Available online at www.sciencedirect.

com

Journal of Hazardous Materials 153 (2008) 1301–1307

Calculation of the upper flammability limit of methane/air

mixtures at elevated pressures and temperatures
F. Van den Schoor a,∗ , F. Verplaetsen b , J. Berghmans a
a Katholieke Universiteit Leuven, Department of Mechanical Engineering, Celestijnenlaan 300A,
B-3000 Leuven, Belgium
b Adinex N.V., Brouwerijstraat 5/3, 2200 Herentals, Belgium

Received 30 July 2007; accepted 24 September 2007

Available online 29 September 2007

Abstract
Four different numerical methods to calculate the upper flammability limit of methane/air mixtures at initial pressures up to 10 bar and initial
temperatures up to 200 ◦ C are evaluated by comparison with experimental data. Planar freely propagating flames are calculated with the inclusion of a
radiation heat loss term in the energy conservation equation to numerically obtain flammability limits. Three different reaction mechanisms are used
in these calculations. At atmospheric pressure, the results of these calculations are satisfactory. At elevated pressures, however, large discrepancies
are found. The spherically expanding flame calculations only show a marginal improvement compared with the planar flame calculations. On the
other hand, the application of a limiting burning velocity with a pressure dependence Su,lim ∼ p−1/2 is found to predict the pressure dependence
of the upper flammability limit very well, whereas the application of a constant limiting flame temperature is found to slightly underestimate the
temperature dependence of the upper flammability limit.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Flammability limits; Methane

1. Introduction spherical [6] flames with the inclusion of a (radiation) heat loss
term in the energy conservation equation, and the application
In a previous study [1], four different methods for the cal- of a limiting burning velocity [4,7] and of a limiting flame
culation of flammability limits were evaluated based both upon temperature [8,9] – will be compared with experimental data
a comparison of their intrinsic capabilities of capturing the dif- [10].
ferent aspects of a near-limit flame relevant to its extinction
(Table 1), as upon a comparison of their results with exper-
2. Numerical methods
imental data on the (lower and upper) flammability limits of
methane/hydrogen/air mixtures at atmospheric pressure and
The calculations are performed using CHEM1D [11], a one-
ambient temperature. It was found among others that the esti-
dimensional flame code capable of solving 1D mass, energy
mates of the flammability limits become more accurate as more
and species conservation equations with detailed transport and
flame aspects are taken into account by the calculations. To fur-
chemical kinetics models. Two different flame geometries are
ther corroborate the previous findings, the focus of this study
used, namely 1D planar premixed flames, both steady and
will be on the upper flammability limits of methane/air mix-
unsteady, and quasi 1D spherically expanding premixed flames.
tures at initial pressures up to 10 bar and initial temperatures
The density, the temperature and the species mass fractions are
up to 200 ◦ C. To this end, the results of the different numer-
specified at the cold boundary, while vanishing gradients are
ical methods – namely the calculation of planar [2–5] and of
imposed at the hot boundary. Throughout the calculations it is
assumed that the gas mixture is ideal and that the pressure is con-
∗ Corresponding author. Tel.: +32 16 32 25 49; fax: +32 16 32 29 85.
stant in space and time. These are valid approximations since the
E-mail address: [email protected] compressibility factor Z ≈ 1.00 for the mixtures under consid-
(F. Van den Schoor). eration [12] and since the pressure increase during near-limit

0304-3894/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.jhazmat.2007.09.088
1302 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307

Table 1
Comparison of the different numerical methods based upon their intrinsic capa-
bilities of capturing the different aspects of a near-limit flame that are relevant
to its extinction
Flame aspects Planar Spherical Su,lim Tf,lim a
√ √ √ √
Chemical kinetics
√ √ √ √
Heat loss
√
Flow strain
√
Flame curvature
√
Natural convection
Flame front instabilities
Preferential diffusion
a Strictly speaking the application of a limiting flame temperature should

not have been included into this comparison since it is empirically based and, Fig. 1. Comparison between experimentally determined and calculated adia-
therefore, difficult to ascertain which flame aspects are taken into account by it. batic burning velocities of rich methane/air flames at 1 atm and 25 ◦ C.

