01-3837.
[26] F. Mallamo, M. Badami, F. Millo, Effect of Compression Ratio and Injection
Pressure on Emissions and Fuel Consumption of a Small Displacement
Common Rail Diesel Engine. SAE International Technical Paper, 2005, 2005-
01-0379.
[27] S. Gan, H.K. Ng, K.M. Pang, Homogeneous Charge Compression Ignition (HCCI)
combustion: Implementation and effects on pollutants in direct injection
diesel engines, Applied Energy 88 (2011) 559e567.
[28] A.G. Gaydon, The Spectroscopy of Flames, second ed. Chapman and Hall,
London, 1974.
[29] J.E. Dec, C. Espey, Chemiluminescence Imaging of Auto-ignition in a DI Diesel
Engine. SAE Technical paper, 1998, 982685.
[30] M.M. Roy, E. Tomita, N. Kawahara, Y. Harada, A. Sakane, An experimental
investigation on engine performance and emissions of a supercharged H2-
diesel dual-fuel engine, International Journal of Hydrogen Energy 35 (2010)
844e853.
[31] R. Kiplimo, E. Tomita, N. Kawahara, S. Zhou, S. Yokobe, in: Effects of Injection
Pressure, Timing and EGR on Combustion and Emissions Characteristics of
a Diesel PCCI Engine, JSAE/SAE International, Kyoto, 2011 SAE 2011-01-1769.
[32] D. Agarwal, S.K. Singh, A.K. Agarwal, Effect of Exhaust Gas Recirculation (EGR)
on performance, emissions, deposits and durability of a constant speed
compression ignition engine, Applied Energy 88 (2011) 2900e2907.
[33] C.A. Idicheria, L.M. Pickett, Effect of EGR on diesel premixed-burn equivalence
ratio, Proceedings of the Combustion Institute 31 (2007) 2931e2938.
[34] M. Zheng, G.T. Reader, J.G. Hawley, Diesel engine exhaust gas
recirculationeea review on advanced and novel concepts, Energy Conversion
and Management 45 (2004) 883e900.
[35] A. Maiboom, X. Tauzia, J.-F. Hé tet, Experimental study of various effects of
exhaust gas recirculation (EGR) on combustion and emissions of an auto-
motive direct injection diesel engine, Energy 33 (2008) 22e34.
[36] T. Fang, C.F. Lee, Low sooting combustion of narrow-angle wall-guided sprays
in an HSDI diesel engine with retarded injection timings, Fuel 90 (2011)
1449e1456.
[37] D. Kim, I. Ekoto, W.F. Colban, P.C. Miles, In-cylinder CO and UHC imaging in
a light-duty diesel engine during PPCI low-temperature combustion, SAE
International, Journal of Fuels and Lubricant 1 (2008) 933e956 2008-01-1602.