Experimental and computational analysis of the
combustion evolution in direct-injection spark-
controlled jet ignition engines fuelled with gaseous fuels
A Boretti
1
*
, R Paudel
1
, and A Tempia
2
1
School of Science and Engineering, University of Ballarat, Ballarat, Victoria, Australia
2
Robert Bosch (Australia) Pty Ltd, Clayton, Victoria, Australia
The manuscript was received on 15 December 2009 and was accepted after revision for publication on 4 May 2010.
DOI: 10.1243/09544070JAUTO1465
Abstract: Jet ignition and direct fuel injection are potential enablers of higher-efficiency,
cleaner internal combustion engines (ICEs), where very lean mixtures of gaseous fuels could be
burned with pollutants formation below Euro 6 levels, efficiencies approaching 50 per cent full
load, and small efficiency penalties operating part load. The lean-burn direct-injection (DI) jet
ignition ICE uses a fuel injection and mixture ignition system consisting of one main-chamber
DI fuel injector and one small jet ignition pre-chamber per engine cylinder. The jet ignition
pre-chamber is connected to the main chamber through calibrated orifices and acco mmodates
a second DI fuel injector. In the spark plug version, the jet ignition pre-chamber includes a
spark plug which ignites the slightly rich pre-chamber mixture which then, in turn, bulk ignites
the ultra-lean stratified main-chamber mixture through the multiple jets of hot reacting gases
entering the in-cylinder volume. The paper uses coupled computer-aided engineering and
computational fluid dynamics (CFD) simulations to provide better details of the operation of
the jet ignition pre-chamber (analysed so far with downstream experiments or stand-alone
CFD simulations), thus resulting in a better understanding of the complex interactions between
chemistry and turbulence that govern the pre-chamber flow and combustion.
Keywords: gas engines, direct injection, jet ignition, lean-burn stratified combustion, bulk
ignition and combustion
1 INTRODUCTION
Development of more energy-efficient and environ-
mentally friendly highway transportation technolo-
gies based on heavy-duty gas engines is a key factor
for reducing fuel consumption, carbon dioxide (CO
2
)
production, and pollutants emissions within Austra-
lia, therefore improving national energy secu rity,
environment, and economy.
Up until the early 1960s, railways dominated all
but the shortest land-based freight task. Since then,
vast improvements in road vehicle productivity and
road infrastructure quality, the gradual removal of
regulations restricting road freight carriage, and the
exponential growth in interstate trade has broa-
dened the range of freight tasks for which road is
more suitable than rail. The Australian domestic
freight task measured 5.21610
14
t km in 2007, with
35 per cent carried by road [1], having roa d trains
covering most of the interstate traffic.
Australia has the largest and heaviest road vehicles
in the world, with some configurations topping out
at close to 200 t. Two-trailer road trains, or B-
doubles, are allowed in most parts of Australia, with
the exception of some urban areas. Three-trailer
road trains or B-triples operate in western New
South Wales, western Queensland, South Australia,
Western Australia, and the Northern Territory, with
the last three states also allowing AB-quads (3.5
trailers). Road trains are used for transporting all
manner of materials. Their cost-effective transport
*Corresponding author: School of Science and Engineering,
University of Ballarat, PO Box 663, Ballarat, Victoria 3353,
Australia.
email: a_boretti@yahoo.com
1241
JAUTO1465 Proc. IMechE Vol. 224 Part D: J. Automobile Engineering
has play ed a significant part in the economic devel-
opment of remote areas, with some communities
totally reliant on a regular service .
The domestic freight task has doubled in size over
the past 20 years, with an average growth of 3.5 per
cent per annum. Bureau of Infrastructure, Transport
and Regional Economics projections [1] suggest that
this trend will continue, although with slight ly
slower growth into the future, growing by approxi-
mately 3.0 per cent per annum until 2030. Over this
period, road freight volumes are projected to more
than double, with domestic demand for manufac-
tured goods sustaining much of the growth, even if
the global financial crisis will certainly dampen
freight growth in the near term. Australia’s annual
greenhouse gases emissions up to the 2009 June
quarter for energy transport amount to 89 MtCO
2
e
[2], or about 14.5 per cent of the total.
Life-cycle emissions anal ysis of alternative fuels
for heavy vehicles [3, 4] has shown the potential of
gaseous fuels for heavy-duty trucks. There have been
major advances in natural-gas engines in recent
years, which means that the present generation of
natural-gas vehicles has significantly lower emis-
sions than the present generation of diesel vehicles.