flame propagation is limited to approximately 5 % of the initial nism includes a number of C3 –C6 species, while the Berkeley
pressure [10]. Further details about the flame code can be found mechanism includes aromatic species in order to better model
elsewhere [13]. soot formation.
The radiation heat loss is modelled by means of the optically These reaction mechanisms were all validated against a vari-
thin limit. Four radiating species are considered, namely CO2 , ety of experimental data, including burning velocities, ignition
H2 O, CO and CH4 . delay times and species concentration profiles [14–17]. The
Since it is expected that the chemical kinetics play an most important parameter of these regarding the calculation
important role in the limit flame behaviour, different reaction of flammability limits is the burning velocity. Therefore, the
mechanisms will be used in the flame calculations: the GRI 3.0 different mechanisms are tested on their ability of reproducing
mechanism [14], the Konnov 0.5 mechanism [15], a mechanism burning velocities of rich methane/air flames. To this end, one-
by Appel et al. [16], which will be called the Berkeley mecha- dimensional planar adiabatic flames are calculated at 1 atm and
nism in the remainder of this paper, and the Leeds 1.5 mechanism 25 ◦ C. As can be seen in Fig. 1, the calculated adiabatic burn-
[17]. ing velocities obtained with the Leeds mechanism are in poor
The GRI 3.0 mechanism is widely used. Since the nitrogen agreement with the experimentally determined values [18,19].
chemistry is unimportant for the limits under consideration, it Consequently, this mechanism is not used in any of the further
is removed from the base mechanism. The resulting mechanism calculations as it would considerably underestimate the upper
contains 219 elementary reactions among 36 species. It can only flammability limit.
be used for the calculation of methane and natural gas flames. The flammability limits for the planar flames are determined
The Konnov mechanism contains 776 elementary reactions by considering whether a transient flame calculation reaches
among 93 species (not including the nitrogen chemistry). It a steady state or not. In the latter case, the maximum flame
can be used for the calculation of hydrogen, carbon monox- temperature and the burning velocity will decrease continuously
ide, methane, methanol, C2 and C3 hydrocarbons and their in time [2]. In the spherical flame calculations, the gas mixture
derivatives. To reduce the computational time and to allow a is ignited by means of a source term in the energy conservation
better comparison with the GRI mechanism, the C3 –C6 species equation. This ignition source is modelled as an energy input
other than C3 H8 and n-C3 H7 , are removed from the Konnov of 10.5 J during a period of 40 ms in a spherical volume with a
mechanism, leading to a reduced Konnov mechanism with 456 diameter of 10 mm. These parameters are chosen to resemble the
elementary reactions among 54 species, which will be used ignition source that is used in the experiments [10] with which
alongside the full mechanism. the calculations will be compared. Mixtures in which flames
The Berkeley mechanism contains 544 elementary reactions propagate a distance of 100 mm are considered to be flammable.
among 101 species. It can be used for the calculation of methane, The planar flame calculations (with the inclusion of the radia-
ethane, ethylene and acetylene flames at 1 bar. There also exists a tion heat loss term in the energy conservation equation) will also
modification of this mechanism for calculating flames at 10 bar, be used to determine values for the burning velocity and the max-
which contains a different set of 546 elementary reactions. imum flame temperature which enable evaluation of the limiting
The Leeds 1.5 mechanism contains 175 elementary reactions burning velocity and limiting flame temperature approaches.
among 37 species. It can be used for the calculation of methane, It is difficult to ascertain the uncertainty in the calculations
ethane and ethylene flames. caused by uncertainties in the reaction rate parameters, in the
The GRI mechanism is chosen because it is widely used. The transport properties, etc. The comparison and the evaluation of
Konnov, Berkeley and Leeds mechanisms are chosen in addition the different numerical methods is, however, not biased by any
to the GRI mechanism to investigate not only the effect of dif- modelling uncertainty.
ferent reaction rate parameters, but also that of a different set of Further details about the numerical methods can be found
species. Compared to the GRI mechanism, the Konnov mecha- elsewhere [1].
 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307 1303

3. Theoretical background isobaric change of state, it is found that

ρu Tf
In the previous study [1], it was found that a constant limiting ∼ . (5)
ρb Tu
burning velocity of 5 cm/s predicts the upper flammability limit
of methane/hydrogen/air mixtures at atmospheric pressure and Since the (adiabatic) flame temperature Tf increases only slightly
ambient temperature very well. At elevated pressures, however, with increasing pressure [23], this density ratio is nearly pressure
Huang et al. [4] found that the application of a constant limiting independent. Consequently, the upward velocity v and hence the
burning velocity significantly underestimates the flammability flame stretch rate K are likewise nearly pressure independent.
limits of highly diluted n-butane/air mixtures. They argued that Since – in addition to the pressure independent flame stretch
this discrepancy is caused by the fact that the mass burning rate, rate – the Lewis number for the near-limit methane/air mixtures
rather than the burning velocity, is the eigenvalue of the flame under investigation is found to be close to one: Le = 1.1, Eq.
equations. They, thus, concluded that a constant limiting burning (3) can be used in this study to calculate the limiting burning
velocity should not be used as a means to calculate the pressure velocity at elevated pressures based upon its value at atmospheric
dependence of the flammability limits. Although they arrived at pressure.
the right conclusion, they did not acknowledge the derivation of
Buckmaster and Mikolaitis [20] for the limiting burning velocity, 4. Results
which clearly shows its pressure dependence.
For a two-reactant mixture and a single-step reaction and for 4.1. Planar and spherical flame calculations
a flame propagating in an axisymmetric stagnation point flow,
Buckmaster and Mikolaitis calculated the limiting (minimum) At an initial pressure of 1 atm, the one-dimensional planar
value of the Karlovitz number Kalim that leads to extinction: flame calculations show satisfactory agreement with the exper-
  iments (Fig. 2). All of the tested reaction mechanisms predict
 
