Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
Transcript of Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
1/8
Characterization of laminar premixed methanolair flames
S.Y. Liaoa,b,*, D.M. Jianga, Z.H. Huanga, K. Zenga
a State Key Laboratory for Multiphase Flow in Power Engineering, Xian Jiaotong University, Xian 710049, Peoples Republic of Chinab Department of Power Engineering, College of Chongqing Communication, Chongqing 400035, Peoples Republic of China
Received 30 September 2005; received in revised form 15 December 2005; accepted 16 December 2005
Available online 7 February 2006
Abstract
This study focuses on the effects of initial temperature and pressure on the propagation characteristics of laminar premixed flame of methanol
air mixtures. Spherically expanding laminar premixed flames, freely propagating from spark ignition sources in initially quiescent methanolairmixtures, are continuously recorded by a high-speed CCD at various equivalence ratios and temperatures. The flames are then analyzed to deduce
the flame speed. The stretch imposed on the spherical flame front is explored experimentally; as a consequence, the unstretched laminar burning
velocities of methanolair flames have been derived. The present measurements are compared with the experimental data reported previously, and
good agreements are obtained. Combined previous results, a correlation in the form ofulZuloTu=Tu0aTPu=Pu0
bp has been developed to describe
the dependences of initial temperature and pressure on the burning velocities of methanolair flames. The global activation temperatures are
determined in terms of the burning mass flux. And then, the Zeldovich numbers for methanolair flames are estimated as a function of equivalence
ratio. On the basis of the mass burning flux, an alternative correlation of laminar burning velocities has been proposed, and agreements can still be
found in the comparison between this alternative forms and the power law correlation above.
q 2006 Elsevier Ltd. All rights reserved.
Keywords: Methanol; Laminar burning velocity; Premixed flame; Mass burning flux
1. Introduction
Methanol has been demonstrated to be one of the
promising alternative fuels for spark ignition engines. The
most important characteristic of methanol is that it is
undoubtedly the cheapest liquid alternative fuel per calorific
unit, which can be produced from the widely available
fossil raw materials including coal, natural gas and bio-
substances, that it can be produced from biological sources
and therefore represents a renewable energy source [1,2].
This essentially means that many countries can solve their
energy imbalance problems due to petroleum shortage by
using methanol as a source of energy.
It is known that, laminar burning velocities are fundamen-
tally important in regard to developing and justifying the
chemical kinetics mechanism of fuel, as well as in regard to
predicting the performance and emissions of the internal and
external combustion system [3]. Generally, chemical kinetics
of flames can be studied by performing numerical calculation
by solving balance equations of one-dimensional laminar
flames, but the chemical kinetic data in such models are not
always sufficiently well known to be used in confidence. In
practice, the measured laminar burning velocities are
commonly used to validate these chemical kinetic schemes.
There are many techniques for experimentally measuring the
laminar burning velocity of combustible mixture, such as
counterflow double flames [4], flat flame burner [5], heat flux
method [6], and closed bomb technique [714]. For spherically
expanding flames in closed combustion bomb, the stretch
imposed on the premixed flame front is well defined.
Furthermore, the asymptotic theories and experimental
measurements have suggested a linear relationship between
flame speeds and flame stretches [15]. As we know, the term of
laminar burning velocity is generally defined for one-
dimensional planar flames in theory, where flames are
unstretched. Therefore, measured flames can then be used
systematically to determine the fundamental burning velo-
citythe unstretched laminar burning velocity by mean of an
extrapolation at zero stretch [3]. Meanwhile, Markstein lengths
can, in principle, also be deduced. Here, Markstein length
characterize the variation in the local flame speed due to the
influence of external stretching, which is important in
Fuel 85 (2006) 13461353
www.fuelfirst.com
0016-2361/$ - see front matter q 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2005.12.015
* Corresponding author.
E-mail address: [email protected] (S.Y. Liao).
http://www.fuelfirst.com/http://www.fuelfirst.com/ -
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
2/8
expressing the onset of flame instabilities and the stretch
influence on flame quenching as well.
