MULTIPOINT SPARK IGNITION ENGINE OPERATING
ON LEAN MIXTURE
Karol Cupiał, Arkadiusz Kociszewski, Arkadiusz Jamrozik
Technical University of Częstochowa, Poland
INTRODUCTION
Experiment on multipoint spark ignition engine operating on lean mixture
was carried out in the Institute of Internal Combustion Engines and Control En-
gineering of Częstochowa University of Technology. The results of research
concerning the influence of spark plugs number and their location in SI engine
on the performance of this engine are presented in the paper. The article also
describes the numerical modelling of a combustion process performed with
KIVA-3V [2], which was compared to an experiment. This software allows
three-dimensional flows modelling in the chamber of any geometry of the IC
engines, taking also turbulence and heat transfer into consideration.
RESEARCH OBJECT
The research object was a diesel engine (S320 ER – WSW ANDORIA),
which was adapted to the spark ignition engine powered by liquid fuel. This
engine, with a 120 mm cylinder bore, had a new cylinder head designed with
four valves and multiple spark plug locations – 8 spark plugs (see Fig. 1) [4].
A carburettor was used for fuel supply and excess air factor was set to
λ = 1.2-2.0. The engine was connected to an electrical brake so the engine could
operate on constant speed 1000 rpm. The ignition system includes eight transis-
tor modules with low-resistance ignition coil per spark plug. This system allows
choosing any spark plug number and location.
The high-speed data acquisition system that was designed and constructed
at the Institute of Internal Combustion Engines and Control Engineering was
used during measurements [3, 5].
MULTIPOINT SPARK IGNITION ENGINE OPERATING ON LEAN MIXTURE 71
0.4 d
3
Exhaust Intake
valve 44 valve
7 8 2 p 1
0.8 d
Exhaust 5 Intake
valve valve
66
Fig. 1. Spark plug location legend (p – pressure transducer – Kistler model 6061B)
RESEARCH COURSE
The research was conducted in full engine load and after heat stabilisation,
which means after boiling the cooling liquid (evaporation based cooling system).
Taking the earlier research into consideration [4], measurements were done for
central spark plug (No 2), four spark plugs (No 2, 4, 5 and 8) and all spark plugs.
Data logging was performed for each one degree of crank angle for 95 consecu-
tive engine cycles with the use of specialised software [5]. Simultaneously
measurements of other quantities, which were essential for further analysis [3],
were done. Measurements were conducted for five values of excess air factor
(λ = 1.2, 1.5, 1.6, 1.8, 2.0). Richer mixtures were not applied because of knock
occurrence, which took place at λ = 1.0-1.1.
At each excess air factor, ignition advance angle was optimised in order to
achieve the best parameters of engine work. Indicated work, indicated efficiency
and non-repeatability of the following cycles were especially taken under con-
sider. Toxic components in exhaust gas concentration were measured during the
research using exhaust-gas analyser RADIOTECHNIKA AI9600.
RESULTS OF THE EXPERIMENT
IGNITION ADVANCE ANGLE OPTIMIZATION
6° CA before TDC at λ = 1.2 was the most suitable ignition advance angle
in the aspect of work, power and indicated efficiency. At λ = 1.5 and 1.6 the
most suitable ignition advance angle was 10° CA before TDC, at λ = 1.8 it was
12° CA before TDC and at λ = 2.0 it was 14° CA before TDC. With the increase
72 Karol Cupiał, Arkadiusz Kociszewski, Arkadiusz Jamrozik
of mixture depletion the ignition must occur earlier in order to achieve a correct
combustion process (Fig. 2).
356
354
Ignition advance angle (°C)
352
350
348
346
344
1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1
λ
Fig. 2. Optimal ignition angle versus excess air factor
The value of ignition advance angle on one hand is limited by engine per-
formance and on the other hand by toxic components concentration in exhaust
gas, especially nitric oxides concentration, which increases with an increase of
ignition advance angle value. In some spark plugs configurations, higher value
of ignition advance angle was favourable but a significant increase in NOx con-
centration also occurred in those cases.
