Professional Documents
Culture Documents
a r t i c l e
i n f o
Article history:
Received 22 September 2008
Received in revised form 27 December 2008
Accepted 25 February 2009
Available online 23 May 2009
Keywords:
Flame propagation
Flame speed
Spray
Flame structure
Laminar ame speed
n-decane
n-heptane
relight
a b s t r a c t
In order to clarify the conditions conducive to propagation of premixed ames in quiescent sprays, a onedimensional code with detailed chemistry and transport was used. n-Heptane and n-decane, distinguished by their volatility, were studied under atmospheric and low temperature, low pressure conditions. The effects of initial droplet diameter, overall equivalence ratio /0 and droplet residence time
before reaching the ame front were examined. Increasing the residence time had an effect only for
n-heptane, with virtually no evaporation occurring before the ame front for n-decane. The trends were
only marginally correlated with the local gaseous equivalence ratio /eff at the location of maximum heat
release rate. /eff could be as low as 0.4 (beyond the lean ammability limit), but the ame speed could
still be 40% of the gaseous stoichiometric ame speed SL,0. For n-heptane, /eff increased towards /0 with
smaller droplets while high ame speeds occurred when /eff was near 1. This implied that the highest
ame speed was achieved with small droplets for /0 6 1 and with relatively large droplets for /0 > 1.
In the latter case, the oxidiser was completely consumed in the reaction zone and droplets nished evaporating behind the ame where the fuel was pyrolysed. The resulting small species, mainly C2H2, C2H4
and H2, diffused back to the oxidation zone and enhanced the reaction rate there. Ultimately, this could
result in ame speeds higher than SL,0 even with /0 = 4. For n-decane, the same trends were followed but
smaller droplets were needed to reach the same /eff due to the slow evaporation rate. Under low pressure
and low temperature, the effects of pressure and temperature on /eff and the ame speed were competitive and resulted in values close to the ones at atmospheric conditions.
2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved.
1. Introduction
Understanding the propagation of a premixed ame in a cloud
of droplets is relevant for many combustion applications including
spark ignition in internal combustion engines and aviation gas turbines. After the generation and growth of a ame kernel, the subsequent ame propagation will determine whether the ignition is
successful or not [1]. Although realistic systems involve polydisperse sprays and turbulent ows, fundamental understanding of
the factors controlling ame propagation can be provided by
studying monodisperse mists of quiescent droplets, which is the
topic examined in this paper.
Flame propagation in a uniformly-dispersed spray has been
widely studied in the past [29]. As discussed in the literature,
the structure of a laminar unstrained spray ame may be divided
in several zones, as shown in Fig. 1. With analytical methods, Silverman et al. [7] distinguished ve parts. The rst three parts are a
primary evaporation zone with a length l, a heating zone, and a
homogeneous reaction zone. The fuel that has evaporated in
the rst two zones reacts in the homogeneous reaction zone. Be* Corresponding author. Fax: +44 1223 3 32662.
E-mail address: em257@eng.cam.ac.uk (E. Mastorakos).
0010-2180/$ - see front matter 2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved.
doi:10.1016/j.combustame.2009.02.014
1628
and the residence time of droplets before the ame (which affects
the degree of pre-evaporation) and we examine the importance of
detailed chemistry. Finally, in view of the importance to aviation,
we look at the effect of relight conditions, i.e. low pressure and
low temperature, on the ame speed. Apparently, few investigations have focused on this regime. In order to examine the effects
of fuel volatility, n-heptane and n-decane are used as typical fuels
with high and low volatility respectively. n-Heptane has been employed in academic congurations of ignition [25] while n-decane
is used as a surrogate for kerosene [26].
The numerical model implemented is presented in the next Section, followed by the results and their discussion. The particular
features revealed by the use of detailed chemistry are emphasized.
The paper closes with a summary of the main conclusions.
2. Numerical formulation
[4,9,1618]. With an overall rich ame, the existence of an optimum droplet size was related to the incomplete evaporation of
large droplets, which produced a gaseous mixture close to stoichiometry that led to high ame propagation speed. The gaseous stoichiometry at the ame front was denoted effective equivalence
ratio /eff by Hayashi et al. [10]. It was suggested that the optimum
conditions were a combination of droplet size and overall equivalence ratio that resulted in /eff near unity [7,14,16].
