Soot Diagnostics for Non-Premixed and Partial Premixed Flames

Joint Meeting of The Scandinavian-Nordic and Italian Sections of The Combustion Institute
Soot Diagnostics for Non-Premixed and Partial Premixed
Flames
Mariusz Choinski, Claudya Arana, Swarnendu Sen, Ishwar K.Puri
Department of Mechanical and Industrial Engineering
University of Illinois at Chicago
Chicago, IL
Measurements of soot concentrations in flames are conducted by applying an optical light extinction
technique. Coannular ethylene/air nonpremixed flames are established to investigate soot formation.
Light extinction (LE) is a non-intrusive method that allows measurements of the soot field without
changing the flame properties or intruding into the flame medium. We use a coherent light source,
namely, a He-Ne laser, to measure the soot volume fraction field. As incident light travels through
the soot-containing region within the flame, light is absorbed and scattered, which results in the
attenuation of the transmitted light intensity. Abel inversion is subsequently used to obtain the soot
volume fraction field distribution once its measured projection distribution is known in an
axisymmetric field. As expected, smaller amounts of soot are found in partially premixed flames as
compared to nonpremixed flames. In partially premixed flames, the soot producing region is in the
annular outer ring along the lower region of the flame. However, soot is observed in the central
regions of the flames farther downstream. Results are validated with published results for a
nonpremixed flame.
Introduction
The exact formation of soot and the response
of soot production to various flame conditions
are not well characterized. Therefore, there is a
need to understand the mechanism of soot
formation and oxidation and to obtain accurate
measurements of soot concentrations under a
variety of conditions. The measurements should
be useful for the development and validation of
evolving numerical models for soot formation,
growth, oxidation and transport [1].
The light extinction (LE) method is nonintrusive
method that allows measurements of
the soot field without changing the flame
properties or intruding into the flame medium.
As the incident light travels through the sootcontaining
region within a flame, light is
absorbed, scattered or transmitted, which results
in the attenuation of the light intensity. The ratio
of the transmitted to incident light intensity can
be used to obtain the soot volume fraction by
employing Bouguer’s law, which relates the
ratio of the transmitted and incident light
5.1.1
intensities to the soot concentration [2]. An
Abel inversion can be used to obtain the field
distribution of soot volume fraction once its
measured projection value distribution is known
in an axisymmetric field. Santoro et al. first
employed the light extinction method at various
points in a soot field to measure the soot
distribution in a nonpremixed co-flow ethyleneair
flame [2]. Greenberg and Ku [3] and Quay et
al. [4] compared the results obtained with the
full field technique with those obtained by
Santoro et al. and found good agreement.
Mitrovic and Lee measured the soot volume
fractions in partially premixed ethylene flames
using laser-induced incandescence (LII) [5].
Experimental Procedure
Steady laminar partially premixed flames
burning ethylene (99.5%) in laboratory air were
established on co-annular axisymmetric burner.
The fuel flow-rate was maintained at 3.85cm 3 /s
and additional air to partially premix it was

varied between 5.65-1.88 cm 3 /s (corresponding
to equivalence ratios between 10 and 27). The
outer airflow was maintained at 999cm 3 /s. A
higher outer air velocity reduces the pulsing
global instability in the flames, but reduces the
visible flame height [8]. The light source is a 30
mW He-Ne laser operating at a wavelength of
632.8 nm. The laser beam passes through a
neutral density filter, a beam expander and a set
of diffusers. The beam through the diffusers is
collected by a condenser lens and directed
through a collimator before it encounters the
soot-laden flame region.. The light beam then
passes through a relay lens, and band-pass and
neutral density filters. A second neutral density
filter is placed after the band-pass filter to
further diminish any influence of the flame
intensity.. Finally, the light is decollimated on to
the CCD camera. The image is digitized and
processed by a computer.