MIXTURE DEPLETION AND ENGINE WORK PARAMETERS
A set of mixture depletion on test engine performance characteristics for
each combination of spark plugs location was made after the determination of
the optimum ignition advance angle values. The variation of indicated work,
indicated efficiency, combustion duration and non-repeatability cycles were
analysed. The non-repeatability factor was determined at the maximum cylinder
pressure using the following dependency:
σ p max
COV p = (1)
pmax
where:
σp – maximum cylinder pressure standard deviation,
p – mean maximum cylinder pressure.
The combustion duration was shown as an angle between the ignition mo-
ment and the time when 100% of fuel was burned. The analysis results are de-
picted in the figures below.
MULTIPOINT SPARK IGNITION ENGINE OPERATING ON LEAN MIXTURE 73
0,65 36
0,6
33
0,55
30
0,5
Li [MJ/m ]
3
η i [%]
0,45 27
0,4
Central spark plug 24
Central spark plug
0,35
Four spark plugs Four spark plugs
21
0,3 All spark plugs All spark plugs
0,25 18
1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 2,1 1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 2,1
λ [-] λ [-]
Fig. 3. Indicated work versus excess air factor Fig. 4. Indicated efficiency versus excess air factor
76 50
Central spark plug Central spark plug
Combustion duration [ CA]
66
Four spark plugs 40 Four spark plugs
o
COVp max [%]
56 All spark plugs All spark plugs
30
46
20
36
26 10
16 0
1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 2,1 1 1,2 1,4 1,6 1,8 2 2,2
λ [-] λ [-]
Fig. 5. Combustion duration versus excess air factor Fig. 6. Non-repeatability of engine cycles versus
excess air factor
From the figures it follows that an engine’s performance decreases with
a mixture depletion increase (Fig. 3) although a distinct work decrease occurs
only at high excess air factor (λ = 2.0). In this case a deterioration of engine
work parameters is caused by significantly longer combustion duration (Fig. 5)
and high cycle non-repeatability with misfire (Fig. 6). Combustion duration was
approximately 70° CA and the pressure dispersion around its mean value was up
to 43.4% (for all spark plugs).
The indicated efficiency of an engine (Fig. 4) increases at the initial range
of λ and achieves its maximum (33.5% – all spark plugs) at λ = 1.8, and then
decreases to the value of 29.1% (four spark plugs) and 19.2% (all spark plugs) at
the maximum excess air factor (λ = 2.0).
The foregoing graphs depict the very favourable setting in the aspect of en-
gine work parameters, which is the central location of four spark plugs at
λ = 1.8. The maximum indicated efficiency of 33,5% occurs in this case and
work and power decrease is low in comparison with richer mixtures.
TOXIC COMPONENTS EMISSION
Toxic components concentration of test engine exhaust gas was measured
during the research. The concentrations of nitric oxides NOx, hydrocarbons HC
and carbon oxides CO were measured.
74 Karol Cupiał, Arkadiusz Kociszewski, Arkadiusz Jamrozik
Carbon oxide concentration in the whole range of mixture depletion was at
a stable level of 0.06%. In order to present the mixture composition’s influence
on nitric oxides and hydrocarbons emission Fig. 7 was presented.
7000
Central spark plug
6000
NOx Four spark plugs
5000 All spark plugs
NOx, CH [ppm]
4000
3000
2000
CH
1000
0
1,1 1,2 1,3 1,4 1,5 1,6 1,7 1,8 1,9 2 2,1
λ [-]
Fig. 7. Mixture composition influence on NOx and HC emission
Nitric oxides concentration in exhaust gas of the test engine in the range of
λ = 1.2-1.6 does not decrease below 1000 ppm. Only at excess air factor
λ = 1.8-2.0, the nitric oxides concentration in exhaust gas is lower than 550 ppm.
The lowest concentration of this compound was obtained at the maximum mix-
ture depletion, and for one centrally located spark plug it was 160 ppm.
The significant increase of unburned hydrocarbons concentration at
λ = 2.0, is the result of a very long-lasting combustion process and high cycle
non-repeatability which occurs in this case. The highest CH emission was ob-
tained at the least favourable configuration of all spark plugs. This effect can-
not be explained on the basis of the conducted research, however, at the front
flame distribution in a cylinder unfavourable influence on combustion process
is probable.
For very lean mixtures of λ > 1.8, the four centrally located spark plugs
configuration gives the lowest CH emission of less than 800 ppm.