The ame speed enhancement has also been predicted numerically with one-dimensional formulations and single-step chemistry for octane [19] for overall rich ames. The explanation given
was that under some conditions the mixture at the ame could become close to stoichiometric. The one-dimensional formulation
does not account for multi-dimensional droplet-scale mixture
inhomogeneities, which is the explanation suggested by experiment, and hence the fact that ame speed enhancement can still
be observed in such simulations suggests that the effective equivalence ratio concept has some merit. However, /eff has not been
discussed yet in the context of detailed chemistry calculations.
Investigations on the minimum energy necessary to ignite
sprays have been reviewed by Aggarwal [20] who pointed out that
for the same reasons, for rich overall stoichiometry, large droplets
may result in a minimum of the ignition energy. Among others, this
was shown experimentally by Danis et al. with n-heptane [21],
Singh and Polymeropoulos with tetralin [22], Dietrich et al. with
n-decane [23] and numerically using single-step chemistry for
n-decane in closed volume by Aggarwal and Sirignano [24]. The
fact that intra-droplet regions in an overall too lean or too rich
spray may be close to stoichiometric also implies a widening of
the ammability limits of ames in sprays [20].
Investigations of the effect of fuel volatility on ame propagation indicate that less volatile fuel leads to a lower ame speed
[16] and a higher ignition energy [20]. This is generally explained
by the low evaporation rate of less volatile fuels resulting in low
fuel concentration in the gas and thus too low /eff. Studies on the
effect of low pressure on spray ames are seldom. Still, Ballal
and Lefebvre [5] provided valuable results for stoichiometric mixtures of diesel oil and air with a Sauter mean drop diameter (SMD)
of 60 lm. They showed that the pressure dependence of the ame
speed varies from p0.3 with pure gas to p0.5 with sprays.
Little work has been done on the numerical study of laminar
monodisperse spray ames with detailed chemistry. Here, we attempt to elucidate the effects of droplet size, equivalence ratio
The coupling between the gas and the liquid phase implemented in Run1DL [27,28,19], the code that was used here, is
essentially the same as the one presented by Abramzon and Sirignano [29]. The code solves a one-dimensional form of the governing equations with a distance coordinate y. A steady solution with
constant pressure is presented. The governing equations are repeated below for completeness.
2.1. Gas phase
The Eulerian gas phase equations include a source term from
the liquid phase. Assuming a dilute spray allows the treatment of
the spray in a simple statistical manner: the source term for a particular scalar is equal to the scalar production from one droplet
multiplied by the number of droplets per unit volume n. The conservation of mass, species, momentum and energy and the ideal
equation of state in the gas phase are given by
@q @qv
Sm
@t
@y
@qY i @qY i v
@qY i V i
wi Sm;i
@t
@y
@y
@qv @qvv
@p @
@v
l
Sv
@t
@y
@y @y
@y
X
@qcp T @qcp T v
@
@T
k
hi wi Se
@t
@y
@y
@y
i
!
X Yi
p qRT
Wi
i
1
2
3
4
5
The gaseous properties used in those equations are the density of the
gas, q, the one-dimensional velocity v, the pressure p, the mass fraction of species i, Yi, the chemical production of species i, wi, the viscosity l, the constant pressure heat capacity of the gas mixture cp,
the temperature T, the conductivity k, the absolute enthalpy of species i, hi, the ideal gas constant R and the molecular mass of species
i, Wi. The diffusion velocities Vi are calculated by V i DX ii rX i V c
where Di is the mixture averaged diffusion coefcient of species i
and Vc is a correction velocity used to enforce the constraint
P
p 3
i qY i V i 0. Moreover, 1 n 6 d is the void volume fraction dened as the volume occupied by the gas only divided by the total volume. For our simulations 0 6 1 6 104 which supports the
assumption of dilute spray. The liquid source terms are given by
_
Sm nm
dv d
_ vd
Sv n md
m
dt
_ pd T T s Lv
Se nq_ mc
6
7
8
1629
9
10
f ln1 BM
_ pdqf Df Sh
m
h
i
f 2 1
Sh
1 Red Sc1=3 max1; Red 0:077 1
FBM
qf djv v d j
Red
lf
BM
Y Fs Y F1
1 Y Fs
11
12
13
14
where qf is the thin lm density and Df is the binary diffusion coefcient of fuel vapour inside the thin lm. Red is the droplet Reynolds
number, always below 100 in our problem, and lf is the lm viscosity. The mass fraction of fuel at the droplet surface Y F s is found
through the ClausiusClapeyron equation by assuming phase equilibrium at the droplet surface
W F XFs
Y Fs
W F X F s 1 X F s W
pref
Lv 1
1
XFs
exp
p
R T ref T s
15
16
where pref and Tref are state reference pressure and temperature, WF
is the molecular mass of the fuel and W is the molecular mass of the
mixture excluding the fuel. Lv is assumed to be constant and equal
to its value at the boiling point. For n-heptane, pref = 1 bar,
Tref = 371.58 K and Lv = 317.927 kJ/kg. For n-decane, pref = 1 bar,
Tref = 447.30 K and Lv = 279.624 kJ/kg.
