Experimental Investigation of Flame Kernel in Turbulent Partial Premixed Flames

A. M. Elbaz, Mohy Mansour & Diaaeldin Mohamed
International Journal of Applied Science (IJAS), Volume (3): Issue (2): 2012 21
Experimental Investigation of Flame Kernel in Turbulent Partial
Premixed Flames
A. M. Elbaz ayman_alhagrasy@m-eng.helwan.edu.eg
Faculty of Engineering/Mechanical Power Engineering Department
Helwan University
Cairo, 11718, Egypt
Mohy Mansour mansour@niles.edu.eg
Faculty of Engineering/Mechanical Power Engineering Department,
Cairo University, Egypt
Diaaeldin Mohamed diaa.eldin@aucegypt.edu
Faculty of Engineering/Mechanical Power Engineering Department,
Cairo, University, Egypt
Abstract
The flame kernel propagation is believed to be influenced by many operating parameters such as mixing
level, turbulent intensity, and the mixture equivalence ratio. The purpose of this study is to investigate the
effect of the mixture equivalence ratio and turbulence intensity on the flame kernel and flow field interlinks
in partially premixed natural gas flames. Three jet equivalence ratios of 1, 1.5, and 2 are considered at
values of jet velocities in the range from 10 to 20 m/s. This study was done under constant degree of
partial premixing. A pulsed Nd: YAG laser is used for the flame ignition, and the turbulent flow field is
captured at several time intervals from ignition using two-dimensional Planar Imaging Velocimetry (PIV).
The mean flow field doesn’t influence with the flame kernel propagation. The turbulent flow field indicates
an increase in the global turbulence intensity in flames associated with the kernel propagation in
comparison with the isothermal case. The jet equivalence ratio of one enhances the flame kernel
propagation and it gives the highest rate of kernel propagation. Increasing the jet equivalence ratio to 1.5
and 2 reduces the intensity of chemical reaction and hence the effect of turbulence becomes the
dominant factor effecting the propagation of the flame kernel .At jet velocity of 20 m/s , an early flame
kernel extinction is recorded without any respect to jet equivalence ratio. At the early stage of the kernel
generation at delay time of 150 µs, linear correlation between the jet velocity and the kernel propagation
is noticed. The chemical reaction is the main factor influences the rate of kernel propagation; it gives
nearly 3.5 times the effect of the flow convection to the maximum rate of the flame kernel propagation at
jet velocity of 20 m/s and equivalence ratio of one.
Keywords: Flame kernel, Partial premixed flame, PIV, Flow field.
1. INTRODUCTION
The early phase of combustion in spark ignited combustion systems affects the flame propagation and
stability, and hence the performance, of the combustion process and the system efficiency. The
developing flame kernel represents this phase and is affected by many parameters, such as spark
energy, rate of energy release, turbulent flow field and the fuel/air mixing. Previous studies have shown
that variations in the initial growth of the flame kernel contribute significantly to cycle-to-cycle variation in
engine performance and emissions [1]. The flame kernel has attracted many experimental research
groups [2, 3-5] and DNS research groups [4, 6-9] interested in studying flame kernel evolution in turbulent
environmental and the main factors that control its characteristics and propagation. Many parameters
have been investigated to study their effects on the flame kernel characteristics and propagation, e.g.,
flame shape, wrinkling and curvature. In the following section a brief review of these studies and their
findings are presented and discussed.