The emissions based on use in original equipment
manufacturer (OEM) vehicles are lower in all
categories, greenhouse gases, important criteria
pollutants, and air toxics. The lower particulate
matter (PM) emissions and noise levels compared
with diesel make it particularly attractive for urban
areas. The major uncertainty relates to upstream and
in-service leakage, which have already been suffi-
ciently reduced in the present generation of OEM
natural-gas vehicles, and also to the lack of sufficient
refuelling stations. The extra weight of compressed
natural-gas (CNG) fuel tanks leads to slightly higher
fuel consumption, or loss of payload in the case of
buses, but this is less of a problem with liquefied
natural-gas vehicles owing to the higher energy
density.
Similarly, a dedicated liquefied petroleum gas
(LPG) bus produces significantly lower emissions of
important criteria pollutants, and lower exbodied
emissions of greenhouse gases. Air toxics from
tailpipe emissions of LPG vehicles are much lower
than those of diesel vehicles, but the greater up-
stream emissions of air toxics causes the exbodied
emissions of air toxics from LPG to be much the
same as those from diesel. HD-5 grade LPG has a
minimum propane (C
3
H
8
) content of 90 per cent
whereas the ratio of C
3
H
8
to butane (C
4
H
10
) varies
widely in autogas LPG. When compared wi th
autogas, HD-5 grade LPG emits more nitrogen
oxides (NO
x
) but less PM. Emissions of hydrocar-
bons (HCs) are similar. The main benefit of HD-5
grade LPG compared with autogas is that the
compression ratio (CR) can be altered to suit this
higher-octane fuel. The lower PM emissions and
lower noise levels than with diesel make it attractive
for use in urban areas. The major disadvantage of
LPG is the lack of market penetration of dedicated
heavy LPG vehicles.
Improving the efficiency of internal combustion
engines (ICEs) is the most promising and cost-
effective approach to increasing vehicle fuel econ-
omy in the next 10–20 years or until the time (which
is still too far off to forecast) when plug-in hybrid
electric or fuel cell hybrid vehicles will dominate the
market [5]. Advanced combustion engines still have
a great potential for achieving dramatic energy
efficiency improvements in heavy-duty vehicle ap-
plications; the primary hurdles that must be over-
come to realize increased use of advanced combus-
tion engines are the higher cost of these engines,
requiring expensive research and development
compared with conventional engines, and compli-
ance with particularly stringent new emission reg-
ulations with catalytic emission control technologies
much less mature than gasoline engine catalysts.
Australian Standard ADR 80/03 requires compli-
ance with Euro 5 standards for heavy-duty trucks
starting 1 January 2010 for new model vehicles and 1
January 2011 for all produced vehicles. Near-future
regulations will very probably follow Euro 6 standards.
The Euro 6 regulation proposal will introduce parti-
cularly thorough emission standards [6]. Procedure
provisions will be defined for test cycles, off-cycle
emissions, particulate number, emissions at idling
speed, smoke opacity, possible introduction of a
nitrogen dioxide (NO
2
) emission limit, correct func-
tioning and regeneration of pollution control devices,
crankcase emissions, on-board diagnostic systems,
in-service performance of pollution control devices,
durability, portable emission measurement system to
verify in-use emissions, CO
2
and fuel consumption,
measurement of engine power, reference fuels, and
specific provisions to ensure correct operation of NO
x
control measures. Implementation of these stringent
emission standards is anticipated to cause a reduction
in the fuel efficiency due to the exhaust emission
control devices needed to meet emissions regulations
for both NO
x
and PM without the introduction of
advanced combustion technologies.
Advanced combustion engine technologies being
developed by the present authors are focused on
1242 A Boretti, R Paudel, and A Tempia
Proc. IMechE Vol. 224 Part D: J. Automobile Engineering JAUTO1465
ICEs fuelled with gaseous fuels, operating in ad-
vanced combustion regime s, including modes of
low-temperature combustion, which increase the
efficiency beyond those of current advanced diesel
engines and reduce engine-out emissions of NO
x
and PM to near-zero levels. In addition to advanced
combustion regimes, a reduction in the heat trans-
fer, control of the load by the quantity of fuel
injected, and a wide range of waste heat recovery
technologies are also being considered to improve
the engine efficiency further and to reduce the fuel
consumption.