The measurements of burning velocities have been
extensively studied for a wide variety of hydrocarbon fuels,
including methane, ethane, propane, butane, and octane etc.
However, as to methanol, it has not been studied so far in the
literatures. Metghalchi and Keck, Gulder, and Saeed and Stonemeasured the burning velocities of methanolair mixture in
closed combustion bombs. However, in those previous studies,
a continuous observation of flame growth were not always
implemented and the laminar burning velocities were
generally, determined by resolving various combustion models
based on combustion pressure traces, but those procedures
commonly ignored the influence of stretch imposed on the
flame, hence the Markstein lengths have not been obtained as
well. Meanwhile, discrepancies still exist, to some extent,
within these experimental results. In addition, Gibbs and
Calcote [16], Muller et al. [17], and Westbrook and Dryer [18]
conducted flame chemical kinetics computations to determinethe laminar flame speed, however, the results obtained still
appear to show somewhat obvious scatterings.
In view of considerations above, the motivation of this work
is mainly due to that the burning velocity data of methanol is
still scare and less accurate. This work presents experimental
measurements of the laminar burning velocities of methanol
air premixed flames. Spherically expanding laminar premixed
flames, freely propagating from spark ignition sources in
initially quiescent methanolair mixtures, are continuously
recorded by a high-speed CCD at various equivalence ratios
and temperature conditions. The flames are analyzed to
estimate flame size, consequently, the flame speeds are derived
from the variations of the flame size against time elapsed.
Following the linear relation between flame speed and flame
stretch, the unstretched laminar burning velocities and
corresponding Markstein lengths of flames have been deduced.
The effects of the fuel/air equivalence ratio, initial temperature
and pressure on the laminar flame propagation have been
studied in detail. As a result, empirical formulas for the laminar
burning velocities are suggested, and comparison with the
previous results is made as well.
2. Experiments and procedures
Shown in Fig. 1 are the schematic diagrams of the
combustion bomb and the optical system used for recording
the flame growth. The combustion bomb has an inside size of
108!108!135 mm, as shown in Fig. 1A. Two sides of this
bomb are transparent to make the inside observable; these sides
are to provide the optical access, and the other four sides are
enclosed with resistance coils to heat the bomb to the desired
preheat temperature. The inlet/outlet valve is used to let fresh
air or combustion product in or out. The liquid fuel needed is
pre-calculated corresponding to the given equivalence ratioand then injected into the combustion chamber using a small
capability injector. The normal function of the motion of
perforated plate is to provide a turbulent combustion
environment if needed, where it is only used to enhance the
vaporization of methanol and make the reactants mixed well.
Two extended stainless steel electrodes are used to form the
spark gap at the center of this bomb. A conventional battery-
coil ignition system is used for producing the spark. The history
of the shape and size of the developing flame kernel is recorded
by a REDLAKE HG-100K high-speed CCD camera, operating
at 5000 pictures per second with a schlieren optical system, as
given in Fig. 1B. Dynamic pressure used to determine burned
gas and unburned gas properties herein, is measured since
spark ignition with a piezoelectric absolute pressure transdu-
cer, model Kistler 4075A, with a calibrating element Kistler
4618A.
Fig. 1. Schematic diagram of experimental system.
S.Y. Liao et al. / Fuel 85 (2006) 13461353 1347
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
3/8
The laminar burning velocity can be deduced from the well-
established expanding flames method as described in [7]. In
this approach, the stretched flame velocity, Sn, is derived from
the flame radius versus time data t as,
SnZdrudt
(1)
where flame size ru is defined as the schlieren flame size. The
definition of the flame stretch, a, of a flame front in a quiescent
mixture is given by,
aZ1
A
dA
dt(2)
where A is the flame front area. For a spherically outward
expanding flame, the flame stretch is well defined as,
aZ1
A
dA
dtZ
2
ru
drudtZ
2
ruSn (3)
Following an early idea of Markstein, it is suggested a linear
relationship between flame speeds and flame stretches, given as
Eq. (4),
SlKS
nZL
ba (4)
Thereby, the unstretched flame speed, Sl, can be obtained as
the intercept value at aZ0, in the plot of Sn against a. The
value of burned gas Markstein length, Lb, can also be obtained
from the plot of experimental values of Sn against a, as
mentioned in Eq. (4), using linear regression shown in
Appendix A.