Katta et al. [10] reported an experimental and numerical investigation using a unique counter flow
diffusion flame with an embedded vortex generator. This study was to understand the local quenching
process associated with the vortex-flame interaction in methane diffusion flames. The results show that,
the high increase in CH3 radicals in the strained flame zone depleted the radical pool (such as OH, H,
and O) and, hence, the flame is quenched locally. They concluded that this quenching process is different
from the quenching observed in steady counter flow flames, where the quenching was due to the gradual
reduction in temperature with increasing strain rate. Renard et al. [11] investigated experimentally (using
OH PLIF) the flame front of a non-pre-mixed flame interacting with a vortex to study the heat release,
extinction and time evolution of the flame surface. They concluded that global intensification or extinction
of the flame is characterized by an increase or decrease in flame surface area because of straining.
Marley et al. [12] investigated experimentally the interaction of spark-ignited flame kernels with a laminar
vortex. They concluded that burning rates for rich flames were decreased, with total flame kernel
extinction occurring in extreme cases. They also proved that small flame kernel–vortex interactions are
dominated by transient stretch effects and thermo-diffusive stability, in agreement with premixed flame
theory. Furthermore, they pointed out that vortex interactions with larger methane–air flame kernels led to
slight flame speed enhancements for both lean and rich flame kernels.
Furthermore, Arcoumanis et al. [13] showed that using a small quantity of rich mixture injected near the
spark gap can yield to formation of a stable and consistent flame kernel after spark ignition. The function
of this local variation in equivalence ratio is to support the flame in a mixture with an overall equivalence
ration as low as 0.39. They concluded that not only the average flame speed could be increased by local
injection at all equivalence ratios [14] but also the fluid dynamic effect alone caused overstretching of the
flame for the ultra lean homogeneous conditions, while rich local stratification in the vicinity of the spark
allowed the suppression of this effect and a reduction of the drivability limit [15]. Roberts et al. [16-18]
investigated the added complexity of flame kernel–vortex interactions compared to some earlier planar
flame–vortex configuration. They observed experimentally the initial flame kernel–vortex interactions
using OH-PLIF in a lean atmospheric-pressure methane–air flame kernel to determine the degree of
flame wrinkling and the ability of vortices of varying size and strength to globally quench combustion
[16].The disturbed flame existed in either the flamelet regime (which has a continuous reaction zone) or
distributed reaction zone regime (which has a coexistence of reactants and products) is strongly
dependent upon: the vortex size, the vortex strength, and the time of the initial flame–vortex interaction.
They investigated the global extinction of the flame kernel with large vortex sizes interacting with small
flame kernels. In addition to that, they concluded that the localized flame front extinction occurred for a
range of vortex sizes and strengths and a range of flame kernel sizes.
Meanwhile, Thevenin et al. [19] and Renard et al. [20] reported an experimental and numerical work of
non-premixed flame interacting with a vortex. They investigated the influences of global mixture ratio and
vortex velocity on changes in the flame surface. Their study concluded that straining effects are
responsible for the extinction of the non-premixed flame front, and the degree of mixing actually increases
at the end of the extinction process. They were able to observe the fuel pocket formation, evolution and
consumption are another important phenomenon during flame-vortex interactions. In addition to that, the
causes for local flame quenching were reported by (Patnaikand Kailasanath [21]) as well. They concluded
that these causes happen due to simultaneously high strain and the heat losses in flame-vortex
interactions of lean methane/air mixture. Harinath Reddy and John Abraham [22] studied the outcomes of
interactions of counter-rotating vortex pairs with developing ignition kernels. They examined quantitatively
and qualitatively the evolution of flame surface area during kernel–vortex interaction. They pointed out
that flame development is accelerated and the net flame surface area growth rate increased with
increasing vortex velocity, besides they proved that increasing the vortex length scale increases the
surface growth rate. They noticed when the vortex velocity is high relative to the flame speed, the vortex
breaks through the ignition kernel carrying with it hot products of combustion, which accelerates growth of
the flame surface area and heat release rates compared to a kernel with no vortex interaction.
Recently, experimental work on the flame kernel structure and propagation in a high turbulent premixed
methane flow was performed by Mansour et al. [23] using combined two-dimensional Rayleigh and LIPF-
OH techniques. They generated the spark of ignition using pulsed Nd: YAG laser. They investigated four
flames at two equivalence ratio of 0.8, and 1, and jet velocity of 6, and 12 m/s.