2 THE ALWAYS-LEAN-BURN DIRECT-INJECTION
JET IGNITION ENGINE
The lean-burn direct-injection jet ignition (DI-JI)
engine is an ICE developed to burn gaseous and
liquid fuels more efficiently and completely within
the cylinder of a four-stroke engine. This engine uses
a fuel injection and mixture ignition system consist-
ing of one main-chamber direct-injection (DI) fuel
injector and one jet ignition (JI) pre-chamber per
engine cylinder. The aim of the system is to increase
the top brake efficiencies as well as to reduce the
efficiency penalty changing the load for a diesel-like
operation by the quantity of fuel injected enabled by
the option to burn extremely lean fuel mixtures.
The small JI pre-chamber is connected to the main
chamber through calibrated orifices and accommo-
dates another DI fuel injector. In the spark plug
version, the pre-chamber also accommodates a
spark plug which ignites a pre-chamber mixture
slightly richer than the bulk which, in turn, ignites
the ultra-lean stratif ied main-chamber mixture
through the multiple jets of hot reacting gases
entering the in-cylinder volume [7].
Figure 1 presents a sectional view of the computer-
aided design (CAD) model of the in-cylinder and JI
pre-chamber volumes on the symmetry plane of a
four-cylinder engine. The piston position is top dead
centre (TDC). Details of the injectors towards the
main-chamber and pre-chamber volumes are not
included. The main-chamber injector is located at the
centre of the combustion chamber close to the pre-
chamber nozzles. A pressure sensor for combustion
studies is also located in the centre. The pre-chamber
injector is located on top of the pre-chamber close to
the spark plug. The bowl in the piston is central to
achieve a main-chamber fuel jet close to the pre-
chamber nozzles. The JI pre-chamber has six nozzles.
It is designed to fit a standard spark plug thread of
diameter 14 mm. It accommodates one racing spark
plug of diameter 10 mm [8]. Space is left to accom-
modate one pre-chamber injector. It features six
equally spaced nozzles of diameter 1.25 mm. The pre-
chamber volume is about 1.5 cm
3
.
The fuel injection and mixture ignition system
operation is as follows. One fuel is injected directly
within the cylinder by a main-chamber DI fuel
injector operating one single injection or multipl e
injections to produce a lean stratified mixture
preferably in the bulk. This inhomogeneous mixture
is mildly lean in the inner region surrounded by air
and some residuals from the previous cycle. The
extension of the inner region may be reduced in size
to achieve mean chamber-averaged mixtures ran-
ging from slightly lean to extremely lean. Experi-
ences so far have been realized with equivalent air-
to-fuel ratios l 5 1tol 5 6.6. This mixture is then
ignited by one or more jets of the reacting gases that
issue from the small pre-chamber connected to the
main chamber via calibrated orifices, sourced from
the same or an alternative fuel that is injected into it
by a second DI fuel injector and then ignited by the
spark plug discharge. Combustion starts slightly fuel
rich in the very-small-volume pre-chamber and then
moves very quic kly to the main chamber through
one or more nozzles, with one or more jets of hot
reacting gases bulk igniting the main-chamber
mixture. The jets of reacting gases enhance combus-
tion of lean stratified mixtures within the main
chamber through a combination of high thermal
energy, multiple ignition points, and the presence of
active radical species.
With reference to the homogeneous DI or PFI and
the main-chamber spark ignition, inhomogeneous
DI and JI offer the advantages of much faster, more
complete, much leaner combustion, less sensitivity
Fig. 1 Sectional view of the in-cylinder and pre-
chamber volumes with the piston at the TDC
position
Combustion evolution in direct-injection spark-controlled jet ignition engines 1243
JAUTO1465 Proc. IMechE Vol. 224 Part D: J. Automobile Engineering
to the mixture state and composition, and reduced
heat losses to the main-chamber walls. This is
because of the better fuel distribution for the same
main-chamber-averaged fuel-to-air equivalence ra-
tio lean conditions, the combustion in the bulk of
the in-cylinder gases, the heat transfer cushion of air
between hot reacting gases and walls, the availability
of a very high ignition energy at multiple simulta-
neous ignition sites which ignite the bulk of the in-
cylinder gases, and the start of main-chamber
combustion aided by large concentrations of par-
tially oxidized combustion products initiated in the
pre-chamber which accelerate the oxidation of fresh
reactants within the main chamber.
The advantages of the system are: a higher brake
efficiency (the ratio of the engine brake power to the
total fuel energy) and therefore a reduced brake
specific fuel consumption (the ratio of the engine
fuel flowrate to the brake power) for improved full-
load operation of stationary and transport engines;
efficient combustion of variable-quality fuel mix-
tures from near stoichiometry to extremely lean for
load control (mostly throttleless) by the quantity of
fuel injected for improved part-load operation of
non-stationary engines; and the opportunity in the
ultra-lean mode to produce near-zero NO
x
in
addition to the nearly zero PM with gaseous fuels.