MODELLING OF COMBUSTION PROCESS USING KIVA-3V
Combustion process modelling was realised for two configurations of spark
plug number and location: one central spark plug (No 2) and two spark plugs
(No 2 and 8). Geometric mesh, which described the shape of combustion cham-
ber, was made in the cartesian co-ordinate system (see Fig. 9). The numerical
study was carried out for excess air factor equal to 1.8 because, according to the
experiment results, it was the most favourable setting of mixture composition.
MULTIPOINT SPARK IGNITION ENGINE OPERATING ON LEAN MIXTURE 75
The numerical simulations began at 220° CA and finished at 494° CA. This
meets the test engine valve timing (the intake valves closure and the start of
exhaust valves opening).
Fig. 8. Geometric mesh in cartesian co-ordinate system depicted for 50° CA after TDC
During the modelling attention was focused on pressure and temperature in
the cylinder and on nitric oxide evolution during the combustion process. KIVA-
3V software presents the analysis results in text files that are [Link], which
concerns pressure and temperature values and otape12, which concerns the nitric
oxide concentration in cylinder gas.
COMPARISON OF EXPERIMENTAL RESULTS WITH NUMERICAL
MODELLING RESULTS
Modelling results are shown graphically using GMV postprocessor [1].
Pressure and temperature (all cylinder volume mean values) in function of crank
angle curves were made. The results of pressure and temperature modelling and
experimental results were compared. The modelled engine nitric oxide distribu-
tion was specified in the function of crank angle. These values are actual and
averaged for all the combustion chamber volume values.
Fig. 9. Temperature distribution in TDC for central Fig. 10. Temperature distribution in TDC for two
spark plug (αignition = 348° CA) spark plugs (αignition = 348° CA)
76 Karol Cupiał, Arkadiusz Kociszewski, Arkadiusz Jamrozik
4 4
KIVA KIVA
Cylinder pressure [MPa]
Cylinder pressure [MPa]
3,2 Research engine 3,2 Research engine
2,4 2,4
1,6 1,6
0,8 0,8
0 0
200 280 360 440 520 200 280 360 440 520
Crank angle [deg] Crank angle [deg]
Fig. 11. Pressure in cylinder for central spark plug Fig. 12. Pressure in cylinder for two spark plugs
2300 2300
KIVA KIVA
1900 Research engine 1900 Research engine
Temperature [K]
Temperature [K]
1500 1500
1100 1100
700 700
300 300
200 280 360 440 520 200 280 360 440 520
Crank angle [deg] Crank angle [deg]
Fig. 13. Temperature in cylinder for central spark Fig. 14. Temperature in cylinder for two spark plugs
plug
400
300
NO [ppm]
Central spark plug
200
Two spark plugs
100
0
340 360 380 400 420 440 460 480 500
Crank angle [deg]
Fig. 15. NO distribution in cylinder versus crank angle for two configurations spark plugs number and location
Significantly faster flame propagation in case of two spark plugs results
from figures 9 and 10. Even at the TDC a greater part of charge has a tempera-
ture higher then 1500 K. The good representation of this effect is the determina-
tion of a percentage share of temperature higher than 1500 K zone in all cylinder
volume – Fig. 16. 50% share of this zone for one spark plug configuration is
achieved at 380° CA and for two spark plugs configuration for 372° CA, which
is 8° CA earlier.
In the two spark plugs configuration, the modelled pressure is higher (Fig. 12)
and its maximum (3.9 MPa) occurs earlier (375° CA) than in one spark plug con-
figuration (pmax = 2.9 MPa at 384° CA) – Fig. 11. Lower pressure values at little
differences of spark plugs configuration were obtained in a real engine. For one
spark plug the pressure is 2.5 MPa at 386° CA, and for two spark plugs it is 2.7 MPa
at 384° CA.
MULTIPOINT SPARK IGNITION ENGINE OPERATING ON LEAN MIXTURE 77
100
90 Central spark plug
80 Two spark plugs
Share of zone with temperature
70
above 1500 K [%]
60
50
40
30
20
10
0
340 350 360 370 380 390 400
Crank angle [deg]
Fig. 16. Percentage share of temperature higher than 1500 K zone in all cylinder volume
The maximal combustion temperature of the modelled engine (Fig. 13 and 14)
are at a higher level than in the experimental engine. The greatest differences are
for two spark plugs configuration, where the temperature of the simulated process
equals to 2130 K at 377° CA, and the combustion temperature calculated on the
basis of the indicated diagram [3] equals to 1830 K at 390° CA. In the case of one
spark plug, the differences are smaller because the modelled temperature is 1960 K
at 390° CA, and in the real engine the temperature is 1880 K at 395° CA.