They showed that the flame kernel structure starts with spherical shape, and then changes gradually to
peanut-like, then changes to mushroom-like and finally the turbulence effectively distributes the kernel.
They concluded that the trends of the flame propagation, flame radius, flame cross section area, and
mean flame temperature are correlated to the jet velocity and equivalence ratio, also lean flames
propagate faster.
Based on this review of previous researchs, it is notable that the interaction between the flow field and
flame kernel propagation hasn’t been investigated experimentally in high turbulent partially premixed
flames. Thus, the aim of the present work is to investigate the flow field associated with flame kernel
propagation history in partial premixing turbulent flames. The experimental program is devoted to study
the effect of the jet equivalence ratio and turbulence intensity on the flame kernel and flow field interlinks
in partially premixed natural gas flames. The mean flow field and turbulence intensity are measured using
two-dimensional Planar Imaging Velocimetry (PIV), in terms of mean axial velocity and rms. The flow field
is first captured for the isothermal field without ignition accompanied, and this could be used as a
reference flow field to those flow fields with ignition. The flow field with ignition is recorded after the start
of ignition at different delay times. Therefore, the flame kernel-turbulent flow field interaction could be
interpreted by the comparison of flow fields of the isothermal and ignited cases.
2. EXPERIMENTAL TECHNIQUE
As shown in Fig.1a, the burner consists of two vertical concentric stainless steel tubes of 6 mm and 10
mm inner diameters and tip thickness of 1 mm. Air is passing through the inner tube while the fuel
(natural gas of 95% CH4 by volume) flows through the annular passage between the inner and outer
tubes. The inner tube exit is lower than the exit of the outer tube by a distance L. This distance can be
varied to generate different degree of partial premixing. Mixing starts at the exit of the inner tube and
continues downstream along the premixing distance L. The burner is sitting at the top of conical
turbulence generator as shown in Fig. 1a which is similar to Videto and Santavicca [24] turbulent
generator, the flow passes through a narrow slit at a diameter of 45 mm and a slit thickness b = 0.8 mm,
followed by an inverted cone with 52
o
base cone angle. The coming flow from the slit provides a ring-like
cylindrical shape, which is broken at the inner cone wall to generate a wide range of eddies, which leads
to higher turbulence intensity. The turbulence intensity due to this burner concept can be as high as 25%
[24]; this is confirmed by the turbulent flow field measurement at the burner exit. The current Planar
Imaging Velocimetry technique, explained in details below, requires to image seeding particles within the
flow. However, due to the narrow slit dimension of the turbulence generator, it isn’t possible to pass all the
seeded flow through the slit. Therefore, 10 % of the total air flow rate is used to carry the seeding
particles and passes through a central concentric tube with the turbulent generator disk, as shown in Fig.
1a, while the rest of the air flow rate is passed through the slit. Then both streams are mixed inside the
cone. Titanium dioxide with a mean diameter of 0.5 µm is used as seeding particles through a fluidized
bed seeder.
The interactions between the flame kernel and flow field is investigated at several delay time intervals
from ignition, under different jet equivalence ratio and jet velocity. This is based on capturing the turbulent
flow field after the ignition at different delay time, namely 150, 300, 500, 1000, 1500, and 2500 µs. These
flow fields are compared with flow field without ignition (henceforth to be termed isothermal flow) to
identify the effects of the evolving flame kernel. A pulsed Nd:YAG laser (Continuum Surelite II) is used for
flame ignition, see Fig.(1-b). Laser ignition provides more stable measurable pulse energy, and overcome
the problems of the effects of spark electrodes on the flame structure, as well as heat loss to the
electrodes. The laser provides a beam of 6 mm diameter at 1064 nm with 230 mJ and is focused to a
beam of waist radius of 5.6 µm using spherical lens of 50 mm focal length.
As shown in Fig.(1-b), the flow field is measured using two-dimensional Particle Imaging Velocimetry
(PIV) technique, of two head Nd:YAG laser with pulse energy of 50 mJ at the second harmonic of 532
nm. The camera is HiSense Mkll PIV CCD cameras (model C8484-5205CP) with 1280 x 1024 CCD light
sensitive array and equal number of storage cells. The objective of the camera is covered with
interference filters at 532 nm with a bandwidth of 10 nm. The laser pulse duration is 6 ns and the inter-
pulse delay between the two laser heads is controlled according to the flow velocity with a minimum of 0.2