For the observation limited to the early stage of the flame
expansion, where the pressure does not vary significantly yet,
there is a simple relationship linking the spatial flame velocity,
Sl, to the fundamental one, i.e. the unstretched laminar burning
velocity, ul, as
ulZrbSl=ru (5)
Here, ru is the density of the unburned gas and rb that of the
burned gas, which can be determined from a quasi-dimensional
two-zone combustion model [19].
3. Results and discussions
After the spark occurs, ignition takes place in the center of
the chamber, and the flame propagates spherically to the whole
mixture. Shown in Fig. 2 is a typical case of flame propagation,
it can be seen that the schlieren flame images show a disk
whose diameter increases with time. As far as the visualization
of flame propagation was concerned, the key aim is to obtain
the flame propagating speed. The flame image analysis can
then be performed to deduce the evolution of flame radius
versus time; hence the spatial flame speed is evaluated as the
rate of variation in flame size (Eq. (1)). In order to characterize
the flame propagation, one can plot the evolution of the flame
speed in function of the flame radius, as shown in Fig. 3. As we
know, the role of spark is to initialize a flame, which is to
overcome the tendency for the flame to quench because of the
high curvature stretch rate, during the early stage of flamepropagation. It is obvious that the flame speeds illustrate a
decrease initially, and then increase gradually as the radius
increases, because of the relatively higher flame stretch and
heat loss during the early stage of the flame propagation [7].
The effects of the fuel/air equivalence ratio and preheated
temperature on the flame speed are studied in this figure. It is
clearly shown that the increase in flame speed with radius is
Fig. 2. Typical growing schlieren flame kernels for methanolair mixture with fZ1.2. The time interval is 4 ms, TuZ358 K, PuZ0.1 Mpa.
Fig. 3. Variations of Sn with flame radius, ru for methanolair mixtures
(0.1 MPa), and (A) equivalence ratio, f, (B) initial temperature, Tu.
S.Y. Liao et al. / Fuel 85 (2006) 134613531348
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
4/8
much greater for the stoichiometric flames and high tempera-
ture can accelerate the flame propagation.
The local stretch imposed on the spherical flame can beevaluated via Eq. (3). Thereby, the spatial flame velocity can
be plotted as a function of the stretch rate, to characterize the
stretch effect on the flame propagation. Shown in Fig. 4 is a
selection of experimental data showing the variations of flame
speed, Sn with total stretch rate, a, at different equivalence
ratios and initial temperatures. According to Eq. (4), the
unstretched flame speed can be obtained by a zero stretch
extrapolation, also shown in Fig. 4 with solid lines, and the
slope of straight line indicates the burned gas Markstein length
Lb. Then, the fundamental burning velocity, unstretched
laminar burning velocity can be derived from the value of
the unstretched spatial flame speed via Eq. (5). Fig. 4 shows
that the flame accelerates with the decrease of stretch rate as the
flame propagates, which denotes positive burned gas Markstein
lengths for all tested flames. This figure also presents that the
flame speed of mixture with fZ0.8 is more sensitive to the
flame stretch than that of richer flames, relatively. However, it
also appear that there is no obvious difference between mixture
offZ1.0 and 1.2. This can be an indication that the curves of
mixture with fZ1.0 and 1.2 are relatively flatter, compared to
that of flames of fZ0.8 (Fig. 4A). The effect of initial
temperature on the flame propagation is investigated as well.
The increasing temperature results in a corresponding increase
in the flame speed. Shown in Fig. 4B is the variation of flame
speed with stretch for stoichiometric flames at different initial
temperature conditions. However, apart from the overall
magnitude of the flame speed, there appears little influence
of temperature on flame behaviour because it is indicated thatthe slopes of flame speed curves do not change obviously.