JI and DI are potential enablers of higher-
efficiency cleaner ICEs, where very lean mixtures of
gaseous fuels could be burned with pollutants
formation below Euro 6 levels, efficiencies ap-
proaching 50 per cent full load, and small efficiency
penalties operating part load.
The concept of the coupling of the DI of the main-
chamber fuel and bulk ignition by multipl e jets of
hot gases from a small pre-chamber fitted with a
second fuel injector and a spark plug has been
covered by many papers and patents, but only by
researchers of Watson’s group at the University of
Melbourne and by the main author of this paper [7–
21]. The concept is an original evolution of the idea
of using jet-style ignition to enable the operation of a
flame propagation engine with very lean mixtures
explored many times, mostly in the large engine
natural-gas industry; the mark ed differences in the
DI fuel injector to the main chamber created
mixtures from lean homogeneous to lean stratified
in order to explore the many options of low-
temperature combustion, and the small pre-cham-
ber fitted with a second fuel injector and a spark plug
enabled start of combustion by multiple jets of hot
reacting gases origina ting from almost stoichio-
metric ignition of a small fraction of the total fuel.
The small size of the pre-chamber (1.5 cm
3
con-
nected to a 1800 cm
3
displaced in-cylinder volume of
the 11-l truck engine considered in the applications)
is assumed to keep the amount of NO
x
produced
low, as otherwise this would be a major detriment of
traditional pre-chamber engines even at very lean
operating conditions. The spark plug life within the
pre-chamber is generally not viewed as a major issue
so long as cooling considerations are taken into
account when designing the pre-chamber housing.
However, the limited amount of space ava ilable to
accommodate a spark plug and the therefore very far
from optimal conditions may suggest replacement of
a spark plug with a much smaller glow plug [9, 22].
In the glow plug version of the ignition pre-chamber,
the spark plug is replaced by a glow plug not only to
increase durability and to reduce maintenance costs
but also to increase packaging and to avoid the
occurrence of locally fuel-rich conditions carefully
controlling auto-ignition and pre-chamber injection.
The operation of the glow plug pre-chamber is, on
the other hand, much more complicated, and some
additional studies are needed to develop the concept
further. The DI-JI engine concept is therefore a novel
idea, of particular relevance for gaseous fuels, and it
definitely has potential for further interesting studies
and developments.
3 SOME NOTES ON PRE-CHAMBER DESIGN AND
INJECTOR SELECTION
If V
PC
is the pre-chamber volume, V
MCC
is the main-
chamber combustion chamber volume, and V
D
is
the displaced volume (V
D
5 p(B
2
/4)S, where B is the
bore and S is the stroke), the true compression ratio
is
CR~
V
PC
zV
MCC
zV
D
V
PC
zV
MCC
while the reference compression ratio is
CR
~
V
MCC
zV
D
V
MCC
If g
V
is the volumetric efficiency
g
V
~
m
a
r
a, i
V
D
where m
a
is the mass of air trapped within the
cylinder when the int ake valves are closed and r
a, i
is
1244 A Boretti, R Paudel, and A Tempia
Proc. IMechE Vol. 224 Part D: J. Automobile Engineering JAUTO1465
the reference air density, then the mass of fuel to be
injected within the pre-chamber is
m
f
, PC
~
f
=
aðÞ
s
l
PC
g
V
r
a, i
CR{1ðÞV
PC
while the quantity of fuel to be injected within the
main chamber is
m
f
, MC
~
f
=
aðÞ
s
l
MC
g
V
r
a, i
V
D
{ CR{1ðÞV
PC
½
where (f/a)
s
is the stoichiometric fuel-to-air ratio
(equal to 0.0642 for C
3
H
8
, 0.0584 for methane (CH
4
),
and 0.02 94 for hydrogen (H
2
)), l
PC
is the pre-
chamber operational air-to-fuel equivalence ratio
(slightly smaller than unity), and l
MC
is the main-
chamber operational air-to-fuel equivalence ratio
(much larger than unity; values of 2.5–6.6 cover the
full-load range with negligible production of NO
x
).
g
V
is approximately unity for naturally aspirated
engines but can reach values up to 2.5 in highly
turbocharged versions with a charge cooler, because
of the recovery of the exhaust energy.