µs for supersonic flow. The laser sheet is created by sheet forming optics that produces expanding laser
sheet. Timing between the laser ignition, the PIV laser sheet forming and the camera capturing was
controlled by 4 channel Stanford Research DG535 pulse delay generator and monitored with a Tektronix
4 channel 200 MHz oscilloscope. The PIV images were processed using an adaptive window offset
cross-correlation algorithm implemented in a commercial analysis package (Dantec Dynamic Studio V
2.30). The final interrogation window was 32×32 pixels with a 50% window overlap, resulting in a spatial
resolution 1.5 mm and vector spacing of 0.5 mm.
FIGURE 1: a) The burner, b) The experimental set up.
3. FLAME STABILITY AND SELECTED CONDITIONS
Flame extinction can be approached by either increasing the jet velocity for the same fueling rate or
reducing the equivalence ratio at the same jet air velocity. In the present work the degree of partial
premixing is fixed where the premixing length L is kept constant at L/D = 2 for all cases, where D is the
inner diameter of the outer burner tube. Figure 2 illustrates the stability characteristics of the burner,
where the stability point is achieved via gradual decrease of the fuel flow until the flame blows off, while
preserving the air flow rate constant. The results are presented as a relation between the jet velocity (Uj)
and the jet equivalence ratio (Φj). The results depicted in Fig. 2 shows that the higher the jet velocity, the
higher the corresponding jet equivalence ratio required for stable flames.
In relevance with the stability limit, the flame kernel and the turbulent flow fields are investigated at three
jet equivalence ratio of 1, 1.5 and 2. For each equivalence ratio, five different jet velocities are applied
and investigated beginning from 10 m/s to 20 m/s with increasing step of 2.5 m/s, see Fig. 2. The flow

conditions are listed in Table 1. The flames are selected on the unstable region, and the flame kernels are
ignited and propagated after each laser pulse. Since this is unstable flame region, this allows the
recording ability of the flow fields at several times from the ignition during the flame kernel propagation,
without the need to shut down the flame. The flow fields of the previous flames are also examined without
ignition (isothermal flow conditions) to be used as references of comparison with those flow fields
accompanied by flame kernel.
Uj(m/s) ΦΦΦΦj=1 ΦΦΦΦj=1.5 ΦΦΦΦj=2
10 √ √ √
12.5 √ √ √
15 √ √ √
17.5 √ √ √
20 √ √ √
Table 1: The selected flames conditions
*The flame designation is FΦj-Uj, where, the symbol Φj after F indicates the jet equivalence ratio, while the second
number represents the jet velocity. For example, F1-10, it means the jet equivalence ratio equals 1 and jet velocity
equals 10 m/s. The isothermal case is indicated by adding a superscript iso to the previous acronym, like F2-10
iso
, it
means at those flow conditions the turbulent flow field is recorded without ignition.
FIGURE 2: Stability limit and the selected flame conditions
4. RESULTS AND DISCUSSION
The contours of the normalized mean axial velocity U/Uj of isothermal jet flow and the ignited cases of
flame F2-10 are illustrated in Fig. 3a at several time intervals from ignition,[the isothermal case is located
at the left of the figure while those on the right are the time history of the mean velocity fields of the
ignited cases]. As shown for the mean flow field, there is no any difference between the isothermal case
and the ignited cases at any delay time from the ignition. For the isothermal and the ignited cases, the
flow field is featured by a central core region extending to a nearly axial distance of 2D (D is the inner
diameter of the outer nozzle), with axial velocity of nearly U/Uj=1.2. The core flow region is surrounded by
annular mixing region, which merges downstream the end of the potential core region. This result
demonstrates that the flame kernel propagation has no effect on the mean flow field.

FIGURE 3: (a) Contours of the normalized mean axial velocity (U/Uj) for isothermal, at the left hand side and the
ignition cases at several time intervals from ignition in flame F2-10, at the right of the isothermal case, (b) Contours of
the normalized rms (Urms/U) for isothermal, at the left hand side and the ignition cases at several time intervals from
ignition in flame F2-10.
On the other hand, as shown in Fig. 3b the contours of the turbulent intensity in terms of Urms/U, shows a
significant difference between the isothermal and the ignited cases. This can be seen by the higher
turbulence intensity along with the jet centerline in case of the ignited cases at any delay time from the
ignition with respect to the isothermal case. This central region of the higher turbulence intensity is
propagated in the range of the delay of 150 to 1500 µs, and shows a maximum propagation axial distance
at 1500 µs. After this time the turbulent flow field returns again to the original turbulence state at time
delay of 2500 µs, as the isothermal case (see the first and the last contours of Fig. 3b). This result
indicates that the flame kernel affects the turbulent flow field. Other feature of the turbulence intensity field
associated with the flame kernel propagation is the increase of the width of the radial distribution of the
turbulence intensity in the range of Urms/U (from 0.16 to 0.56), this turbulence intensity range is spanning
a narrow radial distance in the isothermal case with the comparison of the ignited cases at any axial
distances. This means that the flame kernel propagation leads to the redistribution the turbulent kinetic
energy.
Figure 4 illustrates the radial profiles of the turbulence intensity of flame F2-10 at delay time of 1000 µs at
three axial distances, namely at the point of the sudden decay of turbulence intensity which can be taken
as a trace of flame kernel propagation, X, while the other two locations upstream and downstream X, by 2
mm, X-2, X+2, respectively. The close inspection of the profiles shows that the turbulence intensity is
(a)
(b)