Inter-droplet interactions and gravity being neglected, the drag
force only affects the droplet movement and the momentum equation reduces to
2
md
d xd
dt
3C D q
jv v d jv v d
d qd
@T d
1 @
Td
r2
ad 2
r @r
@t
r
@T d
r 0 0
@r
q_
@T d
d
2
r
2
@r
pd ad qd cp
21
22
d
"
_
q_ m
0:38 #
cpd T T s
T crit T S
Lv
BT
T crit T ref
23
BT 1 BM / 1
24
f a
cp Sh
f
/ d:
:
c pf g
Nu Df
25
g
Nu 2
26
i
1 h
1 Red Pr1=3 max1; Red 0:077 1
FBT
where cpf and af are respectively the lm heat capacity and thermal
diffusivity of the lm mixture. g
Nu is the modied Nusselt number.
The critical temperature Tcrit is 540.15 K for n-heptane and 619.0 K
for n-decane. On the left boundary, droplets have a uniform temperature which is the cold gas temperature and their velocity is the
cold gas velocity. The ame speed is given as the speed of the uid
at the cold boundary. An initial ame speed is provided and iterations are performed until a steady state is established. The procedure for the coupling of the two phases is as follows. First, the
liquid-phase equations are solved using the initial eld of the gas
phase and liquid source terms are calculated. This is done with a
high-order accurate ODE solver whose maximum time-step size is
adapted to be sufciently small for at least two liquid-phase solution points to lie in each gas-phase control volume. This ensures
that in each gas-phase control volume linear interpolation of the liquid-phase source terms on the Eulerian nodes is meaningful and
sufciently accurate. A local source term is attributed to each Eulerian node, without smoothing being added. The gas-phase equations are then solved. The liquid phase is then solved anew based
on the new gas eld. The iterations are repeated until convergence
is achieved. Adaptive gridding based on all variables was used and
grid independence was ensured with about 160 points for all
solutions.
The overall equivalence ratio /0 was determined by the incoming mass uxes of gaseous air, gaseous fuel and liquid fuel. At the
cold boundary, the mass ux of gaseous fuel was set to zero so that
the liquid fuel ux only determined the overall equivalence ratio.
Table 1 summarises the conditions investigated.
Table 1
Range of conditions studied. Case C is referred to as relight.
20
17
2=3
Re
24
where C D Re
1 6d is the drag coefcient [29]. In one-dimen-
nud n0 u0
19
T d r; 0 T d0
dmd
_
m
dt
1 3
md qd pd
6
The internal droplet temperature is calculated by a nite-conductivity model [30]. The temperature within the droplet, assumed
to be a sphere, is given by
18
Case A
Case B
Case C
T (K)
p (bar)
/0
d (lm)
l (mm)
Fuel
298
298
265
1
1
0.4137
0.66
0.6-6
0.66
5140
5140
5140
225
225
225
n-heptane
n-decane
n-decane
2200
1800
1400
d=10 m, o=1
1000
d=50 m, =4
d=50 m, =1
o
o
600
200
5
4
5
3
x 10
y(m)
0.1
0
0.1
Q/(Qgas,max)
gas, =1
o
T(K)
1630
0.2
gas, =1
0.3
d=10 m, o=1
0.4
d=18 m, =1
o
0.5
d=50 m, o=1
0.6
d=50 m, o=4
0.7
0.8
0.9
1
0.5
1.5
2
3
x 10
y(m)
Fig. 3. Heat release through n-heptane ames with l = 5 mm and the indicated
equivalence ratio and droplet size.