CCD Camera
Joint Meeting of The Scandinavian-Nordic and Italian Sections of The Combustion Institute
Neutral Density
Filter(s)
Bandpass filter
(632.8nm)
Figure 1 Experimental set up
Raley Lens
Colliminator
Lens
Burner
Diffuser
Shaker
Beam
Expander
HeNe Laser
30mW
Theoretical Background
The attenuated light intensity Iat is compared
to the intensity of the original light beam Io and
the line of sight fractional absorption is
calculated after making two corrections. The
first involves the flame intensity by acquiring an
image of the flame with the light source turned
off If . The second is due to the background and
is corrected for by considering an image
obtained in the absence of both the flame and
light source Ias. The light transmittance
T = I/Io, where (1)
I=Iat-If, and (2)
I0=Ib-Ias. (3)
The transmittance depends on the soot volume
fraction fv in the path of a ray and the light
wavelength λ. This relationship is described
through Bouguer’s law, i.e.,
K R
ln T = −
e
∫ fv
ds ,
λ −R
where, Ke denotes the dimensionless extinction
coefficient, and R the radius of the flame at the
axial location above the burner. Thus, the
integrated soot volume fraction
R
λ
∫ f vds
= − lnT
. (5)
K
−R
e
The dimensionless extinction coefficient
Ke=Ka(1+αa) (6)
can be calculated by applying Mie’s theory
provided the soot particles have a small optical
dimension so that
36πn
λ
k
λ
Ka
= . (7)
2 2 2 2 2
( n
λ
− k
λ
+ 2)
+ 4n
λ
k
λ
In Eq. (7), nλ and kλ denote the real and
imaginary parts of the complex refractive index
of the particle’s material. Choi et al. have
provided experimentally measured values of the
dimensionless extinction coefficient [6]. They
report that a value of Ke=8.6±1.5 is appropriate
for hydrocarbon fuels at a light wavelength of
632nm. We have chosen the lower limit of this
value (7.1) in our analysis by comparing our
transmittance and soot volume fraction results
with those of Santoro and coworkers [4].
Abel Inversion
An inversion method must be applied to
reconstruct the non-uniform spatial soot
distribution from the path integrated data. The
projection distribution P(x) represents the
projection value distribution that is recorded by
the line of sight optical measurement. The
hypothetical distribution F(r) is the distribution
of the region of interest. Both distributions, P
and F contain the same information regarding
the physical data in the region. The objective is
to obtain the field distribution F given only the
5.1.2

Joint Meeting of The Scandinavian-Nordic and Italian Sections of The Combustion Institute
projection distribution P and the assumption that
F is axisymmetric.
∫ ∞ ∫ = =
−
∞ rF(
r)
dr
P(
x)
= 2 y=
0 F(
r)
dy 2 r x . (8)
2 2
x y
Recognizing that r represents a dummy
integration variable, (that is r=x is the radius of
the line of sight measurement), by change of
variables the Abel transform of the field F(r) can
be obtained as
∫ ∞ =
∫ ∞ mF(
m)
dm
mF ( m)
dm
P x)
= m x
= P(
r)
= m=
r
2 2
2 2
m − x
m − r
( . (9)
The analytical inversion is
dp dm
F r ∫m r dm
m r
∞ 1
( ) = − =
. (10)
π 2 2
−
Equation 10 allows one to calculate the
distribution F(r) if P(x) is known.
The Abel inversion algorithm is very
sensitive to small local changes in areas of low
signal to noise ratio. The Fourier transform
approach is an effective means to filter noise
from optical measurements in flames [2]. This
method is applied to our data prior to a
polynomial high order curve fitting function.
Results and discussion
Figure 2 presents the soot distribution in a
nonpremixed ethylene flame. The outer radial
regions of the flame contain the largest amounts
of soot. These results are in good agreement
with Greenberg and Ku [5]. There is a slight
discrepancy at the centerline, especially at a 50
mm axial displacement, where a fictitious peak
can be observed. This occurs, since the noise
cannot be filtered completely without losing
resolution. The noise creates gradients that
significantly influence the Abel inversion
algorithm. The deviation of the measurements
increases in the regions further than 50 mm
downstream of the burner. This is due to the loss
of parallel rays collimation. The nonparallel
light rays brighten up attenuated areas of the
image, which should have appeared as dark
regions.
Soot volume fractions were obtained for
partially premixed flames established at
equivalence ratios ranging from 5-24 at heights
of 20 mm to 70 mm above the burner. Figures
3-7 present the soot volume fraction as a
function of the radial distance from the burner
axis for different axial displacements. The soot
volume fraction increases when small amounts
of air are added to the mixture. This is true for
equivalence ratios greater than twenty-four at
which soot volume fraction reaches its
maximum value. The soot-producing region is
the annular outer ring in the lower region of the
flame, but the presence of soot is observed in
the central regions of flame farther downstream.
For equivalence ratios below twenty-four and
larger than ten, soot volume fraction slightly
decreases. A drastic reduction in soot
production can be observed when the
equivalence ratio falls below ten. The same
trend is observed at different heights within the
flame region at the same equivalence ratio. Hura
and Glassman [7] suggested that the initial
increase in the soot volume fraction for
equivalence ratios greater than 20 is due to the
chemistry of intermediate carbons that are
significantly altered by the partial premixing.