Fig. 5 exhibits the influences of fuel/air equivalence ratio
and initial temperature on the flame/stretch interaction,
namely, quantified by the burned gas Markstein lengths Lb.
With regard to the values of Lb of methanolair flames, few
experimental results have been reported previously. Our
measurements report that the value of Lb decreases slightly
with the increasing equivalence ratio, however, when fO1.0,
the value ofLb almost appears no change with the variation of
the equivalence ratio. The effect of the preheat temperature on
Lb can be neglected, as it is shown that Lb varies little for
different temperature conditions. Generally, all of the
Markstein lengths are positive. This can be demonstrated bythe fact that the measured flames are smooth, in which no
instability occurs during the atmospheric pressure experiments,
for the present range of fuel/air equivalence ratio.
Shown in Fig. 6 are the measured laminar burning velocities
for methanolair flame at 358 K and the atmospheric pressure,
over a wide equivalence ratio range. In this figure are also
plotted the results obtained previously for comparison.
Obviously, a good agreement of our measurements with
f=0.8
f=1.0
f=1.2
a/s1
Fig. 4. Variations of flame speeds with different stretch rates, and (A)
equivalence ratio, f, (B) initial temperature, Tu.
Fig. 5. Measured Markstein lengths of methanolair flames as a function of
fuelair equivalence ratio. The curves are second order fitting to measured
points.
Fig. 6. Experimental laminar burning velocities for methanolair flames at
358 K and 0.1 MPa.
S.Y. Liao et al. / Fuel 85 (2006) 13461353 1349
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
5/8
literature values of Metghalchi and Keck can be found, while
we slightly underestimate the laminar flame velocity for rich
mixtures, compared to those obtained by Saeed and Stone.
According to these results, at normal pressure of 0.1 MPa and
temperature of 358 K, the maximum burning velocity for
methanolair flame appears to be approximately between 54
and 61 cm/s for Metghalchi et al. and Saeed et al., respectively,whereas the present measured value is about 56.2 cm/s. Over
the measured range of the equivalence ratio, the present results
can be fit by a second order polynomial of the equivalence ratio
as Eq. (6), also shown in Fig. 6 with a solid curve.
ul0ZK195:6f2C419:92fK169:43 (6)
where the subscript o represents the reference conditions, i.e.
358 K and 0.1 MPa in this work. The maximum burning
velocity in Eq. (6) is about 55.9 cm/s at an equivalence ratio of
1.07, corresponding to 56.2 cm/s of experiment. Generally
speaking, one can see that, the fitting burning velocities showgood agreements with the results of experiment over a wide
range of equivalence ratio. In particular, the study of the effect
of temperature on the burning velocities has been emphasized
in this work. Fig. 7 illustrates temperature dependence of the
laminar burning velocities of methanol at atmospherical
pressure. More often, this dependence generally is expressed
as a simple power law relation of nondimensional temperature
(Tu/Tu0) and nondimensional pressure (Pu/Pu0), at the datum
conditions, as
ulZuloTu=Tu0aTPu=Puo
bP (7)
where aT and bP represent the parameters of temperature and
pressure dependences. The power law relation is also presented
in Fig. 7 as solid curves, where initial pressure of premixed
mixture has no change. It can be seen that the plotted curves of
the laminar burning velocities exhibit good agreements with
experimental data over the equivalence ratio range. The
parameter aT is clearly influenced by the fuel/air equivalence
ratio. Table 1 lists the present results for aT. Moreover, in this
table are also presented the results reported previously for
comparison. Metghalchi and Keck reported that aT is function
of the equivalence ratio, which was formulated as aTZ2.18K
0.8 (fK1) at 300 K and 0.1 MPa, while Gulder was given a
constant of 1.75. Although slight differences can be found in
our results against previously studies, partially because of thedifference of the datum temperature, a similar variation of
parameter aT against equivalence ratio can also be obtained, as
given in Eq. (8),
aTZ1:85K0:6fK1 (8)
Fig. 7 also shows the comparisons of the burning
velocities for methanolair flames at various temperatures
derived from Eq. (7) with those derived from the published
empirical formulas. On the whole, it is found that our
predictions are more consistent with those of Metghalchi and
Keck, and especially, better agreements can been obtained
for the mixture at relatively low temperature. However, theresults of Saeed and Stone gradually underestimated the
laminar burning velocities with the increase of initial
temperature.