relatively high along the flame centerline at X-2 and X. More outward displacement from the jet centerline,
the turbulence intensity showed a slight decrease reaching a minimum turbulence intensity point, and this
indicates that the kernel centerline is associated with higher turbulence intensity. Beyond the radial
location of the point of minimum turbulence intensity, the turbulence intensity starts to increase and
merges with the profiles of the turbulence intensity of the axial position at X+2. Where the turbulence
intensity profile at X+2 shows a minimum value at the jet centerline, which is similar to the ordinary jet
flow. This indicates at the axial position of X+2, the kernel is extinguished, which indicates the reasons
behind the radial stretching of the turbulence intensity profiles in the case of the ignited flame kernel with
the comparison to the isothermal one.
FIGURE 4: Radial distribution of axial turbulence intensity at different axial locations: upstream the flame kernels limit
X-2, at the flame kernel limit X, and at downstream the kernel X+2 of flame F2-10 at delay time = 1000 µs.
Fig. 5 shows the centerline turbulence intensity for the jet equivalence ratio of one and jet velocity of 10
m/s at different time of delay from the ignition, with the comparison with the isothermal case. As seen as
the flame kernel propagates as it delays the steep decay point of the turbulence intensity, and by this way
the flame kernel could be traced. At jet velocity of 10 m/s, increasing the jet equivalence ratio from 1 to
1.5 and 2 shows similar qualitative turbulence intensity change, as compared to the corresponding
isothermal cases, see Fig. 6. To quantify the effect of the jet equivalence ratio on the turbulence intensity
accompanied with the flame kernel propagation. This turbulence intensity along the jet centerline for the
three equivalence ratios at delay time of 1000 µs and jet velocity of 10 m/s is illustrated in Fig.7. All of
these centerline turbulence intensities are compared to the isothermal case in the same figure. In general,
centerline turbulence intensity of the isothermal case shows nearly linear decrease in turbulence intensity
along the axial distance. While, for the three equivalence ratios of the ignited cases, the centerline
turbulence intensity is noticed to be higher at the early axial distance, which also shows a mild decrease
in the turbulence intensity along with the increase in the axial distance. This mild decline in the turbulence
intensity is changed into a steep decrease in the turbulence intensity to a lower level of turbulence
intensity; but this level is still higher than those recorded in the isothermal case.
As mentioned before, the axial location of the steep decrease in the centerline turbulence intensity can be
taken as an indication of the flame kernel propagation. The high turbulence regime may be occurred due
to flame kernel fluctuation in this region and this should be associated with large rms and consequently
high turbulence intensity. In addition, the interaction between the flame kernel and flow field likely causes
disturbance and hence another factor affects the turbulence intensity. Thus, the position of the decrease

