Optical analysis of the gas flame development in a RCM using a high-power ignition system

Abstract

The combustion process quality is determined by several factors: the composition of the fuel-air mixture in the vicinity of the spark plug and the discharge conditions on the spark plug. This article assesses a high-power ignition system using optical gas flame propagation analyzes. The tests were carried out in a rapid compression machine, using a fast camera for filming. The spark plug discharge quality assessment was determined indirectly by the flame propagation conditions after the ignition of the mixture (during methane combustion). The size of the flame surface and the rate of its change were assumed as a comparative criterion. It has been found that when using an ignition system with high discharge power the rate of flame development is 14% higher with respect to conventional ignition systems. In addition, the shorter development time of the early flame phase after discharge when using the new ignition system was confirmed. Based on the obtained test results and analyzes, modifications of engine operation settings were indicated, resulting from the use of a high discharge power system.

Full text

Article citation info:
FIEDKIEWICZ, Ł. PIELECHA, I. Optical analysis of the gas flame development in a RCM using a high-power ignition system.
Combustion Engines. 2018, 173(2), 47-54. DOI: 10.19206/CE-2018-208
COMBUSTION ENGINES, 2018, 173(2) 47
Łukasz FIEDKIEWICZ CE-2018-208
Ireneusz PIELECHA
Optical analysis of the gas flame development in a RCM using a high-power
ignition system
The combustion process quality is determined by several factors: the composition of the fuel-air mixture in the vicinity of the spark
plug and the discharge conditions on the spark plug. This article assesses a high-power ignition system using optical gas flame propaga-
tion analyzes. The tests were carried out in a rapid compression machine, using a fast camera for filming. The spark plug discharge
quality assessment was determined indirectly by the flame propagation conditions after the ignition of the mixture (during methane
combustion). The size of the flame surface and the rate of its change were assumed as a comparative criterion. It has been found that
when using an ignition system with high discharge power the rate of flame development is 14% higher with respect to conventional
ignition systems. In addition, the shorter development time of the early flame phase after discharge when using the new ignition system
was confirmed. Based on the obtained test results and analyzes, modifications of engine operation settings were indicated, resulting from
the use of a high discharge power system.
Key words: spark ignition, plasma ignition, optical tests, combustion process diagnostics, ignition systems
1. Introduction
The combustion process in a conventional spark-
ignition engine is triggered by the external thermal energy
generated by the electric arc. Its value and the nature of the
transfer of this energy have a fundamental impact on the
development of the flame front, and thus on the combustion
process parameters (flame development speed, heat release
rate, combustion efficiency) and the ability of the mixture
to ignite.
There are solutions in the form of an increased number
of spark plugs in the combustion chamber (Twin Spark
engine or DTSI system – Digital Twin Spark Ignition) to
increase the ignition energy and multiply the number of
ignition points in the combustion chamber volume. Thanks
to these solutions, it is possible to shorten the duration of
combustion to a minimum [3, 6]. The disadvantage of this
type of ignition systems is the necessity to use a minimum
of two spark plugs, which is not always possible due to the
space restrictions in the engine head. In addition, such sys-
tems have a greater risk of failure.
An alternative method of initiating the ignition using
a spark plug discharge is a laser system. It is characterized
by the multiple times greater energy supplied and has the
capacity of focusing it in a central point of the combustion
chamber [9]. In addition, it is possible to ignite the mixture
in several points at the same time using one "spark plug".
Unfortunately, this system is limited by the large laser costs
and the need for frequent repair of optical elements. This
results in it being used only in laboratory conditions and
prototype engines.
Camilli et al. [2] pointed out the possibility of non-
invasive modification of a conventional ignition system
using a capacitor system to increase the efficiency of the
mixture combustion process. Improved engine performance
indicators in the form of specific fuel consumption, power,
CO2 and NOx emissions, as well as the repeatability of the
engine operation cycles. Such a modernized system, in
combination with a low-resistance spark plug, allows in-
creasing the efficiency of energy transfer from 1% with
a standard solution up to 50% using a system with a larger
electrical capacity [7].
The use of capacitors contributes to the increase of the
peak current in the breakdown phase reaching a value of up
to 1000 A for 5 ns. At the same time, the power released in
the system reaches up to 5 MW. A conventional ignition
system, with the same engine operating conditions, reaches
a value of up to 100 mA, producing 0.125 W of power [8].
Jacobs et al. [4] in collaboration with a certified AVL
research center used optical analysis to demonstrate that
using a spark plug with increased peak power improves
ignition initiation and flame development, resulting in fast-
er combustion of the mixture compared to a conventional
spark plug. This is explained by the formation of a large
volume plasma between the spark plug electrodes in the
first breakdown phase [10].
2. Aim of research
Optical flame development analysis using the ignition
systems with a large, impulse delivered maximum power, is
limited in the modern literature only to the initial phase of
mixture ignition. In addition, it mainly concerns engines
powered by a stoichiometric gasoline-air mixture. The
current state of knowledge does not allow an unambiguous
assessment of the flame development during the combus-
tion of gas mixtures.
The authors of this article proposed the assessment and
comparative optical analysis of flame development during
the combustion of a stoichiometric mixture of natural gas
and air using a conventional ignition system and a system
with an increased electrical capacity.
The analysis of the results of such tests will allow to
supplement the current state of knowledge with information
on the high power maximum ignition system. These works
can significantly contribute to improving the combustion
process control, and thus obtaining more favorable opera-
tional and emission indicators for the operation of spark-
ignition engines fueled with natural gas.