The pressure effect on the laminar burning velocities has not
been investigated experimentally in this work. Metghalchi and
Fig. 7. Laminar burning velocity of methanolair flames on temperatures at
different equivalence ratios.
Table 1
Evolution ofaT with equivalence ratio
aT(Metghalchi and Keck)
aT(Gulder)
aT(present study)
fZ0.8 2.47 1.75 1.98
fZ1.0 2.11 1.83
fZ1.2 1.98 1.75
S.Y. Liao et al. / Fuel 85 (2006) 134613531350
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
6/8
Keck, Saeed and Stone had reported that the pressure
dependence, bP is weakly influenced by the fuel/air equival-
ence ratio. On the basis of the available data reportedpreviously (Fig. 8), we can obtain an empirical relation to
describe the dependence bP on the fuel/air equivalence ratio, as
given in Eq. (9).
bTZK0:1651C0:2fK1 (9)
Thereby, combining Eqs. (6), (8) and (9), Eq. (7)
provides an approach to empirically estimate the laminar
burning velocities of methanolair mixtures by functions of
temperature, pressure and equivalence ratio. In order to
validate these empirical formulas, it is informative to
compare its behaviour over more wide temperature and
pressure ranges, with those reported previously, as shown inFigs. 9 and 10. It is known that, high temperature can
accelerate the rate of combustion chemistry, thus, the
laminar burning velocities increase consistently with the
increase of initial temperature of combustible mixtures. In
Fig. 9, we can find that, the predictions of Metghalchi and
Keck are more sensitive to temperature changes. When
temperature varies from 358 to 480 K, the derivations
among these empirical relations relatively become signifi-
cant, showing that the predicted results of Metghalchi and
Keck are slightly overestimates the laminar burning
velocities than the others at 480 K. The effect of initial
pressure on the laminar burning velocities is presented in
Fig. 10. The general trend is that, the increase of pressure
results in the decrease of the laminar burning velocities. No
direct experimental comparison has been made herein.Similarly, the available predicted results of Metghalchi
and Keck, Saeed and Stone are presented for comparison.
The overall agreement among these results is very good
over the whole range of equivalence ratios.
It is suggested that the expression of the laminar burning
velocity in terms of the square root of an Arrhenius expression
in adiabatic flame temperature Tb with overall activation
energy E, has more significant theoretical basis than does the
power law correlations as given above [2022]. Peters and
Williams [21] has derived an asymptotic structure of the flame
that introduces the inner layer temperature T0 in the fuel
consumption, and express the activation temperature E/R in
terms of the mass flux (ruul) and adiabatic flame temperature Tbas following,
E
RZ2T2b
dlnruul
dTb(10)
where R denotes gas constant and an assumption that T0
remains constant was adopted. In practice, Eq. (10) can also be
rewritten as a alternative form as
E
RZK
d2lnruul
d1=Tb(11)
This indicates that the activation temperature E/R can be
derived directly from the linear plot of 2ln(ruul) against 1/Tb.
Meanwhile, Eq. (11) also suggests that the evaluation of E/R
requires the determination ofruul in advance. It is known that,
for a given mixture with Tu, the adiabatic flame Tb and the mass
burning flux ruul can be determined from chemical equilibrium
calculation and Eqs. (6)(9). That is to say, the power law
relations of Eqs. (6)(9) just provide an approach to determine
activation temperature of flames.