0.25
gas, o=1
d=10 m, o=1
0.2
d=50 m, o=1
d=50 m, o=4
0.15
3. Results
In this Section, we present the dependence of the ame speed
on various factors for the case denoted A in Table 1. The other cases
are qualitatively similar and are mentioned only when signicant
differences exist. Results specic to cases B and C, concerning the
inuence of the fuel and the ambient conditions, are presented in
Sections 3.3 and 3.4.
0.1
C7H16
0.05
0
5
y(m)
4
5
3
x 10
Fig. 4. Reactants mass fractions through n-heptane ames with l = 5 mm and the
indicated equivalence ratio and droplet size.
Before we introduce results regarding the ame speed, we compare, at an overall equivalence ratio /0 = 1, the structure of a gaseous n-heptane ame with two types of n-heptane spray ames:
one, with an initial diameter d = 10 lm, is similar to a gaseous
ame while the other, with d = 50 lm, is more characteristic of
spray combustion. In Figs. 25 we present the variations of the
temperature, the heat release and mass fractions of the main species through the simulated ame. The heat release has been normalised by the maximum value of the heat release of the
stoichiometric gaseous ame. We can see in Fig. 2 that the temperature prole is very close to the gaseous case with d = 10 lm while
1631
0.2
0.15
CO
0.1
gas, o=1
d=10 m, o=1
0.05
d=50 m, =1
o
d=50 m, o=4
0
1
5
3
y(m)
x 10
Fig. 5. CO2 mass fraction through n-heptane ames with l = 5 mm and the indicated
equivalence ratio and droplet size.
Fig. 6 presents the variation of the ame speed with the initial
droplet diameter for three different /0 and for different lengths l
of the primary evaporation zone. For /0 = 1, the general trend is
that the ame speed is reduced as the droplet diameter increases.
As the droplet size falls below 30 lm, the ame speed comes close
to the gaseous ame speed at the same overall equivalence ratio.
This agrees well with previous ndings [2,10,16]. It is also evident
that the ame speed is not a monotonic function of the droplet size
and that a maximum exists at a nite droplet diameter. The corresponding optimum diameter increases from 14 lm to 25 lm when
the residence length l varies from 2 mm to 25 mm. This range of
diameters has been interpreted as a transition range below which
the ame behaves like a gaseous ame [11]. Fig. 6 also compares
our data with a theoretical formula based on considerations of
evaporation rates and chemical reaction rates in the framework
of steady ames and /0 6 1. The expression given by [5] is:
"
C 33 Xqd D232
a2
SL a
8C 1 q ln1 BM S2L;g
o=1
0.5
0.4
gaseous flame, =1
0
0.35
SL(m/s)
S (m/s)
0.45
2mm
5mm
15mm
25mm
Theory
0.15
27
D32 is the Sauter Mean Diameter at the start of the preheat zone and
BM the transfer number. a and q are respectively the thermal diffusivity and the density of the gas at 1200 K. SL,g is the gaseous ame
speed at the same overall stoichiometry. For a monodisperse spray
containing no fuel vapour, the liquid fraction X is equal to 1 and
o=0.6
0.2
#0:5
0.1
2mm
5mm
15mm
25mm
Theory
0.3
0.25
0.2
0.15
0.05
0.1
0.05
0
0
20
40
60
80
100
120
140
20
40
d( m)
60
80
100
120
140
d( m)
=4
o
0.5
0.45
0.4
S (m/s)
0.35
0.3
0.25
0.2
2mm
5mm
15mm
25mm
0.15
0.1
0.05
0
0
20
40
60
80
100
120
140
d( m)
Fig. 6. n-Heptane ame speed as a function of the initial droplet diameter for different lengths l. For /0 6 1, the data are compared with the theoretical curve prescribed by
Ballal and Lefebvre [5]. (a) /0 = 0.6, (b) /0 = 1 and (c) /0 = 4.