of the turbulence intensity can be used as the mean position of the flame kernel, (this will be
demonstrated later in the section of the rate of flame kernel propagation).
FIGURE 5: The centerline Urms/U for F1-10 at different time of delay and the corresponding isothermal case
As shown in Fig. 7, the jet equivalence ratio of 1 gives the highest flame kernel propagation while it
decreases by increasing the equivalence ratio. This is consistent with variation of the flame speed with
the gradient in the equivalence ratio. This reflects the effect of fuel-air chemistry on the flame kernel
propagation, where the partial premixed mixture with jet equivalence ratio of 1 sustains the flame kernel
propagation. This could be explained on the basis of the higher chemical reaction rate of the
stoichiometric mixture to withstand the turbulence effect. On the other hand, increasing the jet
equivalence ratio decreases the chemical reaction rate and hence the effect of turbulent on the flame
kernel becomes predominate, this affects the kernel propagation and its rate.
With the respect to the flame kernel propagation and the rate of propagation, Fig. 8 shows a comparison
between our results and those results of Mansour et al [23]. The results of Mansour et al. [23] are based
on using a combined two-dimensional Rayleigh and LIPF-OH technique. Where, the average axial
locations of the flame kernel center, obtained from all 200 images of LIF-OH measured in each flame.
The results show an excellent agreement between our data and the compared ones up to delay time of
300 µs, while after this delay time the two trends are qualitative similar with a slight increase in our kernel
propagation, and this is believed due to the slight increase in our selected jet velocity by 0.5 m/s, which
enhances the kernel propagation, as well as, the results of Mansour et al [23], were performed on a
premixed mixture.
The effect of both jet velocity and jet equivalence ratio on the flame kernel propagation and the rate of the
flame kernel propagation can be understood from Figs. 9 and 10, respectively. In Fig. 9, the flame kernel
propagation distance Z in mm is represented on the vertical axis versus the jet velocity on the horizontal
axis at each delay time from the ignition. The results indicate that the flame kernel propagation increases
with increasing the jet velocity at any delay time from ignition, and this is attributed to the higher
convection effect of the flow with increasing the jet velocity.
At early stage of the flame kernel generation at time lag of 150 µs a linear correlation between the jet
velocity and flame kernel propagation is recorded for the three jet equivalence ratio. Beyond this time lag
of 150 µs, a deviation from the linear correlation between the jet velocity and the kernel propagation is
noticed and the nonlinearity increases with increasing the delay time. This may be explained based on
the results of Mansour et al [23], where the shape of the flame kernel at the early delay time is not
affected by turbulence and hence this reduces the effect of turbulence level on the flame kernel
propagation at the early time of flame kernel generation. At this early time stage the flame kernel is

almost spherical. After some more time the flame kernel shape is corrugated and takes different shape,
and hence the effect of the turbulence intensity of the flow field is predominate and the flame kernel
boundaries is affected more by the turbulence eddies.
FIGURE 6: Contours of the axial turbulence intensity Urms/U of flame F1-10, flame F1.5-10 and flame F2-10
(from top to bottom, respectively)

FIGURE 7: Centerline axial turbulence intensity at Uj = 10 m/s with Φj = 1, 1.5, and 2, at delay time of 1000 µs with
comparison to the isothermal case.
FIGURE 8: The flame kernel propagation and rate of flame kernel propagation at Φj =1, and Uj=12.5 m/s, compared
to Mansour et al [23], at Φj =1, and Uj=12 m/s.
The above results can be reflected on the flame kernel propagation rate, see Fig. 10, the data provides
an illustration of the flame kernel propagation rate in m/s (vertical axis) at different delay times presented
on the horizontal axis. The results show that the rate of flame propagation increases with respect to the
time in steep manner till reaching a maximum rate at nearly 300 µs for all jet velocity. This maximum rate
increases with increasing the jet velocity due to the higher convection effect. After this point of maximum

rate, the flame kernel propagation rate is attenuated with increasing the time lag. This decay in the rate of
flame propagation is very clear at the higher jet velocity see jet velocities of 17.5 and 20 m/s in Fig. 10.
This indicates that increasing up to these high jet velocities of 17.5 and 20 m/s, increases the turbulence
intensity to a level which leads to the weakness of the chemical reaction, where the chemical reaction
can't withstand these high rates of reactants. At the same time the higher the jet velocity leads to more
local turbulence intensity in the location of the kernel, which in turn increases the heat dissipated to the
surrounding and the kernel extinction occurs. This could be highlighted by the relative early flame kernel
extinction at the jet velocity of 20 m/s without any relevance to the mixture equivalence ratio. If the
maximum rate of flame kernel propagation is normalized by the exit jet velocity, it will indicate the effect of
the chemical reaction for the flame kernel rate of propagation. At the jet velocity of 20 m/s and jet
equivalence ratio of 1, the ratio of the maximum rate of flame kernel propagation to the exit jet velocity is
nearly 3.5; this proves that the chemical reaction enhances the rate of flame kernel propagation more
than the coming from the convective effect of the flow field.
FIGURE 9: Flame kernel propagation at ΦJ =1, 1.5 and 2 for different delay times.

