Extension of Operating Window for Modern

Combustion Systems by High

Performance Ignition

Martin Brandt, Alexander Hettinger, Andreas Schneider,

Hartwig Senftleben(&), and Tim Skowronek

Robert Bosch GmbH, Blaichach, Germany

[email protected]

Abstract. Upcoming legislation motivates further development of modern

combustion systems and engine designs to achieve a reduction of fuel con-
sumption and emissions. Increased power density, charge motion and dilution
drives requirements of the ignition system. Bosch prepared a high performance
ignition device focused on high energy output and long spark duration while
providing high flexibility to select relevant performance parameters. The igni-
tion concept furthermore is designed as a plug and play solution in order to t to
established hardware interfaces and offers high energy efciency.
Based on this technology several engine tests were completed at global OEMs
and inside Bosch. The following paper summarizes the main test results which
cover extension of EGR rates, lean limit and reduction of emissions as well.
Based on this results possibilities how the benets can be used in real appli-
cations are derived and discussed.
Investigations were performed on different engines representing diverse
combustion concepts especially with regard to charge motion level. The con-
cluded tests also conrm that increased ignition performance only in combi-
nation with parameters of the combustion concept itself can lead to the desired
benets on engine level.
Besides the review of mentioned engine tests a new advanced ignition system
called CEI (Controlled Electronic Ignition) is introduced. The main properties of
the used high ignition device are described including a comparison to the visible
market trend for ignition requirements.

1 Introduction

Current development trends for combustion systems in spark ignition engine focus
mainly on exhaust gas recirculation and/or on air dilution (e.g. [24]) as possible
approach for increasing the engine efciency, Fig. 1. For homogeneous lean com-
bustion concepts enleanment needs to be strongly expanded in order to reduce raw
nitrogen oxide emissions to make this approach a valuable solution also in terms of
costs for exhaust gas after treatment.
However, increasing inert gas fraction also means more severe conditions for the
ignition and combustion system itself. The negative influence of the rise in air dilution
is reflected in the form of cyclic variations at engine operation. Therefore, in practice,

Springer International Publishing Switzerland 2017

M. Gnther and M. Sens (eds.), Ignition Systems for Gasoline Engines,
DOI 10.1007/978-3-319-45504-4_2
 Extension of Operating Window for Modern Combustion Systems 27

Fig. 1. Roadmap combustion concepts

further mixture dilution is not applicable when exceeding a smooth running limit of
about 35 % (COV, Coefcient of Variance). The disturbance in running smoothness
is mainly caused by cyclic variations in the in-cylinder flow eld, mixture homogeneity
and flame kernel formation [1, 2]. This leads to sporadic cycles with delayed com-
bustion or even misres. In order to stabilize the initial stage of combustion in diluted
mixtures a wide number of ignition concepts and systems were proposed over time,
recently leading to frequent discussions in expert groups. Figure 2 summarizes the
variety of different proposals using a main classication based on how the flame kernel
volume is activated during the ignition process. For more details about single ignition
proposals see [3, 4] or latest IAV Conference Proceedings [5, 6].

Fig. 2. Classication of ignition systems

Single spot ignition strictly means the deposition of ignition energy within a xed
location inside the combustion chamber, whereby the energy distribution is not actively
controlled by the ignition system itself. Nevertheless, a spark volume is dened by the
spark channel and even enlarged by charge motion, e.g. by deflecting the ignition spark,
which essentially leads to an activation of a flame volume. This is one of the key factors,
which yields the potential to ignite diluted mixtures using spark ignition systems.
28 M. Brandt et al.

Volumetric ignition concepts, such as corona or plasma-jet ignition actively

provide a flame kernel volume even in low charge motion conditions at ignition timing.
Despite of the tremendous variety of ignition concepts and their competition, the
TCI (Transistor Coil Ignition) is still the dominating product in series application of
passenger cars. Its simplicity, low cost and proven robustness, especially compared to
innovative, often complex systems made TCI to an unrivaled product. However, the
ignition of diluted mixtures using TCI-Systems gets more severe and the question
arises how further potential for spark ignition concepts can be accessed. To address
these challenges Bosch is about to extend the spark ignition portfolio by proposing new
high performance products as shown in Fig. 3.

Fig. 3. Spark ignition portfolio incl. extension by new high performance products XR and CEI

A straightforward method to promote the ignition process is to boost the ignition

energy by increasing spark current and/or spark duration. For conventional TCIs this
approach is typically limited as negative side effects do increase e.g. spark plug wear,
low ignitability in low spark current phases, size and weight of the ignition device and
the electrical efciency (ratio of output spark energy to input energy delivered from the
board net). Investigations showed efciencies in the range of 2030 % under real
engine conditions [7]. Due to the described conflict it is assumed a conventional TCI
can be designed for a spark energy of some 140 mJ but not signicantly higher. This is
covered by a new proposed XR coil (see Fig. 3).
To solve some of the drawbacks as listed above connected with conventional TCIs
and also to avoid the negative impact of decreasing spark current over time on flame
kernel formation a new high performance ignition approach called CEI (Controlled
Electronic Ignition) is described in this paper. This includes in a detailed way: iden-
tication of requirements for advanced combustion concepts (main aspect: highly
diluted mixtures), discussion of target requirements for high performance ignition,
review of present state of the art ignition systems, description of CEI with key
 Extension of Operating Window for Modern Combustion Systems 29

properties, summary of test results and benets for different engine/combustion types.
At the end of the article short summary and nal remarks are presented.

2 Requirements on Ignition System for Modern Combustion

Concepts

2.1 Requirements for Efcient Combustion

Depending on the combustion concept and engine operation point different require-
ments on ignition system may arise. For stratied combustion one of the key factors is
to cover the temporal and geometrical tolerances with regards to local mixture com-
position at spark plug location [8, 18]. A particular importance is assigned to the
injection process, in order to create a preferably stoichiometric mixture in the vicinity
of the spark plug. However, temporal fluctuations in the mixture formation process
need to be covered, which can be achieved by long spark durations. Similar require-
ments on the ignition system apply at catalyst heating operation, so called homoge-
neous split injection (HSP). As mentioned before, long spark durations are limited
when using conventional TCI. Even though in-house tests showed advantages of CEI
for stratied operation the focus was put on homogeneous combustion concepts to
derive requirements, since no market indication is obvious implying stratied operation
as a mainstream trend.
At full load operation main requirements arise from high gas densities at ignition
timing, leading to high ignition voltage demands of up to 40 kV. After spark break-
down the conditions to promote the flame kernel formation are more favorable than at
low load conditions, since high gas densities and high flow velocities are present,
leading to efcient energy transfer into the surrounding mixture. One straightforward
method to decrease the ignition voltage demand would be to decrease the electrode gap
distance. However, this would result in negative consequences at low load conditions
due to quenching losses, especially with higher dilution of the mixture. Here, ignition

injection system
and parameters

local mixture
composition

local flow spark

field parameters
combustion
concept, ignition
compr. ratio concept
speed, load (TCI, Corona,
ignition angle spark plug Laser,.
tumble geometry
p, T