Optical analysis of the gas flame development in a RCM using a high-power ignition system
48 COMBUSTION ENGINES, 2018, 173(2)
3. Research method
3.1. Test object
Experimental research was performed on a rapid com-
pression machine (RCM) that performs a piston cycle of an
internal combustion engine at defined thermodynamic con-
ditions. The choice of the test object was dictated by the
possibility of full optical access to the combustion chamber
(Fig. 1). The RCM technical parameters are shown in
Table 1.
Fig. 1. Diagram of the RCM setup [5]
Table 1. Characteristics of a rapid compression machine
Parameter
Unit
Value
Bore × stroke
mm
80 × 90
Compression ratio
–
14.7
Simulated engine speed
rpm
up to 500
Ignition system
–
spark ignition
Valves system
–
electromagnetic
Fuel system
–
direct gas-injection
(electromagnetic injector)
Air system
–
naturally aspirated
The air supply system for the piston enables obtaining
the average piston rod linear speed corresponding to the
average linear velocity of the piston of the internal combus-
tion engine (at its rotational speed of 500 rpm). Compressed
air is supplied to the chamber under the piston rod, which
expands the piston rod towards the combustion chamber.
The solenoid valves used for controlling the air inlet and
exhaust gas outlet in the RCM combustion chamber ensure
the required reaction time. The special design of the piston
rod in conjunction with a mirror and a transparent piston
crown (quartz glass) allows the phenomena occurring in the
combustion chamber to be observed.
The natural gas-air mixture is prepared using direct injec-
tion with Bosch injectors. It is ignited by a spark plug
placed centrally in the combustion chamber. The ignition
controller (produced by Mechatronics Kędzia) enables
setting the ignition advance angle and the energy discharge
value with the specified charging time of the coil. A high-
voltage ceramic capacitor connected in parallel was used to
execute the high-power ignition (Fig. 2). The order of de-
vices activating along with the time of their activation is
controlled by the microcontroller (sequencer) with the trig-
ger uncertainty equal to 1 ns.
Fig. 2. Setup diagram of a high power ignition system
3.2. Measurement apparatus
To determine the phenomena occurring in the combus-
tion chamber and to compare the ignition systems used, an
engine indication system, necessary sensors and a camera
for fast filming were used.
A LaVision camera model HSS5 that allows recording
images at a maximum frequency of 250,000 Hz was used.
The monochrome CMOS sensor used allows to record
images at a maximum resolution of 1024 × 1024 pixels.
The tests used a 5 kHz filming speed with an image size of
1024 × 624 pixels (the active filming area was 510 × 510
pixels). The camera was controlled by an external computer
with the manufacturer's software (DaVis 7.2) allowing
advanced image processing.
A piezoelectric AVL GH14D pressure sensor with
a measuring range of 0-250 bar was used to record pressure
in the RCM combustion chamber. The linear position of the
RCM piston was determined using the Megatron LSR 150
ST R5k linear potentiometer. The charging current of the
primary ignition coil as well as the discharge voltage on the
spark plug was measured using the current clamp (Pico
Technology) and the high voltage probe (Chauvin Arnoux
SHT40kV), respectively. The above signals were recorded
using the induction system – AVL IndiCom with data ac-
quisition software – AVL Concerto.
3.3. Test conditions
A comparison of the proposed methods of spark igni-
tion, i.e. a conventional solution and a high-power ignition
required the adoption of a comparative criterion. Thus, the
area of the flame in relation to the combustion chamber
volume was taken as an indicator. The second indicator
used was the flame surface increase rate over time. Due to
the method of obtaining research data (access to the com-
bustion chamber from the bottom of the piston – Fig. 1), the
flame area is understood as a flat exposure of the image (in
contrast to the spatial distribution of the flame in the com-
bustion chamber). Image analysis was performed with 15
repetitions of the combustion process carried out by a rapid
compression machine for both ignition systems.
The stoichiometric fuel mixture in the cylinder was de-
termined based on the previously performed characteristics
of the injector fuel outflow and the total volume of the
coil
control system
spark
plug
ceramic
capacitor