From previous relations, the main-chamber DI
fuel injector has to deliver much larger quantities of
fuel than the pre-chamber DI fuel injector does. The
injection pressures required are 200–300 bar for the
high-pressure (HP) main-chamber DI fuel injector,
and 25–50 bar for the low-pressure (LP) pre-cha mber
DI fuel injector. In the case of the 1.5-l four-cylinder
turbocharged engine, with volumetric efficiencies
approaching g
V
5 2, the amounts of C
3
H
8
and CH
4
fuels introduced within the main chamber are about
24.5–12.25 mg and 22.5–11.25 mg respectively when
running with l
MC
5 2.25–4.5. Conversely, the
amounts of fuel introduced within the pre-chamber
are roughly 2.8 mg and 2.6 mg respectively with C
3
H
8
and CH
4
fuels, i.e. 10–20 per cent of the main-
chamber fuel.
Injectors are preferably of the multi-hole type and,
if possible, operate with choked flow through
nozzles. The isobaric properties of C
3
H
8
,CH
4
, and
H
2
at a pressure of 200 bar and temperatures of 300–
400 K [23] show that C
3
H
8
is a liquid at 300–370 K,
with a density of 501– 380 kg/m
3
, and then super-
critical [23], while CH
4
and H
2
are always super-
critical with much lower densities of 155.3–98.5 kg/
m
3
and 14.4–11.06 kg/m
3
respectively. The isobaric
properties of C
3
H
8
,CH
4
, and H
2
at a pressure of
50 bar and the same temperatures of 300–400 K show
that C
3
H
8
is a liquid at 300–370 K, with a density of
531–450 kg/m
3
, and then supercritical [23], while
CH
4
and H
2
are always supercritical with much
lower densities of 35–24.6 kg/m
3
and 3.43–2.96 kg/m
3
respectively. Despite the fact that the speed of sound
is higher in H
2
than in CH
4
and in C
3
H
8
(equal to
1318.4 m/s for H
2
, 449.5 m/s for CH
4
, and 252.3 m/s
for C
3
H
8
in typical injection conditions), the mass
flowrates are lower with H
2
than with CH
4
or C
3
H
8
,
thus requiring larger flow passages.
Ideally, the main-chamber and pr e-chamber in-
jections should complete on approaching TDC,
when the fast combustion process should start,
initiated by the spark discharge. The fast-actuating
HP high-flowrate main-chamber DI fuel injector
must produce a bulk lean stratified mixture, fully jet
controlled or mixed jet–w all controlled, and injec-
tion should occur with the valves closed. Because of
the load control by the quantity of fuel injected, the
injection times may vary considerably from lean
l 5 2.5 to extremely lean l 5 5 and above. The LP
low-flowrate pre-chamber DI fuel injector is in
principle less demanding because it must introduce
a smaller amount of fuel that does not change too
much with load or speed without time constraints to
produce slightly fuel-rich conditions within the
small 1.5 cm
3
pre-chamber.
Prototype hardware has been defined so far using
normal production injectors. However, better perfor-
mances are possible by developing ad hoc solutions
for both the main-chamber and the pre-chamber DI
fuel injectors. Whereas one last-generation fast-
actuating HP gasoline direct-injection (GDI) fuel
injector could be used as the main-chamber injector
for C
3
H
8
for prototype applications where durability
and dry-run capability are not an issue (e.g. one of the
HP fast single-coil or piezo GDI injectors proposed in
references [24] and [25]), a specific injector (e.g. the
HP injector proposed in reference [26]) must be used
for the short-injection-time, high-temperature, high-
injection-pressure, high-durability, dry-run capability
required with CH
4
and moreover with H
2
.
The fast single-coil DI fuel injector reported in
reference [24] provides (with gasoline fuel) flowrates
up to 40 g/s at 200 bar, up to three multiple
injections, hydraulic separation at multiple injec-
tions of not longer than 0.2 ms, and a hollow cone
pattern with cone angles as required by the applica-
tion. The double-acting multi-hole DI fuel inject or
used in the work in reference [26] has a much larger
area, with an equivalent flow area of 0.7 mm
2
through 16 holes, and may deliver up to 20 g/s with
C
3
H
8
at 200 bar, with response times within 0.1 ms
and a minimum injection duration of 0.5 ms. This
latter inject or has also been used with H
2
in
prototype applications [27].
Combustion evolution in direct-injection spark-controlled jet ignition engines 1245
JAUTO1465 Proc. IMechE Vol. 224 Part D: J. Automobile Engineering