Fig. 4. Influencing factors on flame kernel formation

30 M. Brandt et al.

voltages are rather low but the requirements on ignition system to promote the flame
kernel formation are high. Besides the influence of ignition system, the initial stage of
combustion is influenced by many other factors, see Fig. 4.
To exploit further potentials of spark ignition systems at diluted engine operation,
the interaction of all shown parameters needs to be considered. Despite global
homogeneous operation, local mixture inhomogeneity may arise at spark plug location.
These can be covered by long spark durations [16]. The local flow eld is of particular
importance, since it affects the spark deflection. High local flow velocities lead to long
plasma arcs, which help to distribute the thermal spark power in the mixture. This leads
to high energy transfer efciencies [9, 17].
Basic investigations in an optical combustion vessel showed disadvantages with
regard to maximum spark deflection when using a TCI System with its typical
decreasing shape spark current. Figure 5 shows the evolution of spark length as
function of time at local flow velocity of 10 m/s for 50 consecutive cycles (gray lines)
and two different spark current proles (left & right). The spark length was determined
by optical analysis of captured high speed images. Based on single cycles and on mean
spark length it can be observed that for current levels below approximately 4050 mA
the spark length steadily decreases. The reason for this is the steady increase of the
number of so called restrikes with the decreasing spark current. A restrike occurs if a
certain value of spark length is exceeded, leading to formation of a short circuit arc
between the plasma channels. In this case the deflection process is interrupted and the
channel length is abruptly shortened [10]. Single restrike events can be identied in the
waveforms (gray lines) showing steep spark length gradients. With regard to ignition at
engine operation it can be expected that below a certain current level the ignition
plasma remains in the immediate vicinity of the spark plug electrodes preventing good
access to fresh mixture. Furthermore, higher heat losses from the flame kernel to the
relatively cold electrodes are promoted [10]. The right picture of Fig. 5. shows the
evolution of spark length if the spark current is kept constant over the entire discharge
duration. This shape was achieved by using a laboratory ignition system. The number
of restrikes decreases signicantly, leading to stable spark deflection. Therefore,

36 240 36 240
Ispark: decreasing (TCI) Ispark= 70 mA, continious
30 Restrikes: 3.1 200 30 Restrikes: 1.8 200
Current [-mA]
Current [-mA]
Spark Lenght [mm]

Spark Lenght [mm]

24 160 24 160
Funkenstrom
Funkenstrom

18 120 18 120

12 80 12 80
Spark
Spark

6 40 6 40

0 0 0 0
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5
Zeit [ms]
Time Zeit [ms]
Time

Fig. 5. Evolution of spark length for a decreasing (left) and continuous (right) spark current for
50 cycles (gray lines) at v = 10 m/s. Red line indicates mean spark length.
 Extension of Operating Window for Modern Combustion Systems 31

an ignition system with continuous spark current level would be favorable. The
maximum spark deflection also depends on local value of flow velocity.
High flow velocities lead to long plasma arcs. However, the deflection process
cannot be sustained endlessly, since a restrike becomes more probable as the plasma
arc becomes longer. The reason for that is the increasing spark voltage (burn voltage)
with spark length, leading to high electric eld strengths between the plasma channels
[9]. Engine measurements at different operating points revealed, that restrike voltages
higher than 6 kV are less likely. However, at highest flow velocities a so called spark
blowout is possible. In such cases the spark plasma can detach from the electrodes,
therefore requiring new spark breakdown between the electrodes. The ignition system
therefore needs to establish high voltage build up and provide a new spark breakdown
within a short time after spark blowout.
image sequence

spark deflection activated flame kernel

strong cycle

after spark discharge

poor cycles
strong cycles

poor cycles

COV-hp = 13.6
a) COV-hp_opt =2.2 b)
image sequence

dissipation of flame
kernel
poor cycle

Fig. 6. (a) imep-hp as f(cycle #) at IgnAng = 42 before TDC (Top Dead Center), dashed line
separates poor cycles from strong cycles, (b) Uspark as f(t). Image sequence at the bottom:
formation of a misre. Image sequence at the top: flame kernel formation in case of strong cycle

The meaning of flow velocity at spark plug location on spark deflection and flame
kernel formation in condition of lean combustion had been discussed in recent paper by
Schneider et al. [10]. Based on optical measurements on a single cylinder research
engine it was shown, that misres and delayed combustion cycles refer to low velocity
of air-fuel-mixture in the vicinity of spark plug. Figure 6(a) shows the indicated mean
effective pressure imep-hp (high pressure loop) at lean limit ( = 1.6) for 600 cycles,
spark timing being at 42 cad before TDC (Top Dead Center). A continuous spark
32 M. Brandt et al.

current prole was used for the measurement. A limit (imep-hp = 2 bar) is dened to
indicate the delayed combustion cycles and misres (dashed line). The cycles below
this limit are classied as poor cycles. The COV-hp (COV of high pressure loop)
includes all 600 cycles, while COV-hp_opt includes only those cycles being above the
dened limit (strong cycles), see gure legend. Figure 6(b) shows the corresponding
spark voltage waveforms. Poor cycles (red lines) clearly indicate low spark voltages
and so can be referred to cycles with low flow velocities. The image sequence at the
bottom shows how a misre develops. The reduced flow leads to static arcs which in
turn lead to high quenching losses. It has to be pointed out that in such conditions,
spark ignition systems cannot release their potential, even if ignition energy is
increased, either by high spark current levels or long spark durations. According to
these ndings requirements for the ignition system and also for the Operating Point:
imep = 2 bar, n = 2000 rpm, = 1.6 [10]. Combustion system can be derived [10].
Furthermore, when evaluating the dilution limit one has to consider the temporal
phases of the combustion process, consisting of flame kernel formation (mfb0-5), rst
half of combustion (mfb5-50, duration in [cad] from 5 % till 50 % of total mass fuel
burned) and the last half of combustion (mfb50-95). The COV, often used as criterion
for evaluating ignition systems or ignition parameters contains the variation of all three
phases. Figure 7 shows the described phases of combustion obtained from pressure
trace analysis as function of in cylinder residual gas fraction.