Plots of 2ln(ruul) against 1/Tb are shown in Fig. 11. The
activation temperature at each pressure is given by the slope of
appropriate straight fitting line. It is found that E/R of
methanolair flames is a weak function of temperature, while
strongly depends on initial pressure of mixtures, which appears
Fig. 8. Parameter bP for methanolair mixtures, the straight line shows the least
squares fit for Eq. (9) to points.
Fig. 9. Effect of temperature on burning velocities for methanolair flames.
Fig. 10. Effect of pressure on burning velocities for methanolair flames.
S.Y. Liao et al. / Fuel 85 (2006) 13461353 1351
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
7/8
to have a similar behaviour for methaneair [7] and iso-octane
air flames [1]. Over the studied range, E/R can be
approximately expressed by,
E
RZ2566PuC16085 (12)
With the pressure Pu in MPa, and the intercept value of
fitting lines at 1/TbZ0, C is found to be,
CZ10:88P0:253u (13)
Integration of Eq. (11) gives:
lnruulZKE
2R
1
TbC0:5C (14)
Or rearrangement gives:
ulZ exp KE
2R
1
Tb
exp0:5C
ru(15)
When E/R and Cpresumed from Eqs. (12) and (13), Eq. (15)
gives an alternative correlation of laminar burning velocities to
that presented by Eq. (7). The calculated laminar burning
velocities using Eq. (15) together with Eqs. (12) and (13) are
also plotted in Figs. 9 and 10. We can see that, the predicted
burning velocities of Eq. (15) exhibit good agreements withthose of Eq. (7), and still has no significant deviation with the
predicted results by Metghalchi and Keck, and Saeed and
Stone. Certainly, Eq. (15) is directly derived from the empirical
correlation of Eq. (7), rather than experimental measurements,
that is to say, the former can only be thought as an alternative
form of the later, the good agreement for Eq. (15) just
demonstrates that the derivation in deduction of Eq. (15) from
Eq. (7) is acceptable. However, as mentioned above, the former
has more physical relevance, which just is the significance of
this procedure.
From the values of E/R, the Zeldovich number, b can be
determined [7] as,
bZE
2RT2bTbKTu (16)
Here b represents the sensitivity of chemical reactions to the
variation of the maximum flame temperature and the inverse of
it denotes an effective dimensionless width of the reaction zone
[20]. Shown in Fig. 12 are the variations ofb against pressure
for methanolair mixtures. The values ofb is clearly influenced
by equivalence ratio, temperature and pressure. The increase of
pressure results in increasing b, because of the increase of inner
flame temperature T0 [17]. Also, it is out of question that, b
attains a minimum for increased Tu and near-stoichiometric
mixture, which has a maximum Tb. More often, another form of
Zeldovich number, Ze, defined as
ZeZbTbKTu
T0KTu(17)
The predicted results of Ze for methanolair are plotted in
Fig. 13, and the values of methaneair and iso-octaneair
mixtures are presented for comparison. The present results and
those of methaneair generally show weak dependence on
pressure than those computed values for iso-octaneair
mixtures. And methanolair gives a similar tendency against
the variation of equivalence ratio with methaneair flames,
which is obviously contrast to iso-octane mixture in air.
Fig. 12. Variations of b with the pressure for methanolair mixture with
different equivalence ratio.
Fig. 11. Plots of 2ln(ruul) against 1/Tb for methanolair mixtures. Where full
lines are linear fits through the points obtained from Eqs. (6)(9), and dashed
lines denote burning mass flux correlation of Eq. (14).Fig. 13. Variations of Ze with the pressure for methanolair mixture with