Optical analysis of the gas flame development in a RCM using a high-power ignition system
COMBUSTION ENGINES, 2018, 173(2) 49
RCM chamber. Before the next cycle, the volume of the
cylinder was flushed with fresh air to remove the remaining
exhaust gases. The injected gas dose was kept at qo = 32
mg. The obtained average linear velocity corresponded to
the average linear velocity of the piston of an internal com-
bustion engine with the rotational speed equal to n = 360
rpm.
The maximum charging current of the primary winding
of the ignition coil (Imax av) was 7 A, which resulted in an
average maximum voltage on the secondary winding (Umax av)
of 4.3 kV (where the distance between electrodes was d =
0.4 mm). The discharge on the spark plug was on average
10 ms before TDC (tav). The remaining test conditions are
shown in Table 2.
Table 2. Test conditions
Ignition type
conventional ignition
high discharge energy
ignition
nśr [rpm]
360
q0 śr [mg]
32
Imax śr [A]
7
Umax śr [kV]
4.6
3.9
tśr [ms before TDC]
10.6
9.6
C [pF]
0
480
R [k]
1.6
d [mm]
0.4
camera
f = 5 kHz
1024 × 624 px
4. Data selection criteria
Initial analysis of indicated data, in the form of indicat-
ed pressure, showed a high non-repeatability of the RCM
work cycles. The issue of the non-repeatability of RCM's
work is presented in [1]. The reasons for such operating
conditions of RCM are seen in the injection system (injec-
tion of gas into the combustion chamber) and the nature of
fuel injection.
Further analysis of the cylinder pressure characteristics
and the piston position allowed to determine the linear
velocity, which was characterized by a very low repeatabil-
ity between cycles. This was probably due to the instability
of the air pressure supplied to the chamber under the piston
rod, the limited air tightness of this chamber and the sole-
noid valves. The linear velocity of the piston, different for
each cycle, contributed to the differentiation in the moment
of ignition, which was performed as a function of time
alone. In addition, each ignition was carried out under dif-
ferent thermodynamic conditions (pressure, temperature)
due to different piston positions.
In the next stage of research work, a criterion allowing
the elimination of incorrect cycles was adopted. The criteri-
on was the standard deviation of pressure to the time when
the combustion took place so that the thermodynamic con-
ditions during the discharge were similar.
Standard deviation values for conventional and high-
power ignition were 1.61 and 1.21 bar, respectively (Fig.
3). The pressure characteristics were considered abnormal
if their absolute value exceeded the average pressure set for
all cycles by 1.21 bar (standard deviation value for high
power ignition – see the three-sigma rule). As a result of
such criterion, 6 out of 15 cycles for each ignition system
were eliminated as unreliable. Standard deviation pressure
value for the remaining operating cycles was 0.57 for con-
ventional ignition and 0.83 for ignition with a high maxi-
mum discharge power. Differences between mean values of
pressure at the moment of ignition for both solutions in-
cluding selected cycles did not exceed 8%.
Fig. 3. Sequences of indicated pressure P_cyl (black), mean value (red) and standard deviation (blue) for the compression curve using conventional igni-
tion. On the left before the selection, on the right after the exclusion of incorrect RCM work cycles; the characteristic peak on the pressure curve was
caused by interferences from the ignition system (t = 0 ms)
5. Optical analysis algorithms
Pictures taken using the HSS5 camera included more
than 140 ms of the RCM work cycle at 5000 fps. Such
a long recording time resulted from the uncertainty of the
moment of ignition. In order to achieve the desired results,
further processing of the photos was necessary. For this
purpose, the camera manufacturer software LaVision –
DaVis 7.2 was used. The algorithm for optical analysis is
shown in Fig. 4.
In the first stage, the number of photos was limited to
the minimum value, containing only the necessary infor-
mation. The analysis was limited to approximately 300
0
1
2
3
4
5
6
7
0
5
10
15
20
25
-48 -42 -36 -30 -24 -18 -12 -6 0 6
[bar]
P_cyl [bar]
t [ms]
0
1
2
3
4
5
6
7
0
5
10
15
20
25
-48 -42 -36 -30 -24 -18 -12 -6 0 6
[bar]
P_cyl [bar]
t [ms]