60
CAD a. TDC [cad]

40 Ignition
ZZP Timing
(mfb50-95)
20 mfb5
(mfb5-50) mfb50
0
mfb95
-20 (mfb0-5)
-40
5 10 15 20 25 30 35 40 45 50
Residual Gas Fraction [%]

Fig. 7. Different combustion phases as function of residual gas fraction. Operating Point:
imep = 3.5 bar, n = 2000 rpm, ext. EGR sweep, High Tumble

The optimum of isfc (indicated specic fuel consumption) is obtained when the
combustion phasing (mfb50) is approximately at 68 cad a. TDC. The dilution of
mixture leads to lower laminar burning velocities and therefore longer overall burn
durations. Figure 7 shows, that most time of the combustion process is consumed for
flame kernel formation (mfb0-5) and the second half of combustion (mfb50-95). The
shortest time is consumed for mfb5-50, since this combustion phase takes place around
TDC, where, especially for high tumble concepts, highest TKEs (Turbulent Kinetic
Energies) are present. The biggest trade-off in combustion of diluted mixtures arises
between the phase mfb50-95 and mfb0-5. The best ignition performance is achieved in
case of late ignition timing (typically near TDC), where high gas temperature and high
flow velocity are present. However, retarding the ignition timing towards TDC will
 Extension of Operating Window for Modern Combustion Systems 33

shift the main combustion phase (mfb5-95) and especially the mfb50-95 as well,
leading to delayed combustion cycles. Consequently, the COV increases. The ignition
system is only capable to shorten the rst phase of combustion (mfb0-5).
However, for spark ignition systems this means that high flow velocities at ignition
timing must be present in order to achieve maximum energy transfer efciency. Since
ignition timing cannot be shifted towards TDC for the reasons stated above, the main
alternative is to shorten the burn duration (mfb5-95), by e. g. higher tumble level or
higher compression ratio [11, 23], which in consequence will allow later ignition
timing. Therefore, when evaluating the COV, attention must be laid on the whole
combustion process, especially when comparing different ignition systems. Strong
fluctuations in the main combustion phase may either mask the potential of ignition
systems or can lead to unfavorable ignition conditions in case of long burn durations. In
the latter case substantial improvements of spark ignition systems may not be possible.
Nakata et al. [11] show that increasing spark current can lead to signicant improve-
ments with regard to flame kernel growth if velocities at spark gap are increased by
engine set-up. If the above mentioned requirements for combustion system are met,
high performance spark ignition systems can provide stable flame kernel formation
even in diluted conditions. To sum up, the following requirements for spark ignition
systems can be dened:
Spark current level above 50 mA over the entire discharge duration to stabilize
spark deflection by avoiding frequently occurring restrikes in strong flow elds.
Spark duration of 3 ms at low engine speed (10002000 rpm) to provide high spark
energy level and to cover possible mixture inhomogeneity at spark plug location.
Ignition voltages of 40 kV to provide secure ignition at high loads.
The ignition system must sustain spark voltages (burn voltages) of at least 6 kV, at
conditions where high spark elongation is present.
Fast re-ignition capability in case of spark blowout, typically occurring in strong
flow elds.

2.2 Derived Component Requirements

Taking a conventional ignition transformer (so called ignition coil) as a reference, the
following requirements for a new ignition component are derived:
1. Target performance values acc. to ndings in Sect. 2.1:
(a) Spark duration: 3 ms
(b) Minimum spark current: 50 mA
(c) Secondary voltage (depending on load): 40 kV
2. Denition of spark energy and secondary voltage offer independent from each
other: Demand for both parameters has different reasons. In case of conventional
ignition coils, both parameters depend on each other since both are delivered from
the same inductive energy store. That means for a given value of stored energy the
available spark energy is influenced by the energy consumption needed to establish
34 M. Brandt et al.

the breakdown voltage at the spark plug electrodes which further depends on
cylinder pressure, spark gap and spark plug fouling.
3. Spark current trace with optimized energy content: A triangle shaped spark current
typical for conventional ignition coils represents only 50 % compared to a rect-
angular trace. Ways to increase total spark energy in conventional coils usually are
limited with respect to increased spark plug electrode wear or increasing the low
current area not contributing to an efcient combustion process.
4. Denition of spark duration and spark current level independent from each other:
acc. to above requirements, a long spark duration with a spark current above a
certain threshold enables ignition of diluted mixtures. In the same way high spark
current levels/peak values can be limited only for challenging operation conditions
of the engine and beyond that be avoided in order to minimize spark plug electrode
wear.
5. Spark stabilization: enable spark to withstand turbulences or in case of blowout
enable quick re-ignition.
Besides the requirements for secondary performance driven by the combustion
concept there are further aspects to be considered in view of mass production
engines:
6. High energy efciency: important in order to really contribute to reduce fuel
consumption and also limiting heat dissipation for high energy ignition. The latter
becomes more severe in combination with engine concepts connected with high
ambient temperatures for the ignition device.
7. Fit to established hardware and software interfaces: of course this is not an absolute
must, but seen as a success factor for acceptance and project implementation.
Reasons: ignition components are mainly handled as commodity; high require-
ments for adaptation of a complete new system will lower the chances for market
entry.
8. Small package size/high energy density: growing packing density or downsizing
trends realized in modern engine concepts will most probably not allow that
package size of the ignition device can proportionally grow to the increase of
performance output.
9. Robustness and durability: mandatory requirements anyway but efforts to safe-
guard robustness and durability are expected to be tremendous in case of com-
pletely new technological approaches due to technological challenges or risks on
component and system level.
10. Cost: Denitely one of the all-time key parameters in the highly competitive
market of automotive ignition. In terms of high performance ignition, additional
costs for this higher performance usually need to be balanced with according
benets for the combustion system and have to consider implementation efforts in
the system environment (see also requirement no. 7).
 Extension of Operating Window for Modern Combustion Systems 35