FIGURE 10: Rate of Flame kernel propagation at Φj =1, 1.5 and 2 for different jet velocities.
5. CONCLUSIONS
An experimental work is conducted on the flame kernel propagation of partial premixed natural gas
turbulent flames under a constant degree of partial premixing. The study is oriented to study the effect of
jet equivalence ratio and jet velocity on the flow field-flame kernel interaction. The mean flow field and
turbulence intensity are measured using two-dimensional Planar Imaging Velocimetry (PIV). The flow field
is captured for the isothermal field without ignition in order to be used as a reference to those flow fields
with flame kernel. The flow field with ignition is recorded after the start of ignition at different delay times.
The data show that the mean flow field isn't affected by the flame kernel if it is compared to the isothermal
one. It is found that the centerline turbulence intensity increases during the flame kernel propagation, and
then followed by a sudden decrease. The flame kernel propagation and its rate in the stoichiometric
flames are occurred at higher rates compared to the richer flames. This is consistent with the flame speed

data where the maximum flame speed occurs at the stoichiometric cases. This leads to more significant
effect of turbulence on the flame kernel propagation at richer conditions. During the early period of the
flame kernel generation, within the first 150 µs period, a linear correlation between the jet velocity and
flame kernel propagation is recorded for all equivalence ratios. Beyond this period of 150 µs, a deviation
from the linear correlation between the jet velocity and the kernel propagation is noticed and the
nonlinearity increases with increasing the delay time. This indicates that flame kernel propagation is
accelerated. Moreover, increasing the jet velocity up to nearly 17.5 m/s and more, increases the
turbulence intensity to a limit which weakens the chemical reaction, which can't withstand this higher rates
of reactants. The relative early flame kernel extinction at the jet velocity of 20 m/s is occurred without any
relevance to the value of jet equivalence ratio.
6. ACKNOWLEDGEMENTS
This work is financially supported by the joint project between Cairo University, Egypt, and North Carolina
State University, USA. The project title is "Computational and Experimental Studies of Turbulent
Premixed Flame Kernels". The project ID is 422.
7. FUTURE WORK
To go more depth about the flame kernel and flow field interaction, planar laser-induced fluorescence
(PLIF) and stereoscopic particle image velocimetry (PIV) will applied at 10 KHz repletion-rate to acquire
time-resolved of the flow field flame kernel interaction.
8. REFERENCES
[1] R.R. Maly, in: J.C. Hilliard, G.S. Springer (Eds.), "Flow and Combustion in Reciprocating Engines,"
Plenum Press, New York, 1983.
[2] C.F. Kaminski, J. Hult, M. Alden, S. Lindenmaier, A. Dreizler, U. Mass, M. Baum, "Complex
turbulence/chemistry interactions revealed by time resolved fluorescence and direct numerical
simulations," Proceedings of the Combustion Institute 28:399–405(2000).
[3] A. Dreizler, S. Lindenmaier, U. Maas, J. Hult, M. Alden, C.F. Kaminski, "Characterization of a spark
ignition system by planar laser induced fluorescence of high repetition rates and comparison with
chemical kinetic calculations," Applied Physics B70:287–294 (2000).
[4] S. Gashi, J. Hult, K.W. Jenkins, N. Chakraborty, R.S. Cant, C.F.Kaminski, "Curvature and wrinkling of
premixed flame kernels – comparisons of OH PLIF and DNS data," Proceedings of Combustion
Institute 30:809–817 (2005).
[5] C.C. Huang, S.S. Shy, C.C. Liu, A. Yan, "A transition on minimum ignition energy for lean turbulent
methane combustion in flamelet and distributed regimes," Proceedings of the Combustion Institute
31:1401–1409(2007).
[6] K.W. Jenkins, R.S. Cant, "Curvature effects on flame kernels in a turbulent environment," Proceedings
of the Combustion Institute 29:2023–2029(2002).
[7] D. Thevenin, O. Gicquel, J. de Charentenay, R. Hilbert, D. Veynante, "Two versus three dimensional
direct simulations of turbulent methane flame kernels using realistic chemistry," Proceedings of the
Combustion Institute 29:2031–2039 (2003).
[8] N. Chakraborty, M. Klein, R.S. Cant, "Stretch effects on displacement speed in turbulent premixed
flame kernels in the thin reaction zones regime," Proceedings of the Combustion Institute 31: 1385–
1392(2007).
[9] K.W. Jenkins, M. Klein, N. Chakraborty, R.S. Cant, "Effects of strain rate and curvature on the
propagation of a spherical flame kernel in the thin reaction zones regime," Combustion and Flame
145:415–434 (2006).
[10] V.R. Katta, K.Y. Iisu, and W.M. Roquemore, "Local Extinction in an unsteady methane-air jet diffusion
flame," Proceedings of the Combustion Institute 27:1121-1129 (1998).
[11] P.H. Renard, J.C. Rolon, D. Thevenin, , and S. Candel, "Investigations of heat release, extinction,
and time evolution of the flame surface, for a non-premixed flame interacting with a vortex,"
Combustion and Flame 117:189–205 (1999).