Optical analysis of the gas flame development in a RCM using a high-power ignition system
50 COMBUSTION ENGINES, 2018, 173(2)
consecutive images. Then, the background was subtracted
(the so-called reference image – a black frame). This pro-
cedure was used to eliminate noises and reflections on other
pictures.
In the next stage, a mask was used to limit the analyzed
area. The circle-shaped mask was the same size as the com-
bustion chamber and insulated the remaining area outside
the chamber (the bottom of the piston, cylinder walls).
The last stage of the analysis was to calculate the flame
area in the RCM combustion chamber. A program was
written in the internal language of the DaVis software by
the article authors. For its operation it was necessary to:
1) determine the minimum luminance value attributed to
the analyzed pixel (above the given luminance value the
pixel was treated as the flame surface);
2) determine the cylinder diameter expressed in pixels.
Images showing the beginning of an electric discharge
with different discharge powers are shown in Fig. 5.
The image sequences processed using the algorithm
shown in Figure 4 are depicted in Fig. 6.
The obtained values of the flame surface and their anal-
ysis are presented in Chapter 6.
Fig. 4. Optical analysis algorithm using DaVis 7.2
Conventional ignition
t aSOI [us]
0
200
400
600
800
t aSOI [us]
1000
1200
1400
1600
1800
High discharge power ignition
t aSOI [us]
0
200
400
600
800
t aSOI [us]
1000
1200
1400
1600
1800
Fig. 5. Images of the ignition phase start (f = 5 kHz, the first image represents SOI – start of ignition)
Calculation of the
surface area
Range selection
Difference to first
image
Applying the
mask
Min pixel value
Max pixel
value
Cylinder
diameter

Optical analysis of the gas flame development in a RCM using a high-power ignition system
COMBUSTION ENGINES, 2018, 173(2) 51
t
[ms]
Conventional
ignition
t
[ms]
Conventional
ignition
t
[ms]
High discharge
power ignition
t
[ms]
High discharge
power ignition
0
20
0
20
2
22
2
22
4
24
4
24
6
26
6
26
8
28
8
28
10
30
10
30
12
32
12
32
14
34
14
34
16
36
16
36
18
38
18
38
Fig. 6. Selected images of the combustion process using conventional ignition and high discharge power ignition (f = 5 kHz, the first image represents SOI
– start of ignition)