1632
Flame speeds calculated with n-decane at atmospheric conditions (Flames B) generally show the same trends as presented
above. However, some differences, which may be attributed to
the lower volatility of n-decane, are worth pointing out. In Fig. 9a
we remark again the inverse dependence of the ame speed on
the initial diameter. However, the diameter below which the ame
speed is very close to the gaseous ame speed is around 15 lm,
which is much smaller than for n-heptane. In fact, the reduction
of the ame speed with larger diameters is stronger in the case
of n-decane. This agrees with results reviewed previously. Furthermore, the value of the laminar ame speed is virtually independent
o=1
L,0
0.4
0.35
0.3
0.25
0.2
1
0.8
SL/SL,0
pure gas
5m
18 m
50 m
100 m
140 m
0.45
0.5
S (m/s)
30 m
0.6
50 m
0.4
100 m
0.15
0.2
0.1
140 m
0.05
0
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
0.25
0.5
0.75
Fig. 7. n-Heptane ame speed as a function of the overall equivalence ratio for
different initial droplet diameters and l = 5 mm.
1633
0.5
0.5
0.45
0.45
0.4
o=4
0.2
0.3
o=1
0.25
0.15
2mm
5mm
15mm
25mm
o=4
=1
gaseous flame, =1
0.35
S (m/s)
2mm
5mm
15mm
25mm
0.3
0.25
S (m/s)
0.35
0.4
0.2
0.15
0.1
0.1
=0.6
0.05
o=0.6
0.05
0
0
20
40
60
80
100
120
140
20
d( m)
40
60
80
100
120
140
d( m)
Fig. 9. n-Decane ame speed at various /0 as a function of the initial droplet diameter for different lengths l. The gaseous ame does not exist at /0 = 4. (a) Atmospheric
conditions and (b) relight conditions.
0.5
0.5
0.45
0.45
0.4
0.4
0.35
L,0
0.25
0.2
0.15
0.1
0.05
pure gas
5m
10 m
30 m
50 m
100 m
0.3
S (m/s)
SL(m/s)
0.35
L,0
pure gas
5m
15 m
30 m
50 m
100 m
0.3
0.25
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Fig. 10. n-Decane ame speed as a function of the overall equivalence ratio for different initial droplet diameters and l = 5 mm. (a) Atmospheric conditions and (b) relight
conditions.
S /S
L,0
1634
Numerical
Theory
Experiment
1.1
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0
20
40
60
80
100
120
140
d( m)
Fig. 11. Comparison between the theoretical curve, experiments with heavy fuel oil
(SL,0 = 0.41 m/s) [5] and our results with n-decane at an /0 = 1 and l = 2 mm.
0.5
P=100kPa,T=298K
P=41.37kPa,T=265K
P=30kPa,T=230K
0.45
0.4
S (m/s)
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
0
0.5
1.5
Fig. 12. Flame speed of gaseous n-decane at three different conditions of pressure
and initial temperature.
1635
o=0.6
o=1
1.2
1
1.2
2mm
5mm
15mm
25mm
2mm
5mm
15mm
25mm
0.8
eff
eff
0.8
0.6
0.6
0.4
0.4
0.2
0.2
0
0
20
40
60
80
100
d( m)
120
140
20
40
60
80
d( m)
100
120
140
o=4
1.2
2mm
5mm
15mm
25mm
eff
0.8
0.6
0.4
0.2
0
0
20
40
60
80
100
120
140
d( m)
Fig. 13. Effective equivalence ratio for n-heptane as a function of the initial droplet diameter for different lengths l. (a) /0 = 0.6, (b) /0 = 1 and (c) /0 = 4.
1.2
0=1
=4
0.8
1.2
2mm
5mm
15mm
25mm
=4
0
0.8
2mm
5mm
15mm
25mm
0=1
eff
eff
0.6
0.6
=0.6
0
0.4
0.4
=0.6
0
0.2
0.2
0
0
20
40
60
80
100
120
140
d( m)
20
40
60
80
100
120
140
d( m)
Fig. 14. Effective equivalence ratio for n-decane as a function of the initial droplet diameters for different lengths l. (a) Atmospheric conditions and (b) relight conditions.