π 1 1 Ea the slope of the temperature dependence of the upper flamma-
Kalim = exp − 1− (Tf − Tu ) (1) bility limit well, whereas the observed differences between the
4 2 Le RTf2
calculated and the experimentally determined values are approx-
with Le the Lewis number, Ea the overall activation energy, imately 2 mol.%. At an initial pressure of 10 bar, however, larger
R the universal gas constant, Tf the flame temperature and Tu differences are found between the numerical and experimen-
the temperature of the unburned gases. Based upon Eq. (1), an tal results as well as large discrepancies between the different
expression for the limiting burning velocity Su,lim can be derived reaction mechanisms (Fig. 3). Moreover, all of the reaction
[20,21]: mechanisms overestimate the slope of the temperature depen-
dence. At an initial temperature of 25 ◦ C, it can be seen that the
   
1 1 Ea Konnov mechanism reproduces the slope of the pressure depen-
Su,lim ∼ exp 1− (Tf − Tu ) α1/2 K1/2 (2) dence fairly well, whereas the results of the GRI mechanism
4 Le RTf2 substantially diverge from the experimental ones (Fig. 4). These
large differences between the calculated and the experimentally
with α the thermal diffusivity and K the flame stretch rate. determined values at higher pressures and temperatures imply
By assuming in Eq. (2) that the Zeldovich number Ze = that the estimates of the upper flammability limits obtained by
Ea /[RTf2 (Tf − Tu )] is (nearly) pressure independent or that Le one-dimensional planar flame calculations are of limited value
is (close to) 1, and that K is (nearly) pressure independent, the at these conditions.
pressure dependence of the limiting burning velocity Su,lim is Because of computational difficulties, only a limited number
given by of spherical flame calculations have been performed using the
GRI mechanism (Table 2). They show only a slight lowering
Su,lim ∼ α1/2 ∼ p−1/2 (3)

with p the pressure.

The assumption of a pressure independent flame stretch rate
K is based upon the following considerations. For near-limit
flames, which only propagate upwards, the flame stretch rate
is proportional to the upward velocity of the flame kernel.
Crescitelli et al. [22] found that the upward velocity v of a
spherical flame kernel with neglect of the drag force is given
by

((ρu /ρb ) − 1)gt

v= (4)
4(1 + (1/2)(ρu /ρb ))

with ρu and ρb the densities of the unburned and burned gases Fig. 2. Comparison between experimentally determined and calculated upper
respectively, g the gravitational acceleration and t time. For an flammability limits of methane/air mixtures at 1 atm.
1304 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307

Table 2
Comparison of upper flammability limit values (mol.%) of methane/air mixtures
obtained from planar and spherical flame calculations with experimental data
Planar flame Spherical flame Experimental [10]

1 atm 25 ◦ C 17.9 16.6 ± 0.4a 16.0 ± 0.4

3 bar 25 ◦ C 21.5 20.8 ± 0.3 17.4 ± 0.4
6 bar 25 ◦ C 28.9 28.0 ± 0.3 19.4 ± 0.4
1 atm 100 ◦ C 19.2 18.1 ± 0.4 17.0 ± 0.4
3 bar 100 ◦ C 23.4 22.7 ± 0.3 19.0 ± 0.4
1 atm 200 ◦ C 21.0 20.1 ± 0.3 18.8 ± 0.4
a The uncertainties on the results of the spherical flame calculations stem from

the concentration step size used in the calculations.