Optical analysis of the gas flame development in a RCM using a high-power ignition system
52 COMBUSTION ENGINES, 2018, 173(2)
6. Results
The flame area values in the RCM combustion chamber
were determined as a result of image analysis using the
DaVis software. The specific nature of the optical access to
the combustion chamber and the software application of
image masking made it possible to analyze over 1800 mm2
of the filmed combustion chamber area. The results of ana-
lyzes of all selected cycles along with the standard devia-
tion value are shown in Fig. 7. The moment of the electric
discharge between the spark plugs was taken as the point
where time = 0.
Standard deviation values of the surface area in the
flame development phase (up to 10 ms) and in the decay
phase (from 25 ms) for conventional ignition exceed the
values of this deviation when using a high-capacity igni-
tion. Alternative ignition is characterized by greater stabil-
ity and process repeatability, which is probably due to the
higher value of current flowing in the breakdown phase, in
which the ignition is initiated. In the case of a conventional
solution, the ignition may be initiated at random during one
of three stages of ignition: breakdown, arc or glow. In addi-
tion, a shorter time gap is observed between breakdown
point and the development of the flame (Fig. 8) with a mix-
ture ignited with high power ignition. The greater flame
surface area value in the combustion chamber using con-
ventional ignition in the first period after the start of the
process (up to 2 ms) result from the constraints of the creat-
ed program. The high luminance value attributed to the
pixels at the time of the electric breakdown was treated as
the flame surface. Although, in reality the higher brightness
(luminance of pixels) was caused by the visible light emit-
ted by the electric arc. Nevertheless, on Fig. 8, it is possible
to notice a greater emission of visible light during a stand-
ard ignition, which causes additional energy losses in the
system.
Fig. 7. Flame surface area as a function of time (solid line) and its standard deviation (dashed line); on the left – conventional ignition, on the right – high-
power ignition
Fig. 8. Flame surface area as a function of time (solid line) and its standard deviation (dashed line) for the early stage of flame development; on the left -
conventional ignition, on the right - high-power ignition
0
40
80
120
160
200
240
280
320
360
400
0
200
400
600
800
1000
1200
1400
1600
1800
2000
010 20 30 40 50
[mm2]
A [mm2]
t [ms]
conventional
0
40
80
120
160
200
240
280
320
360
400
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0 5 10 15 20 25 30 35 40
[mm2]
A [mm2]
t [ms]
high power
0
2
4
6
8
10
12
14
16
18
20
0
20
40
60
80
100
120
140
160
180
200
0 0.5 1 1.5 2 2.5 3 3.5 4
[mm2]
A [mm2]
t [ms]
conventional
0
2
4
6
8
10
12
14
16
18
20
0
20
40
60
80
100
120
140
160
180
200
0 0.5 1 1.5 2 2.5 3 3.5 4
[mm2]
A [mm2]
t [ms]
high power

Optical analysis of the gas flame development in a RCM using a high-power ignition system
COMBUSTION ENGINES, 2018, 173(2) 53
The mean values of the flame surface area and the time
derivative of this field inform indirectly about the flame
development rate. The results of these analyzes are present-
ed in Fig. 9. From the mean value analyzes it was found,
that the flame presence duration in the combustion chamber
for high-power ignition was reduced by about 20% in rela-
tion to the conventional ignition. In addition, the maximum
flame velocity, expressed as a derivative of the surface area
of the flame, was found to be 14% higher. These values
indicate the possibility of increasing the thermal efficiency
of the engine using a high-power ignition.
The analysis of the initial flame development phase in
a short time after the electric breakdown allowed to deter-
mine the average time delay from the electric breakdown to
the moment when the flame front begins developing. This
time was approximately 2 ms for standard ignition and 1.5
ms for high-power ignition. These values are directly refer-
enced in the voltage values on the secondary winding of the
ignition coil, where the voltage value after this time drops
to zero, thus ending the glow stage in the discharge process.
With regard to the second solution, its use in an internal
combustion engine may require correction of the ignition
advance angle.
Fig. 9. Average flame area in the combustion chamber as a function of time (solid line) and its derivative (dashed line) for the entire combustion process
(left), and early stage of flame development (right)
7. Conclusions
The analysis presented in the article concerned the opti-
cal evaluation of flame development using two ignition
methods: conventional and high discharge power, when
burning gaseous fuel (methane) using a stoichiometric
mixture.
Based on the specific comparative criteria, which were:
the flame surface area and its rate of change, the following
conclusions were formulated for ignition systems (classical
and high ignition power):
1. Based on the mean values analysis of the flame surface
averaged over many cycles, the flame duration in the
combustion chamber was found to be reduced by about
20% when using high-power ignition relative to the
conventional ignition.
2. The maximum flame velocity expressed as a derivative
of the surface area of the flame was determined to be
approximately 14% greater.
3. During the analysis of the initial flame development
phase, the average time delay between the electric
breakdown and the development of the flame front was
determined (the value of the secondary voltage in the
system was used as a confirmation). This time was ap-
proximately 2 ms for standard ignition and 1.5 ms for
high-power ignition.
The values indicated above show the possibility of in-
creasing the thermal efficiency of the engine with the use of
a high-capacity ignition.
The research presented in the article was conducted as part of the
statutory work no 05/52/DSPB/0261.
Nomenclature
A flame area
f frequency
RCM rapid compression machine
P_cyl cylinder indicating pressure
 standard deviation
t time
SOI start of ignition
U voltage
I current
qo fuel dose
n engine speed
TDC top dead centre
DTSI Digital Twin Spark Ignition
-1400
-1160
-920
-680
-440
-200
40
280
520
760
1000
1240
1480
1720
1960
2200
2440
0
200
400
600
800
1000
1200
1400
1600
1800
2000
010 20 30 40 50
dA/dt [mm2/ms]
A [mm2]
t [ms]
high power
conventional
-1400
-1160
-920
-680
-440
-200
40
280
520
760
1000
0
20
40
60
80
100
120
140
160
180
200
0 0.5 1 1.5 2 2.5 3 3.5 4
dA/dt [mm2/ms]
A [mm2]
t [ms]
high power
conventional