1636
1.4
1.4
1.2
1.2
SL/SL,0
SL/SL,0
1
0.8
0.6
1
0.8
0.6
0.4
0.4
0.2
0.2
0
0
0
0.2 0.4 0.6 0.8
eff
eff
1.4
1.2
SL/SL,0
1
0.8
0.6
0.4
0.2
0
0
eff
Fig. 15. Normalised ame speeds as a function of /eff. (a) n-Heptane, (b) n-decane atmospheric and (c) n-decane, relight. Line: gas. Symbols: spray.
ame speed is nite at very low /eff, even beyond the lean ammability limit. Although the governing equations used here do not
fully account for ammability limits as there is no heat loss mechanism, the code converges for gaseous ames only in a range of
equivalence ratios and gives very low speeds (SL/SL,0 < 0.2) at
/0 0.5 and /0 2. The spray ames, in contrast, show high ame
speeds for /eff as low as 0.4. This widening of the ammability limits in sprays has been attributed to inter-droplet inhomogeneities,
where locally the mixture may reach ammable values [2,17,9]
and hence sustain ame droplet-to-droplet propagation. In the
next section, we provide an alternative explanation for this phe-
x 10
YOH
3
gas, =1
o
d=10 m, =1
d=50 m, =1
o
d=50 m, =4
0
1
y(m)
5
3
x 10
Fig. 16. OH mass fraction through n-heptane ames with l = 5 mm and the
indicated equivalence ratio and droplet size.
1637
x 10
gas, o=1
d=50 m, =1
d=10 m, =1
YC
0.03
gas, =1
2 2
0.02
d=10 m, =1
d=50 m, =4
d=50 m, o=1
d=50 m, =4
0.01
1
0
1
0
1
y(m)
5
3
y(m)
x 10
x 10
0.05
0.04
gas, =1
C H
0.03
d=10 m, =1
C H
d=10 m, =1
2 4
gas, =1
x 10
3 6
d=50 m, =1
o
d=50 m, =1
d=50 m, =4
d=50 m, =4
o
0.02
0.01
0
1
0
1
x 10
y(m)
y(m)
5
3
x 10
x 10
gas, =1
o
d=10 m, o=1
d=50 m, =1
Y1C
d=50 m, o=4
4 8
0
1
y(m)
5
3
x 10
Fig. 17. Mass fraction of intermediate species through n-heptane ames with l = 5 mm and the indicated equivalence ratio and droplet size. (a) H2, (b) C2H2, (c) C2H4, (d) C3H6
and (e) 1 C4H8.
1638
400
700
gas, o=1
d=18 m, =1
o
w(mol/m3/s)
300
d=50 m, o=4
200
w(mol/m3/s)
d=50 m, o=1
gas, o=1
600
d=18 m, =1
500
d=50 m, o=1
d=50 m, o=4
400
300
200
100
100
0
0
2
12
17
12
17
x 10
y(m)
1600
x 10
y(m)
gas, o=1
d=18 m, =1
2800
gas, o=1
2400
d=50 m, =1
2000
d=50 m, =4
d=18 m, =1
o
o
d=50 m, =4
o
o
w(mol/m /s)
w(mol/m3/s)
d=50 m, =1
1200
800
1600
1200
800
400
400
0
0
2
12
17
12
17
4
x 10
y(m)
x 10
y(m)
Fig. 18. Rates of elementary reactions through n-heptane ames with l = 5 mm and the indicated equivalence ratio and droplet size. (a) C7H16 + O , C7H15 + OH,
(b) C7H16 + OH , C7H15 + H2O, (c) CH2O + OH , HCO + H2O and (d) HCO + O2 , CO + HO2.
200
200
gas, o=1
d=18 m, =1
o
d=50 m, o=1
100
50
150
w(mol/m3/s)
d=50 m, o=4
w(mol/m /s)
150
gas, =1
100
d=18 m, o=1
d=50 m, o=1
50
d=50 m, o=4
0
2
y(m)
12
17
4
x 10
y(m)
12
17
4
x 10
Fig. 19. Pyrolysis reaction rate through n-heptane ames with l = 5 mm and the indicated equivalence ratio and droplet size. (a) C7H16 , C5H11 + C2H5 and (b)
C7H16 , pC4H9 + nC3H7.
vive far beyond the ame. The temperature and heat release
proles presented in Figs. 2 and 3 are consistent with the endothermic nature of pyrolysis: the temperature prole presents a
negative slope behind the ame front, coincident with the change
of the heat release sign.