different equivalence ratio, and the results of methaneair [7] and iso-octane
air [17] are presented for comparison, where &, 0.1 MPa; B, 0.5MPa; 6,
1.0 MPa.
S.Y. Liao et al. / Fuel 85 (2006) 134613531352
-
7/30/2019 Characterization+of+Laminar+Premixed+MethanolE28093air+Flames
8/8
4. Conclusions
In this paper, centrally ignited, spherically expanding
flames have been measured using schlieren photography to
determine the influence of fuel/air equivalence ratio and
initial temperature on the laminar burning velocities of
methanolair mixtures. The stretch imposed at the flamefront has been explored experimentally and the Markstein
lengths are estimated to characterize its effect. Combined
the measurements and previous results, an empirical relation
by the function form ulZuloTu=Tu0aTPu=Pu0
bT , has been
obtained, to formulate the laminar burning velocity
dependencies on the equivalence ratio, initial temperature
and pressure. On the basis of the asymptotic theory of flame
structure, an alternative burning velocities correlation has
been derived from above power law formula, and the
expressions for mass burning flux, Zeldovich numbers have
been obtained. The comparisons of our empirical corre-
lations for burning velocities with the results reported
previously show a good agreement, which validates the
present study.
Acknowledgements
This work is supported by the state key project of
fundamental research plan of Peoples Republic of China
(No. 2001CB209206), and the key project of NSFC (No.
50156040).
Appendix A. Linear regression
Consider Npairs of computed or measured values of (a, Sn)i.
They satisfy Eq. (4):
SlKSnZLba (A1)
The value ofLb is calculated from
LbZK
PNiZ1
aiK aSniKSn
PNiZ1
aiK a2
(A2)
in which
Sn Z1
N
XNiZ1
Sni; aZ1
N
XNiZ1
ai; SlZ SnCLb a (A3)
The standard derivation is given by
eZ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
NK2
XNiZ1
SnKSn 2CLb
XNiZ1
aiK aSniKSn
" #vuut(A4)
References
[1] Lin TC, Chao MR. Sci Total Environ 2002;284:6174.
[2] Heinrich W, Marquardt KJ, Schaefer AJ. SAE paper No. 861581; 1986.
p. 9981010.
[3] Bradley D, Hicks RA, Lawes M, Sheppard CGW, Woolle R. Combust
Flame 1998;115:12644.
[4] Yamaoka I, Tsuji H. Twenty second symposium (international) on
combustion. The Combustion Institute Pittsburgh, PA; 1988. p. 1883.
[5] Haniff MS, Melvin A, Smith DB, Williams AJ. Inst Energy 1989;62:229.
[6] Bosschaart KJ, de Goey LPH. Combust Flame 2003;132:17080.
[7] Gu XJ, Haq MZ, Lawes M, Woolley R. Combust Flame 2000;121:4158.
[8] Bechtold JK, Matalon M. Combust Flame 2001;127:190613.
[9] Liao SY, Jiang DM, Gao J, Huang ZH, Cheng Q. Fuel 2004;83(10):
12818.
[10] Davis SG, Law CK. Combust Sci Technol 1998;140:42749.
[11] Metghalchi M, Keck JC. Combust Flame 1982;48:191210.
[12] Saeed K, Stone CR. Combust Flame 2004;139:15266.
[13] Gulder OL. Nineteenth symposium (international) on combustion. The
Combustion Institute Pittsburgh, PA; 1982. p. 27581.
[14] Liao SY, Jiang DM, Gao J, Huang ZH. Energy Fuels 2004;18:31626.
[15] Markstein G. Nonsteadyflame propagation. New York: MacMillan;1964.[16] Gibbs GJ, Calcote HF. J Chem Eng Data 1959;4:226.
[17] Muller UC, Bollig M, Peters N. Combust Flame 1997;108:34956.
[18] WestbrookCK, Dryer FL. Third international symposiumon alcohol fuels
techn, vol. 1; 1979.
[19] Liao SY, Jiang DM,Gao J, Zeng K. Proc InsMech EngPart D: J Automob
Eng 2003;217(D11):102330.
[20] Zeldovich YB, Barenblatt GI, Librovich VB, Makhviladze GM. The
mathematical theory of combustion and explosions. New York:
Consultants Bureau; 1985.
[21] Peters N, Williams FA. Combust Flame 1987;68:185.
[22] Gottgens J, Mauss F, Peters N. Twenty-fourth symposium (international)
on combustion. Combust Inst 1992;129.
S.Y. Liao et al. / Fuel 85 (2006) 13461353 1353