Fig. 3. Comparison between experimentally determined and calculated upper

flammability limits of methane/air mixtures at 10 bar. Fig. 5 shows the reactions that exhibit the largest positive
and negative sensitivities, meaning that variations in their reac-
(of about 1 mol.%) of the calculated upper flammability limits, tion rates have the largest impact on the mass burning rate. The
when referred to those of the planar flame calculations. At higher important chain branching reaction
pressures and temperatures, however, this small improvement
H + O2  O + OH
does not outweigh the increased computational effort.
The observed discrepancies between the results obtained is found to have a large positive sensitivity for all three reaction
with the different reaction mechanisms merit further analysis. mechanisms. Nevertheless, it is not the most important reaction
Therefore, a sensitivity analysis and a reaction path analysis are in the Berkeley and GRI mechanisms as the reaction
performed for a methane/air flame with a methane concentra-
tion of 25.7 mol.% at an initial pressure of 10 bar and an initial CH3 + HO2  CH3 O + OH
temperature of 25 ◦ C.
Sensitivity analysis can be used to identify the reactions that has a larger positive sensitivity. The complete sensitivity anal-
influence the flame behaviour the most when the flame temper- ysis shows an increased importance of reactions involving the
ature is lowered. This is done by calculating the sensitivity of HO2 radical at elevated initial pressures. Egolfopoulos et al.
the mass burning rate ṁ on the pre-exponential reaction rate [5] reached the same conclusion for lean methane/air flames.
parameter A for all reactions Furthermore, it can be seen that the reaction
∂ ln(ṁ) CH3 + O2  CH2 O + OH
.
∂ ln(A)
plays an important role in the Konnov mechanism, whereas
The influence of the flame temperature on the reaction rate is, its role in the Berkeley and GRI mechanisms is rather lim-
thus, simulated by a change in A. ited. This finding is corroborated by the results of the reaction
In a reaction path analysis, the contributions of all reactions path analysis, which show that in the Konnov mechanism
to the consumption (or formation) of a chemical species are CH3 → CH3 O → CH2 O and CH3 → CH2 O are equally impor-
determined by the production rates. This allows the identification tant pathways, whereas in the GRI mechanism only the first is
of the key reactions in the transformation of the reactants into important (Fig. 6).
the products and the subsequent construction of a reaction path
diagram.

Fig. 5. Sensitivity of the mass burning rate on the reaction rate parameters of
Fig. 4. Comparison between experimentally determined and calculated upper different reaction mechanisms for a methane/air flame with a methane concen-
flammability limits of methane/air mixtures at 25 ◦ C. tration of 25.7 mol.% at 10 bar and 25 ◦ C.
 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307 1305

Fig. 7. Comparison of the experimentally determined upper flammability limit

(markers) [10] and the pressure dependence estimated based upon the limiting
burning velocity approach (solid lines) for methane/air flames.

imental data. This implies that the GRI mechanism can only
Fig. 6. Comparison of the main reaction pathways of the GRI (a) and Konnov
(b) mechanism in the oxidation of CH4 to CO and CO2 for a methane/air flame
be used confidently within the pressure and temperature range
with a methane concentration of 25.7 mol.% at 10 bar and 25 ◦ C. of the experimental data used for its construction. The Konnov
mechanism, however, can be used with more confidence outside
The reactions that exhibit the largest negative sensitivity are its validation range. This is the primary reason why the Konnov
the same for all three-reaction mechanisms, namely mechanism will be used in the subsequent calculations.
CH3 + CH3 (+M)  C2 H6 (+M)
4.2. Limiting burning velocity
and
At atmospheric pressure and ambient temperature, a value of
HO2 + HO2  H2 O2 + O2 .
4.8 cm/s is found for the burning velocity at the experimentally
There is, however, a large difference between the GRI mecha- determined upper flammability limit using the reduced Konnov
nism and the Konnov mechanism regarding the sensitivity of the mechanism. The theoretical derivation in Section 3 gives the
first reaction. pressure dependence of the limiting burning velocity (Eq. (3)).
The reaction path analysis shows that despite the large dif- Using this equation, an estimate of the upper flammability limit
ference in the number of species between the GRI (36) and the at elevated pressures can be obtained starting from the exper-
Konnov (93) mechanism, the main pathways for the oxidation imentally determined value at atmospheric pressure. First, the
of CH4 to CO and CO2 are very similar (Fig. 6). The pres- burning velocity at the experimentally determined flammabil-
ence of C3 –C6 species in the Konnov mechanism is, thus, of ity limit at atmospheric pressure is calculated. Next, Eq. (3) is
minor importance for the calculation of rich methane/air flames. used to calculate the limiting burning velocity at the desired ini-
Therefore, it is not surprising that the results of the reduced Kon- tial pressure and finally, the corresponding fuel concentration
nov mechanism – from which these species were removed – are is calculated. Fig. 7 shows a comparison between the experi-
nearly equal to those of the full mechanism (Figs. 2–4). This also mental data, indicated by the markers, and the results of this
means that the large differences between the upper flammability calculation, indicated by the full lines. Given the inaccuracies
limits calculated with the GRI and the Konnov mechanism that in the experimental determination of the flammability limits and
are observed at elevated pressures (Fig. 4) are probably caused in the calculation of burning velocities, as well as the assump-
by a difference in reaction rate parameters, rather than by a tions made in the derivation of the limiting burning velocity, the
difference in the set of elementary reactions. agreement is very good.
The observation that the results obtained with the Konnov Whereas the pressure dependence of the limiting burning
mechanism agree better with the experimental data does not velocity can easily be deduced, its temperature dependence
necessarily mean that this reaction mechanism is to be preferred is more complicated. For example, the flame stretch rate K
over the GRI mechanism, given the large differences between experienced by a flame kernel propagating upwards cannot
the numerical and experimental flames. Unfortunately, no exper- be assumed to be temperature independent. Therefore, another
imental data are available on planar or spherical methane/air approach is used to ascertain whether the exponent a of the
flames near the upper flammability limit at elevated pressures. power dependence Su,lim ∼ Ta is nearly constant for different
However, there is an important difference in how these mech- pressures by calculating the burning velocity at the experimen-
anisms were constructed that determines their applicability at tally determined upper flammability limits and by subsequently
elevated pressures. Whereas the reaction rate parameters of the fitting the power dependence to these data points by means
GRI mechanism were fitted to the experimental data, those of the least squares method. a is found to lie in the range
of the Konnov mechanism were determined a priori and only 0.96–1.18 for pressures up to 10 bar. Compared to the pres-
afterwards was the mechanism validated by means of the exper- sure dependence for which a similar exercise with Su,lim ∼ pb
1306 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307