In Fig. 20 we present the net formation rate of some products of
the decomposition of fuel: H2 and C2H2, both of which are highly
reactive. In the case d = 50 lm, /0 = 4, the production rates of these
species are very high in the pyrolysis zone because large amounts
1639
30
20
20
w(mol/m /s)
w(mol/m /s)
10
10
0
gas, =1
o
10
d=18 m, =1
o
0
gas, =1
o
d=18 m, o=1
10
d=50 m, =1
d=50 m, =1
20
d=50 m, =4
d=50 m, =4
30
12
y(m)
17
20
12
17
4
y(m)
x 10
x 10
Fig. 20. Formation rate of some products of fuel decomposition through n-heptane ames with l = 5 mm and the indicated equivalence ratio and droplet size. (a) C2H2 and
(b) H2.
10000
H+O O+OH
2
3
w(mol/m /s)
8000
6000
H +OHH O+H
2
4000
HCO+MH+CO+M
2000
0
0.2
0.7
1.2
y(m)
Fig. 21. Normalised sensitivity coefcient of the velocity on reaction rates. The
particular reactions selected have the highest degree of sensitivity.
1.7
3
x 10
Fig. 22. Rates of the indicated elementary reaction, selected as reactions with high
degree of sensitivity through n-heptane ames with l = 5 mm and the indicated
equivalence ratio and droplet size. Lines: gas, /0 = 1. Circles: d = 50 lm, /0 = 4.
1640
d = 18 lm, /0 = 1, this implies that a large part of oxidiser is consumed in the homogeneous reaction zone while droplets still exist
behind it. Furthermore, Fig. 19 shows that behind the homogeneous reaction zone, the fuel evaporated by droplets is intensively
pyrolysed eventually resulting in non-negligible generation of H2,
see Fig. 20. Although this is not a proper pyrolysis zone because
combustion reactions still occur, see Fig. 18, we can infer that a
pool of intermediates such as H2 is created. Those reactive species
can then diffuse back and enhance the reaction rates, as in the
d = 50 lm, /0 = 4 ame. To corroborate this statement, we calculated the integrated heat release for the case d = 18 lm, /0 = 1
and found a value of 5.32 107 W/m3 which was higher than both
cases. The
the d = 10 lm, /0 = 1 and the gaseous stoichiometric
p
R
a
overall result
was
that
the
ratio
between
Qdy
for
d = 18 lm,
p
R
/0 = 1 and a Qdy for d = 10 lm, /0 = 1 was 1.07, while the ratio
of the ame speeds was 1.06. This close agreement supports the
fact that incomplete evaporation at the reaction zone may be benecial to the ame propagation.
5. Conclusions
Calculations of one-dimensional laminar ames in sprays of
n-heptane and n-decane under atmospheric and relight conditions
have been presented. The effects of droplet diameter, overall
equivalence ratio and residence time before the ame zone were
examined. The trends were only marginally correlated with the
effective equivalence ratio /eff, based on the total gaseous carbon
mass fraction at the location of maximum heat release rate. /eff
was close to the total equivalence ratio /0 for small droplets and
long residence times. High ame speeds were more likely to occur
when /eff was close to unity. This implies that the maximum ame
speed is achieved with small diameters and long residence time
under lean conditions. Under rich conditions, the ame speed
was greater for relatively large droplet sizes that caused /eff to
be near unity. This was in agreement with previous studies.
Three different structures of a spray ame were observed. In the
rst one, droplets evaporated completely before the ame front
and the structure was identical to a gaseous ame. In the second,
droplets evaporated very little so that /eff was low. Little oxidiser
reacted in the homogeneous reaction zone and the rest was consumed in a long droplet burning zone providing heat release. The
third ame structure occurred for rich sprays where /eff was close
to unity and the droplets large enough. In this case, the oxidiser
was consumed in a relatively short reaction zone. The fuel evaporated behind it by the surviving droplets was pyrolysed and the
reactive species generated (mainly H2, C2H2, C2H4) diffused back
towards the oxidation zone and enhanced the reaction rates. This
could result in enhanced ame speed, above the maximum gaseous
ame speed even for /0 = 4.
The same trends were followed with the less volatile fuel,
n-decane. However, the low evaporation rate generally induced a
lower /eff than with n-heptane at atmospheric conditions. At high
altitude relight conditions, the residence time had no effect on /eff
and the ame speed. Despite the low temperature, the value of /eff
was not markedly affected by the relight conditions thanks to the
low pressure that enhanced the evaporation rate in the preheat
zone. The ame speed was comparable to that at atmospheric con-
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]