Here, the increase in soot production results
from oxygen addition, which enhances the local
radical pool and alters the hydrocarbon
chemistry.
The flame height follows the same trend as
the soot volume fraction. It first increases
(Φ>24) and then decreases (Φ

Soot Volume Fraction (ppm)
10
9
8
7
6
5
4
3
2
1
0
0.0
Soot Volume Fraction (ppm)
Soot Volume Fraction (ppm)
Joint Meeting of The Scandinavian-Nordic and Italian Sections of The Combustion Institute
fraction of partially premixed ethylene flames.
The light extinction method is a relatively
inexpensive method to measure the soot volume
fraction in flames. Our results show that soot
volume fraction is higher at outer radii as
compared to the centerline. The soot volume
fraction decreases with decreasing equivalent
ratios. As more air is added to the mixture, the
concentration of soot and the height of the flame
tend to decrease.
Height Above Burner 20mm
-0.2
-0.5
-0.7
-0.9
-1.1
-1.4
-1.6
-1.8
-2.1
-2.3
-2.5
-2.7
-3.0
-3.2
-3.4
-3.6
-3.9
-4.1
-4.3
-4.6
-4.8
10
9
8
7
6
5
4
3
2
1
0
0.0
10
9
8
7
6
5
4
3
2
1
0
-0.2
-0.5
-0.7
-0.9
Radial Distance (mm)
Fig 3: Soot Volume Fraction Φ=∞,24,20,10,5
0.0
-0.2
-0.5
-0.7
Height Above Burner 30mm
-1.1
-1.4
-1.6
-1.8
-2.1
-2.3
-2.5
-2.7
Radial Distance (mm)
Fig 4: Soot Volume Fraction Φ=∞,24,20,10,5
Height Above Burner 40mm
-0.9
-1.1
-1.4
-1.6
-1.8
-2.1
Radial Distance (mm)
Fig 5: Soot Volume Fraction Φ=∞,24,20,10,5
-3.0
-2.3
-3.2
-3.4
-2.5
-3.6
-2.7
-3.9
-3.0
eqratio infinite
eqratio 24
eqratio 20
eqratio 10
eqratio 5
-3.2
eqratio infinite
eqratio 24
eqratio 20
eqratio 10
eqratio 5
eqratio 24
eqratio 20
eqratio 10
eqratio 5
eqratio infinite
5.1.4
References:
1. Xiao, X., Markov, I.M., Puri K.P., and Megaridis C.M.,
“Light Extinction Soot Measurements in Axissymmetric
Flames using a Synthetic Data Processing Approach” ,
Communicated in Applied Optics (2002)
2. Santoro, R.J., Semerjian, H.G., and Dobbins, R.A,
Combust. Flame 51:203-218 (1983).
3. Greenberg, P.S, and Ku, J.C., Applied Optics 36:5514-
5522 (1997).
4. Quay, B., Lee T.-W., NI, T., and Santoro R.J.,
Combust. Flame 97: 384-392 (1994)
5. Mitrovic, A., and Lee, T.-W., Combust. Flame 115:
437-442 (1998).
6. Choi M.Y., Mulholland G.W., Hamins A., and
Kashiwagi T, Combust. Flame 102: 161-169 (1995)
7. Hura, H.S., and Glassman, I., Twenty –Second
Symposium (International) on Combustion, The
Combustion Institute, 1988, p.371.
10
9
So
ot
Vo
lu
m
e
Fr
ac
tio
n
(p
p
8
7
6
5
4
3
2
1
Height Above Burner 60mm
0 0.0 -0.2 -0.5 -0.7 -0.9 -1.1 -1.4 -1.6 -1.8 -2.1 -2.3 -2.5 -2.7 -3.0 0.0
Radial Distance (mm)
Fig 6: Soot Volume Fraction Φ=∞,24,20,10
So
ot
Vo
lu
m
e
Fr
ac
tio
n
(p
p
10
9
8
7
6
5
4
3
2
1
0
0.
0
-
0.
-
0.
-
0.
Height Above Burner 70mm
-
0.
- - - - -
1. 1. 1. 1. 2.
Radial Distance (mm)
Fig 7: Soot Volume Fraction Φ=∞,24,20,10
-
2.
-
2.
-
2.
-
3.
eqratio infinite
eqratio 24
eqratio 20
eqratio10
eqratio infinite
eqratio 24
eqratio 20
eqratio 10