2.3 Requirements Development vs. Ignition Solutions

Conventional ignition coils clearly dominate the automotive market segment of pas-
senger cars. A rough differentiation may separate clusters acc. to spark energy
3060 mJ used in PFI applications for several decades
6590 mJ used in DI applications for some 20 years
approx. 100 mJ appearing in latest DI applications with increased requirements.
Besides these standard solutions Bosch started early to work on solutions with
higher performance output:
(a) Proposal for a high power conventional ignition coil with 140 mJ spark energy:
Over many years (since 2007) only built as a laboratory sample known as
PowerPlus and widely spread in the community as enabler for testing of new
combustion concepts due to the outstanding high spark energy. Results were
published acc. to a study [4] where PowerPlus was rated as a reference ignition
system representing conventional technology. Meanwhile Bosch started a study to
prepare industrialization of an ignition coil with 140 mJ nominal output at room
temperature, now called XR (=eXtended Range) coil. Focus here is also on
providing a reasonable high output of min. 120 mJ at high ambient temperatures.
In contrast to CEI, the XR coil can be optionally available with or without
integrated power stage/electronics. Motivation is a high number of applications
having the power stage placed in the ECU, since high ambient temperature for the
ignition device make integration of electronics difcult or even impossible.
(b) First industrialization of MSI (multi spark ignition) for lean stratied DI appli-
cations (realized in 2009): Based on a standard ignition coil regularly provid-
ing 70 mJ in single spark mode, a special closed loop control for multiple
discharge and recharge enables a signicant increase of spark duration and spark
energy output while keeping the small package size unchanged. MSI functionality
is achieved with an electronic control inside the ignition coil. Signicant fuel
consumption benet could be realized for a serial engine supported by MSI
technology [12].
Further progress in development of new high performance ignition concepts in
general can be separated in two segments:
Concepts mainly based on conventional technological approaches, extending the
basic idea of MSI but avoiding the spark interruption during the recharge phase.
Concepts with a signicant different technological approach like high frequency,
microwave, Laser or Corona ignition (see Fig. 2).
Recent publications [6] report progress in both segments with regard to high per-
formance output and according combustion benets. In view of the derived component
requirements listed in Sect. 2.2 this article is mainly focused on the rst segment of
conventional concepts. Although own evaluations have been performed including
samples from the second segment of concepts, the received results do not indicate that
these approaches will be relevant for introduction in mass production for automotive
passenger car applications within the next 5 to 10 years.
36 M. Brandt et al.

With respect to the rst segment the proposed solutions are connected more to a
potential industrialization. In order to allow a continuous, uninterrupted spark, long
spark duration and as a result scalable, high spark energy output, main proposals
include
DCI (Dual Coil Ignition) [13],
DCO (Dual Coil Offset) [14] and
CMC (Coupled Multi-Charge Ignition) [15].
All three are based on repeated recharge and discharge of an inductive store.
Therefore (similar to MSI) there is no limitation to spark duration and spark energy due
to the working principle itself. Limitation due to self-heating vs. maximum ratings of
used components is of course a relevant aspect.
Brief description of concepts:
DCI [13]: Combining two ignition coils being charged and discharged alternating
and by superposing of both high voltage outputs an uninterrupted spark is achieved.
DCO [14]: Basically similar to DCI but additionally aiming on reduction of
overlapping of secondary current traces coming from two ignition coils.
CMC [15]: Basically similar to above described concept with two ignition coils but
in addition a step-down converter and a more advanced control procedure is used to
achieve further reduction of high secondary current peak duration and reduced
voltage load for high voltage diodes.
Main pros and cons:
Pro (DCI, DCO, CMC):
technical feasibility
high and scalable secondary performance.
Con:
high spark current peaks: DCO: >0.4 A measured with sample, CMC: 0.20.4 A,
high primary current: DCO: 20 A measured with sample, CMC 3570 A,
necessity of high voltage diodes (4050 kV) for DCI and DCO to decouple
ignition coils secondary circuits
big package size and weight due to double coil design (expected to be signicantly
higher compared to CEI (Sect. 3)).
Though a lot of good results have been collected in the ignition community, the
remaining cons connected to the described solutions above motivate to look for further
solutions. A new Bosch approach is described in the following chapter.

3 CEI Working Principle and Sample Status

3.1 CEI Working Principle
Based on the set of requirements above and evaluation of alternative approaches known
from publications and own studies a new proposal for high performance ignition is
 Extension of Operating Window for Modern Combustion Systems 37

presented in the following. The essential idea for the design is the separation of the two
main functions of an ignition device, which is to provide
high breakdown voltage and
spark energy
The name CEI = Controlled Electronic Ignition is chosen according to the working
principle. The basic approach is more an evolutionary one, based on technical design
components which are in general available and already proven in other applications.
This includes:
(a) ignition coil optimized for high voltage generation,
(b) step-up converter to generate the spark energy,
(c) electronic unit to control operation of (a) and (b) triggered by ECU input.
All three functional elements are combined in one package and represent the CEI
device (Fig. 4).
The selected functional design shows the benet of high performance output, which
in terms of spark energy and spark duration is not limited due to the working principal.
Of course, in practical applications a limitation has to be considered based on heat
dissipation, component maximum rating and ambient temperature. Main features of
CEI based on the shown design and related to a standard conventional coil are (Fig. 8):

Fig. 8. Block diagram of CEI (Controlled Electronic Ignition) and interface structure suitable
for operation with established ECUs and spark plugs

Widely extended range for applicable spark energy and spark duration.
Increased ignition voltage and electrical efciency.
38 M. Brandt et al.

Flexible spark performance data and ignition voltage applicable on demand and
independent from each other.
A more detailed comparison between the performance of a standard conventional
ignition coil and CEI is summarized in Fig. 9. Especially the change of the spark
current shape by means of the functional principle of CEI leads to several benets like
high performance and the flexibility of selecting the single performance values inde-
pendent from each other. The electrical performance data are dened acc. to internal
evaluations of requirements connected with combustion concepts as discussed above
(Sect. 2.1).

Fig. 9. Performance data: CEI vs. conventional ignition coil, incl. comparison of spark current
traces (schematic view)

In addition to the secondary performance, the geometrical design parameters shown

in Table 1 are relevant for the application. Dimensions refer to the functional volume
without connectors and xing point, since these elements usually are specic to their
application.

Table 1. CEI Packaging dimensions and weight

 Extension of Operating Window for Modern Combustion Systems 39

3.2 CEI Sample Status and Performance Measurement Results

Bosch activities with CEI are running as a study based on an idea created in a
brainstorming session in 2012. Besides theoretical considerations, simulations and
measurements were started in the early beginning to verify the idea of CEI. On this
bases the development and buildup of samples were realized. Figure 10 shows the
sample history.