The numerical analysis of a combustion process shows an increase of nitric
oxide concentration in exhaust gas in case of higher number of spark plugs –
Fig. 15. For this mixture composition centrally located one spark plug is charac-
terized by NO concentration of 80 ppm, and for two spark plugs the concentra-
tion is 310 ppm at 494° CA (exhaust valves opening). It is because of a higher
maximum combustion temperature values.
Nitric oxides concentration in the experimental engine exhaust gas
was measured and it was 280 ppm for one central spark plug and 400 ppm for
two spark plugs.
CONCLUSIONS
Research into SI engine powered by mixture of λ = 1.2-2.0 show that in-
crease in number of active spark plugs causes decrease of non-repeatability rate,
increase of indicated pressure and efficiency. It also shortens the combustion
duration. Significant improvement is obtained for four spark plugs (No 2, 4, 5
and 8) located centrally in combustion chamber. At this spark plug configura-
tion, excess air factor value of 1.8 allows maximal indicated efficiency of
33.5%, indicated work equals 0.552 MJ/m3 and is a little lower than 0.6 MJ/m3
obtained at richer mixtures. Non-repeatability of the following cycles with refer-
ence to the maximum pressure in the cylinder equals to 6.13%. Exhaust gas tox-
icity is at a low level.
78 Karol Cupiał, Arkadiusz Kociszewski, Arkadiusz Jamrozik
Numerical modelling of the combustion process in SI engine confirms
a good influence of applying two spark plugs in a cylinder (standard version of
KIVA does not allow modelling of more than two spark plugs). It increases
flame propagation velocity. It is important during lean mixture combustion be-
cause applying that kind of mixture causes a decrease in combustion speed and
deterioration of engine parameters. Figures 11 and 14 show a comparison of
experimental results with numerical modelling results. In two spark plugs con-
figuration the numerical model reveals higher pressures in the cylinder and
a faster increase of those pressures with relation to one spark plug configuration.
The maximum pressure angle amounts to 384° CA (central spark plug in KIVA)
and 375° CA (two spark plugs in KIVA). The discrepancy between pressures
and between temperatures during combustion can be caused by a different expo-
nent of the compression polytropic curve values, which gained higher values in
the numerical model than in a real engine. Figure 15 depicts nitric oxide concen-
tration in the cylinder of KIVA engine. It shows that an increase in fire points
number causes an increase in nitric oxide concentration.
Good results were obtained during the verification of the mathematical
model using KIVA-3V software with a real engine. Processes in experimental
engine combustion chamber were well reflected in computer simulations. Mod-
elling results and experimental results confirm a good influence of a spark plugs
number increase especially for lean mixtures burning.
REFERENCES
1. Abraham J., Magi V.: GMV, General Mesh Viewer, Los Alamos National Laboratory LA-UR-
952986, March 1995.
2. Amsden A. A.: KIVA-3V, a Block-Structured KIVA Program for Engines with Vertical or 3. Canted
Valve, Los Alamos National Laboratory LA-UR-97-689, February 1997.
3. Cupiał K.: „SILNIK” wersja 2001.5 – program do opracowywania wykresów indykatorowych,
Instytut Maszyn Tłokowych i Techniki Sterowania, Częstochowa 2002.
4. Cupiał K., Kociszewski A.: Wielopunktowy zapłon w silniku ZI, Międzynarodowa Konferencja
Motoryzacyjna AUTOPROGRES-KONMOT 2002.
5. Gruca M.: „LCTXR” – program do rejestracji i analizy sygnałów, Instytut Maszyn Tłokowych
i Techniki Sterowania, Częstochowa 2001.
SUMMARY
The results of an experiment on multipoint spark ignition engine operating on lean mixture (λ = 1.2-2.0)
were presented in the paper. The research concerned the influence of spark plugs number and location on the
following performance characteristics of this engine: the indicated work, indicated efficiency and the cycle’s
non-repeatability. The article also described a modelling of the combustion process in this engine, which was
performed with KIVA 3V. Those simulations were compared to the experiment.