[12] S.K. Marley, S.J. Danby, W.L. Roberts, M.C. Drake, T.D. Fansler, "Quantification of transient stretch
effects on kernel–vortex interactions in premixed methane–air flames," Combustion and Flame 154
(2008) 296–309.
[13] C. Arcoumanis, D.R. Hall, and J. H. Whitelaw, "An approach to charge stratification in lean-burn
spark-ignition engines," SAE technical paper 941878 (1994).
[14] C. Arcoumanis, D.R. Hall, and J.H. Whitelaw, "Optimizing local charge stratification in a lean-burn
spark ignition engine," Proc. Instn. Mech. Engrs, Part D: J. Auto. Eng. 211:145–154(1997).
[15] C. Arcoumanis, M.R. Gold, J.H. Whitelaw, and H.M. Xu, "Local mixture injection to extend the lean
limit of spark-ignition engines," Exper. Fluids 26:126–135 (1999).
[16] D.A. Eichenberger, W.L. Roberts, "Effect of unsteady stretch on spark-ignited flame kernel survival,"
Combustion and Flame 118:469–478(1999).
[17] Y. Xiong, W.L. Roberts, M.C. Drake, T.D. Fansler, "Investigation of pre-mixed flame-kernel/vortex
interactions via high-speed imaging," Combustion and Flame 126:1827–1844 (2001).
[18] Y. Xiong, W.L. Roberts, "Observations on the interaction between a premixed flame kernel and a
vortex of different equivalence ratio," Proceedings of the Combustion Institute. 29:1687–1693(2002).
[19] D. Thevenin, P.H. Renard, J.C. Rolon, and S. Candel, "Extinction processes during a non-premixed
flame / vortex interaction," Proceedings of the Combustion Institute. 27:719–726 (1998).
[20] P.H. Renard, J.C. Rolon, D. Thevenin, and S. Candel, "Wrinkling, pocket formation and double
premixed flame interaction processes," Proceedings of the Combustion Institute. 27:659–666 (1998).
[21] G. Patnaik, and K. Kailasanath, "A computational study of local quenching in flame-vortex
interactions with radiative losses," Proceedings of the Combustion Institute. 27:711–717 (1998).
[22] H. Reddy and J. Abraham, "A Numerical Study of Vortex Interactions with Flames Developing from
Ignition Kernels in Lean Methane/Air Mixtures," Combustion and Flame 158 (2011) 401–415.
[23] M.S. Mansour, , N. Peters, L.U. Schrader, "Experimental study of turbulent flame kernel propagation,"
Experimental Thermal and Fluid Science. 32:1396–1404(2008).
[24] B.D. Videto, D.A. Santavicca, "A turbulent flow system for studying turbulent combustion processes,"
Combustion Science and Technology 76 (1991) 159–164.

Experimental Investigation of Flame Kernel in Turbulent Partial Premixed Flames

More Related Content

What's hot (18)

Viewers also liked (14)

Similar to Experimental Investigation of Flame Kernel in Turbulent Partial Premixed Flames (20)

More from Waqas Tariq (20)

Recently uploaded (20)

Experimental Investigation of Flame Kernel in Turbulent Partial Premixed Flames