Fig. 10. CEI sample development and present status

Measurements of secondary performance data are presented in Figs. 11, 12 and 13.
Figure 11 starts with measurement results for the spark current showing the flexibility
to achieve very long spark duration (variation achieved by changing the control signal)
and very high spark energy values. In addition the spark current level can also be

Fig. 11. Measured CEI spark current traces and resulting spark energy levels obtained by
variation of input control signal: with CEI spark current is kept at high level for long spark
duration
40 M. Brandt et al.

Fig. 12. Measured CEI spark current and voltage traces at imep = 10 bar, n = 2000 rpm
showing behavior at (a) moderate and (b) strong turbulence with (a) stable burning spark and
(b) spark re-ignition in case of blowout due to special feature of CEI

Fig. 13. Measured CEI secondary voltage traces at laboratory load conditions as indicated

adjusted in a flexible way approx. between 60 mA and 120 mA to allow spark energy
performance on demand and avoid unnecessary spark plug wear and heat dissipation in
the CEI device itself.
As an additional feature - besides showing a reasonable good spark stability - CEI
provides the ability of quick re-ignition in case of spark blowout. This is illustrated in
Fig. 12 comparing two spark current traces at imep = 10 bar/2000 rpm, one without
(a) and one with re-ignition feature.
Finally Fig. 13 shows secondary high voltage offer of CEI acc. to different load
conditions in laboratory.
 Extension of Operating Window for Modern Combustion Systems 41

4 Engine Results

Performance tests with CEI have been evaluated on many engines over the last years, to
study the impact of continuously fed high ignition energy under different cylinder flow
respectively charge motion conditions. A summary of these results will be presented in
this section to show the bandwidth of the achieved increase of engine performance.
Tests have been carried out mainly for EGR and homogenous lean combustion
process but also for stratied lean and special catalyst heating modes. Nevertheless the
focus in this paper will be on EGR and homogenous lean combustion since EGR
dilution is already widely used and homogenous lean combustion tends to be a
promising development step for gasoline engines.
Table 2 lists the different engines used for evaluation of the CEI potential. All
engines are modern turbocharged (T/C) engines with inlet-/outlet cam-phasing
(D-VVT), direct injection system (DI) but differ in the charge motion. All engines
are equipped with central mounted injector except engine D where a side mounted
injector is applied. Mounting position of the injector has some impact on the spark plug
position, typically engines equipped with central mounted injectors have a spark plug
with a slight offset to the combustion chamber center. Even for tumble charge motion
concepts this can affect the flow eld in the vicinity of the spark plug. Furthermore, the
investigated engines follow different charge motion layouts from medium to high level.

Table 2. Engines used for CEI performance investigation

Engine Fuel system Air system Charge motion
A GDI central T/C, D-VVT + VVLin High
Full engine Int. EGR
B GDI central T/C, D-VVT Medium
Single cyl. Ext. EGR
C GDI central T/C, D-VVT Medium - High
Single cyl. Ext. EGR
D GDI side T/C, D-VVT Medium
Full engine Ext. EGR

Typically dilution by EGR respectively air is limited when the combustion stability
becomes worse which in general is described by the CoV of the mean effective pressure
(CoVpmi). A typical limit of CoV = 5 % for single cylinder engines and CoV = 3 % at
full engine application (in view of series application) is chosen. This covers the fact that
at full engine CoV is evaluated as a mean value over all cylinders. With the lower CoV
limit for a full engine it is ensured, that a single cylinder (e.g. due to cylinder-cylinder
variation of charge and flow) not exceeds values of CoV > 5 %. It is widely known
[1922] that by an increase of ignition energy combustion stability limits can be shifted
to higher dilution ratios. Current study therefore focuses on the potential of CEI
extending the dilution limits compared to conventional TCI Systems.
42 M. Brandt et al.

4.1 Potential Study for EGR Combustion Concepts

First evaluations of the CEI potential have been carried out on single cylinder engine B
with external EGR. Since engine B is equipped with a variable charge motion system
(tumble flap) it is possible to study also the effect of charge motion on combustion and
ignition. An operating point (2000 rpm/2 bari) critical for ignition has been chosen.
CEI is compared with two transformer coils, providing 90 mJ resp. 140 mJ nominal
ignition energy. The 140 mJ coil is the formerly known PowerPlus prototype from
Bosch [17]. Furthermore the 140 mJ coil was operated at different energy levels
between 96 mJ and 167 mJ.
Figure 14 shows results with active tumble flaps which would be applied for this
operating point if a variable tumble flap is installed. The application of gas exchange
analysis allowed for estimation of residual gas fraction at inlet valve closing. The
x-Axis therefore indicates the total in-cylinder residual gas fraction. Following the rule
of CoV limit = 5 % for single cylinder engines it can be observed that CEI enables
highest internal EGR rates. Compared to the nominal operation of 140 mJ, the
EGR-rate can be increased by roughly EGR = 5 %, whereas compared to the
nowadays typically standard energy level of around 90 mJ the shift of combustion limit
is even higher EGR * 8 %. Only the overloaded 140 mJ coil providing an ignition
energy of 167 mJ allows an extension of the combustion stability limit close to the CEI
performance. Nevertheless this operation would not be possible under vehicle
respectively series conditions with this coil.

Fig. 14. External EGR variation and stability limits on engine B with activated tumble flap at
2000 rpm/2 bari. x-axis indicates total in cylinder residual gas fraction.