Optical analysis of the gas flame development in a RCM using a high-power ignition system
54 COMBUSTION ENGINES, 2018, 173(2)
Bibliography
[1] BOROWSKI, P., CIESLIK, W., PIELECHA, I.,
WISŁOCKI, K. Evaluation of the repeatability of combus-
tion process in rapid compression machine using optical re-
search. XXII International Symposium on Combustion Pro-
cesses. 22-25.9.2015, Poland.
[2] CAMILLI, L.S., GONNELLA, J.E. Improvement in spark-
ignition engine fuel consumption and cyclic variability with
pulsed energy spark plug. SAE Technical Paper. 2012; DOI:
10.4271/2012-01-1151.
[3] FORTE, C., BIANCHI, G.M., CORTI, E., FANTONI, S.
Evaluation of the effects of a Twin Spark Ignition System on
combustion stability of a high performance PFI engine. En-
ergy Procedia. 2015, 81, DOI: 10.1016/j.egypro.
2015.12.143.
[4] JACOBS, T.J., CAMILLI, L., NEUBAUER, M. High power
discharge combustion effects on fuel consumption, emis-
sions, and catalyst heating. SAE Technical Paper. 2014,
DOI: 10.4271/2014-01-2626.
[5] PIELECHA, I. Optyczne metody diagnostyki wtrysku
i spalania benzyny. Wydawnictwo Politechniki Poznańskiej.
Poznań, 2017.
[6] RAMTILAK, A., JOSEPH, A., SIVAKUMAR, G., BHAT,
S.S. Digital Twin Spark Ignition for improved fuel economy
and emissions on four stroke engines. SAE Technical Paper.
2005; DOI: 10.4271/2005-26-008.
[7] ROHWEIN, G.J. An efficient, power-enhanced ignition
system. IEEE Transactions on Plasma Science. 1997, 25, 2.
[8] ROHWEIN, G.J., CAMILLI, L.S. Automotive ignition
transfer efficiency. SAE Technical Paper. 2002; DOI:
10.4271/2002-01-2839.
[9] TSUNEKANE, M., INOHARA, T., KANEHARA, K.,
TAIRA, T. Micro-solid-state laser for ignition of automobile
engines. Advances in Solid State Lasers Development and
Applications. 2010.
[10] http://pulstarnatgas.com/testing/existence-of-plasma
Łukasz Fiedkiewicz, MEng. – Faculty of Machines
and Transport, Poznan University of Technology.
e-mail: Lukasz.M.Fiedkiewicz@doctorate.put.poznan.pl
Prof. Ireneusz Pielecha, DSc., DEng. – Faculty
of Machines and Transport, Poznan University of
Technology.
e-mail: Ireneusz.Pielecha@put.poznan.pl