Additional to the CoV evaluation ignition delay values (ign5) are plotted in the
lower left plot. Clearly lowest values can be observed with the CEI system, a more
detailed discussion will follow below. On the other hand it can be observed that the
effect of the investigated ignition systems on the combustion duration (BD5-95) is small
until corresponding combustion stability limits are reached. Since the spatial influence
 Extension of Operating Window for Modern Combustion Systems 43

24
cEGR
20
2000 rpm / 2bar, withoutTumble
COVpmi [%]

16 140 mj- 5.5 ms (167 mj)

140 mj- 4.6 ms (140 mj)
12
140 mj- 1.9 ms (96 mj)
8 90 mj- 1.9 ms (96 mj)
CEI
4
0
60 75
55 70
65
Ign-5 [cad]

50

BD5-95 [cad]
60
45 55
40 50
35 45
40
30
35
25 30
20 25
10 15 20 25 30 35 40 10 15 20 25 30 35 40
EGR [%] EGR [%]

Fig. 15. External EGR variation and stability limits on engine B with not active tumble flap at
2000 rpm/2 bari, x-axes indicates total in cylinder residual gas fraction.

of spark plug ignition systems is limited to a small volume (see Sect. 2.1) this behavior
is expected. More investigations have been carried out on engine B with deactivated
tumble flap. Results for the same operation point are summarized in Fig. 15. Here
combustion stability limit is reached at much lower internal EGR values EGRLi-
mit * 28 % compared to EGRLimit 32 % 38 % for the case with active tumble flap. In
addition the impact of the ignition system on stability limit is rather small. It is assumed
that this behavior relates to the rather low charge motion values in case of deactivated
tumble flaps. Spark elongation by the flow eld in the near eld of the spark plug is
very small. At such conditions a large ratio of ignition energy provided to the spark
plug is lost via the spark plug electrodes (see Chap. 2.1).
Evaluating ignition delay times (ign5) also shows on average a slower flame
kernel growth compared to the active tumble flap case. Nevertheless also here CEI
shows signicant faster flame kernel growth. This in general leads to stabilization of the
combustion process but due to the very long burn durations in case of the deactivated
tumble flaps the CoV value indicates, that the overall stability of the combustion cannot
be signicantly improved by higher ignition performance.
As mentioned at the presentation of results on engine B with active tumble flaps, a
detailed view on the combustion stability will be discussed in terms of combustion
statistic plots shown in Fig. 16. The data corresponds to the operation point with
EGR-rates * 33 % in Fig. 14. Investigations have been carried out using standard
ignition transformers (TCI) with nominal energy levels of 96 mJ respectively 140 mJ as
well with CEI. The statistic plots show the cycle individual values of indicated mean
effective pressure (high pressure loop, imep-hp) and the ignition delay dened as the
interval between ignition timing and 5 % heat release (fhr0-5). Increase of ignition
energy applied by standard transformers results only in a slight decrease of the ignition
delay whereas CEI leads to signicant reduction of ignition delay. The continuous
feeding of high energy provided by the CEI system enables a faster establishment of
44 M. Brandt et al.

Fig. 16. Combustion statistics on engine B for different ignition systems for operation point
2000 rpm/2 bari/EGR * 33 %, activated tumble flaps

the flame kernel. Assuming a constant combustion phasing (50 % heat release mfb50)
at mfb50 = 8 CA, this results in later ignition timings for the CEI system. Later
ignition timing then prot from improved ignition conditions, clearly indicating a
considerably lower scatter band width CEI.
Full engine results are shown for engine A in Fig. 17 for the operation point
1000 rpm/2 bari. Fuel consumption is plotted in terms of inlet and exhaust cam timing
position, whereby timings are given with respect to the gas exchange TDC (top dead
center). Variation of cam timings results in different internal EGR rates. Highest
internal EGR ratios are achieved with late outlet and early intake cam timings
resp. high valve overlap. Combustion stability limits for different ignition systems are
included (solid lines) as well as the series calibration point (dashed circle) as function
of valve timings.
Stability limits based on CoV < 3 % are plotted for the series coil with 90 mJ, the
140 mJ PowerPlus coil and the CEI system. In the relevant area stability limits of series
coil and PowerPlus coil behave quite similar. However CEI shifts the stability limit to
signicantly higher valve overlap which means higher internal EGR rates. Series
calibration for valve train target positions is somewhat away from the stability limit
which reflects the safety margin needed to cover tolerances in the system (engine
resp. valve train tolerances, transient behavior ). If an identical safety margin is
applied to CEI (red solid circle), the series calibration point (dashed circle) can be
shifted towards higher EGR rates respectively larger valve overlaps. Fuel economy
 Extension of Operating Window for Modern Combustion Systems 45

benet accounted for this operating point was around be = 11 g/kWh which corre-
sponds to roughly 3 %.
It should be mentioned that additional test have been carried out with a reduced
electrode gap (l = 0.2 mm). In general reduction of electrode gap leads to even worse
ignition conditions especially at low engine speed and low load, where charge motion
is weak. With CEI it was possible to maintain the stability limit found for the series coil
with the larger electrode gap. Smaller electrode gaps in general reduce the requirements
of ignition voltage demand.

Fig. 17. Cam timing variation and stability limits achieved with different ignition systems on
engine A at 1000 rpm/2 bar

Modern and especially future engines tend to have increased charge motion level.
Hence experiments have been carried out on single cylinder engine C to investigate also
the performance of CEI for higher charge motion levels on representative D-VVT
engines. Here for the operating point 2000 rpm/3 bari a variation of the external EGR
was investigated. In these tests CEI is compared against the MSI-System (Multi Spark
Ignition) and the series coil with 90 mJ. Here CEI enables a signicant shift of the
combustion stability of EGR = 12 % compared to both the MSI and the series system.
In this case the stability limit reached with CEI is in the range where fuel consumption
already starts to increase again. This is related to due to increasing burn durations.
However, also here safety margins need to be considered (shaded areas in left gure). If
the target EGR rate is at minimum fuel consumption, applying the safety margin leads to
fuel consumption benet with CEI of roughly 1 % (dashed vertical lines in right gure).
Furthermore with the increased EGR limit also NOx raw emissions are reduced but this
is not main topic for stoichiometric homogenous operation (Fig. 18).
More benchmark results based on measurements on full engine D are presented for
the operation point 1500 rpm/3.2 bari in Fig. 19. Here a comparison between
46 M. Brandt et al.

Fig. 18. External EGR variation on single cylinder engine C at 2000 rpm/3 bari; combustion
stability, fuel consumption and NOx raw emissions

a conventional transformer coil operated at two energy levels, CEI and a DCO system
is shown. DCO system has been introduced in Sect. 2.3. It offers a quite high energy of
450 mJ but also shows very high spark current pulses which need to be considered
when evaluating spark plug wear. An external EGR variation was performed to study
the influence of the ignition systems on combustion stability limit. CEI as well as DCO
enables shifting the stability limit by around EGR = 5 % applying a COV limit of
5 % but benet vanishes compared to the 80 mJ case if the full engine COV limit of
3 % is chosen.
Nevertheless CEI and DCO show a benet when evaluating which safety margin
need to be applied. This will be evaluated based on the LNV value [20]. Evaluating a
certain number of cycles (200 500) LNV is dened as:

minimep
LNV 100 % 1
meanimep

This property is of great importance for transient operation, since it reflects better
the driver feeling during acceleration. During transient operation, increased tolerance of
the cylinder charge can be observed since the gas exchange flow varies from stationary
conditions. Temperatures as well as pressures relevant for the gas exchange are away
from their equilibrium state. This can lead to EGR-rate overshoots nevertheless a
continuous evolution of the torque is expected. Typically the LNV value should drop
below 80 %.
Above presented results for EGR combustion concepts can be summarized as
follows:
Depending on engine combustion concept and load point EGR can be increased up
to EGR = 12 % compared to a nowadays ignition system ranging between 70
90 mJ.
Operating points and engine types have been observed, where the impact of a more
powerful ignition system is rather small. However, considering system tolerances
higher robustness can be achieved with CEI.
 Extension of Operating Window for Modern Combustion Systems 47

Fig. 19. External EGR variation on engine D at 1500 rpm/3.2 bari

Higher charge motion levels generally lead to higher benets, since spark is
elongated stronger and the ratio of spark energy lost via the spark plug electrodes is
reduced.
Especially at low load conditions, improved combustion stability limits enable
increased EGR rates allowing fuel economy benets typically around 1 %, up to
3 % at specic operating points and engines.

4.2 Potential Study for Lean Combustion Concepts

In the future, homogenous lean combustion concepts will play a much more signicant
role than today. Therefore investigations have been carried out evaluating the impact of
the CEI ignition system on the air dilution limits.
Since conventional 3-way catalyst does not allow sufcient high conversion rates of
NOx at lean operation conditions, optimization of NOx raw emissions to lowest values
is of great importance. The level of NOx raw emissions denes what type of NOx
aftertreatment must be applied to full emission legislation. Hence, additionally to the
design of the combustion process itself, a powerful and suited ignition system enabling
high dilution rates is a key technology to allow simple exhaust gas treatment e.g.
passive systems for lean combustion engines.
Results for lambda variation on engine B are presented in Fig. 20. Here, for CEI
different spark durations between 1.5 and 3.5 ms are chosen, a PowerPlus coil operated
at 140 mJ is used as reference. It can be found that CEI enables highest lambda
(air-fuel equivalence ratio) values. Compared to the PowerPlus coil, lambda can be
increased by = 0.1. It should be mentioned that comparing against todays coils with
around 80 mJ the benet would be even higher.
Additional study of the CEI potential for homogenous lean combustion concept
was performed on single cylinder engine C and is presented in Fig. 21. This engine has
a higher charge motion level than engine B used for the investigation shown in Fig. 20.
On engine C CEI was benchmarked against MSI and the series coil. With CEI it was
48 M. Brandt et al.

Fig. 20. Lambda variation for engine B at operating point 2000 rpm/2 bari

possible to signicantly increase the lambda limit. Compared to the other ignition
systems, a gain of = 0.17 was found and a remarkable lambda value around = 1.9
has been achieved for the operating point 2000 rpm/9 bari. Indeed, also the other
ignition systems enable an operation close to optimal fuel consumption (bi-best point)
since the bi-curve becomes quite flat in this range. Nevertheless, as described above,
NOx raw emissions are of big interest for lean combustion concepts. In this case with
CEI it was possible to reduce the NOx raw emissions by 50 % compared to series coil
as wells as the MSI system. These results conrm the basic investigations presented in
Sect. 2.1 that a continuous energy fed into the spark is benecial to achieve a long and
stable spark elongation which nally leads to improved ignition of lean mixtures.

Fig. 21. Homogenous lean combustion lambda variation on engine C at 1250 rpm/9 bari
 Extension of Operating Window for Modern Combustion Systems 49

Considering that for series calibration the target values for lambda would be chosen
with a certain safety margin to the stability limit, the impact on NOx would be even
higher since the gradient dNOx/d increase for lower lambda values.
Nearly the same benet in terms of enleanment extension as well as raw NOx
decrease was found at engine C for other operation points e.g. 2000 rpm/3 bari.
Above presented results for lean homogenous combustion concepts can be sum-
marized as follows:
Depending on engine combustion concept and operating point the lean limit can be
extended by up to = 0.17 compared to conventional TCI-Systems ranging
between 70 90 mJ and MSI.
A continuous energy fed into the spark is essential for reaching highest air dilution
values.
Even if benets in terms of fuel economy at some operation points are small, the
main focus is put on signicant reduction of the NOx raw emissions. Those sig-
nicantly determine, if a simpler/cheaper NOx aftertreatment can be chosen.

5 Summary

CEI (Controlled Electronic Ignition) as a new approach for high performance ignition
is presented; being derived from generic analysis of requirements driven by from
market and combustion trends but also considering aspects of industrialization for
automotive high volume products. CEI offers high performance up to 300 mJ spark
energy and also high flexibility to select a set of secondary performance parameters.
Further benets are high energy efciency, a balanced package size and weight
resulting in an outstanding value of energy density as ratio of spark energy per package
volume. Present samples of CEI allow operation at engine test bench and are prepared
for vehicle testing as well.
Basic requirements on ignition and combustion system were described, in order to
extend the dilution limits. Measurements in a flow combustion vessel have shown that
the decreasing spark current prole of TCI-Systems leads to frequently occurring
restrikes if the spark current level drops below 50 mA. Restrikes shorten the plasma
channel, which impede further deflection. Therefore, the current threshold must be
considered when designing high performance spark ignition systems. However, besides
electrical spark parameters the combustion system must provide adequate boundary
conditions at ignition timing. Optical measurements on a single cylinder engine
revealed, that the reduced flow velocity at ignition timing lead to inefcient energy
transfer from spark to mixture increasing the probability of misres and delayed
combustion cycles. Furthermore, attention must not only be paid to ignition but also to
the entire combustion process when evaluating the lean limits. The ignition system is
specically capable to influence the very early phase of combustion (typically up to
35 % heat release point). The acceleration of second half of combustion (mfb50-95)
is one of the key factors to extend the lean limits.
Several engine tests have been carried out on different engines to benchmark the
CEI system against standard transformer type coils as well as against the MSI system.
50 M. Brandt et al.

Here the main focus was on evaluating the benet in terms of higher dilution rates for
EGR and for homogenous lean combustion concepts.
Variations of internal and external EGR rates show in general an increased com-
bustion robustness enabling higher EGR rates. Depending on engine and reference
ignition, EGR rates could be extended with CEI by EGR 512 %. This in general
enables fuel economy benets of around 1 % but at some specic points up to 3 %.
Compared to MSI a benet was observed which supports the conclusions made in
Sect. 2.1, that a continuous feeding of high spark energy is necessary to create a stable
spark providing further elongation in the flow elds. Nevertheless, it must be mentioned
that besides the ignition system the layout of the combustion concepts (charge motion)
influences the inflammation signicantly and is limited for extended EGR rates.
Evaluations carried out to analyze the potential of CEI for homogenous lean
combustions concepts also showed promising results. Depending on engine and ref-
erence ignition system it was possible to increase the air dilution in the range of
= 0.10 0.17. While fuel consumption is already close to optimal values for the
investigated operation points with the reference ignition systems, NOx raw emissions
are further reduced by 50 % using CEI. This is of big interest since NOx raw emissions
determine which type of NOx aftertreatment becomes necessary. Hence, a powerful
ignition system as CEI is a key technology to enable homogenous lean combustion
concepts with cost optimized NOx aftertreatment.

References
1. Wenig, M.: Simulation der ottomotorischen Zyklenschwankungen. Dissertation, Universitt
Stuttgart (2013)
2. Ball, J.K.: Cycle-by-cycle variation in spark ignition internal combustion engines. Ph.D.
theses, University of Oxford (1999)
3. Kettner, M.: Technologiebewertung Zndsysteme fr magere Gemische. Forschungsvere-
inigung Verbrennungskraftmaschinen (FVV), 834 (2004)
4. Graf, J., Lauer, T.: Innovative Zndung. Forschungsvereinigung Verbrennungs-
kraftmaschinen (FVV), 1008 (2011)
5. Gnter, M., Trger, R., Ignition systems for gasoline engines. In: 1st International
Conference on Ignition Systems for Gasoline Engines, Berlin (2012)
6. Gnter, M., Trger, R.: Ignition systems for gasoline engines. In: 2nd International
Conference on Ignition Systems for Gasoline Engines, Berlin (2014)
7. Schneider, A., Hettinger A., Kufferath, A., Rottengruber, H.: Requirements on inductive
ignitions systems under engine and steady state conditions. In: 2nd International Conference
on Ignition Systems for Gasoline Engines, Berlin (2014)
8. Kiefer, A., Lippisch, M., Storch, A., Kufferath, A.: Entwicklung eines strahlgefhrten
Magerbrennverfahrens mit Mehrloch-Magnetventil-Injektoren fr turboaufgeladene Moto-
ren. 9. Tagung Diesel- und Benzindirekteinspritzung, Berlin (2014)
9. Schneider, A., Hettinger, A., Schnemann, E., Rottengruber, H.: Optische Untersuchungen
zur Flammenkernbildung im Ottomotor bei Ladungsverdnnung. 12 Internationales Sym-
posium fr Verbrennungs-diagnostik, Baden Baden (2016)
 Extension of Operating Window for Modern Combustion Systems 51

10. Schneider, A., Hettinger, A., Leick, P., Rottengruber, H.: experimental studies on spark
stability in an optical combustion vessel under flowing conditions. 3 Internationaler
Motorenkongress, Baden Baden (2016)
11. Nakata, K., Nogawa, S., Takahashi, D., Yoshihara, Y., et al.: Engine technologies for
achieving 45 % thermal efciency of S.I. engine. SAE Technical Paper 2015-01-1896
(2016)
12. Lckert, P., Doll, G., Merdes, N., Waltner, A., Eder, T.: Die Schwerpunkte der
Otto-Motoren-Entwicklung von Mercedes Benz, 19. Aachener Kolloquium Fahrzeug-
und Motorentechnik, Aachen (2010)
13. Sackmann, M., Adolf, M.; Becker, M.: High energy dual coil ignition system for gasoline
engines with EGR or lean combustion. In: SIA International Conference and Exhibition: The
Spark Ignition Engine of the Future, Strasbourg (2013)
14. Alger, T., Gingrich, J., Mangold, B., Roberts, C.: A continuous discharge ignition system for
EGR limit extension in SI engines. SAE Int. J. Engines 4(1), 677692 (2011)
15. Weyand, P., Lorenz, F., Schilling, S., Geringer, B., Hofmann, P., Graf, J., Hofherr, T.:
Adaptive continuous spark ignition as enabler for high dilution EGR operation. In: 2nd
International Conference on Ignition Systems for Gasoline Engines, Berlin, pp. 199228
(2014)
16. Anderson, R.W.: The effect of ignition system power on fast burn engine combustion. SAE
Paper, 870549, Society of Automotive Engineers, Warrendale, PA (1987)
17. Pischinger, S., Geiger, J., Neff, W., Bwing, R., Thiemann, J.: Einfluss von Zndung und
Zylinderinnenstrmung auf die ottomotorische Verbrennung bei hoher Ladungsverdnnung.
MTZ Motortechnische Zeitschrift 63(5), 388399 (2003)
18. Kufferath, A., Stroch, A., Hese, M., Tschke, H.: Potenziale des Zndsystems beim
strahlgefhrten Brennverfahren mit Piezo-Einspritztechnik. Der Arbeitsprozess des Ver-
brennungsmotors, Graz (2009)
19. Chiriac, R., Radu, R., Apostolescu, N.: A study of ignition discharge parameters effects on
homogeneous mixtures combustion in engine. SAE Paper, 2001-01-1953 (2001)
20. Chen, W., Madison, D., Dice, P., Naber, J., et al.: impact of ignition energy phasing and
spark gap on combustion in a homogenous direct injection gasoline SI engine near the EGR
limit. SAE Paper, 2013-01-1630 (2013)
21. Gukelberger, R., Alger, T., Gingrich, J., Mangold, B.: Impact of operating parameters on
ignition system energy consumption. SAE Paper, 2014-01-1233 (2014)
22. Alger, T., Gingrich, J., Roberts, C., Mangold, B., et al.: A high-energy continuous discharge
ignition system for dilute engine applications. SAE Technical Paper 2013-01-1628 (2013)
23. Ayala, F.A., Gerty, M.D., Heywood, J.B.: Effects of combustion phasing, relative air-fuel
ratio, compression ratio, and load on SI engine efciency. SAE Technical Paper
2006-01-0229 (2006)
24. Nakata, K., Shimizu, R.: Toyotas new combustion technology for high engine thermal
efciency and high engine output performance. 37. Internationales Wiener Motorensymposium
http://www.springer.com/978-3-319-45503-7