Table 3 flammability limits decrease by approximately 150 ◦ C in the

Calculated flame temperature (K) at the experimentally determined upper pressure range 1–10 bar (Table 3).
flammability limits of methane/air mixtures at different initial pressures and
temperatures
5. Discussion
298 K 373 K 473 K

1 atm 1692 1678 1667 The findings of this study corroborate those of the previous
3 bar 1648 1622 1612 study [1]. Specifically, they show that the concept of a limit-
6 bar 1597 1583 1584 ing burning velocity (derived from a limiting Karlovitz number)
10 bar 1545 1533 1524
has the best potential for accurately calculating the flammability
limits. This is not surprising since – as was described in detail in
the previous study – this approach takes most of the aspects of a
gives b = −0.50 ± 0.02, the spread on a is too large to be use-
near-limit flame that are relevant to its extinction into account,
ful. This comparison, however, only shows that the temperature
as can be seen in Table 1. It must, however, be borne in mind –
dependence of the limiting burning velocity does not follow a
as can be deduced from Table 1 – that none of the methods will
simple power dependence, as opposed to its pressure depen-
give satisfactory results if flame front instabilities are present.
dence. Consequently, more research is necessary to explore the
Nevertheless, this study has shown that the concept of a limiting
applicability of the limiting burning velocity approach to pre-
burning velocity cannot only be used to estimate the flammabil-
dict the temperature dependence of the flammability limits. In
ity limits at atmospheric pressure and ambient temperature, but
the next section, it will be shown that the limiting flame tempera-
also at elevated pressures (and temperatures).
ture concept offers a better potential to estimate the temperature
dependence.
6. Conclusions

4.3. Limiting flame temperature It is found that for rich methane/air mixtures:

Calculation of the maximum flame temperatures at the exper- (1) At atmospheric pressure, the calculation of the upper
imentally determined upper flammability limits shows that the flammability limit by calculating planar flames with the
limit flame temperature remains nearly constant for a given inclusion of radiation heat loss is satisfactory. The differ-
initial pressure, independent of initial temperature (Table 3). ences between the reaction mechanisms are minimal.
Using the concept of a constant limiting flame temperature, (2) At elevated pressures, the calculated upper flammability
an estimate of the upper flammability limit at elevated tem- limit values are significantly too high and large differences
peratures can be obtained starting from the experimentally are found between the different reaction mechanisms. The
determined value at ambient temperature. Fig. 8 shows a compar- results obtained with the (reduced) Konnov mechanism
ison between the experimental data, indicated by the markers, show the best agreement with the experimental data.
and the results of this calculation, indicated by the full lines. (3) The spherical flame calculations show a marginal improve-
It can be seen that the use of a constant limiting flame tem- ment compared with the planar flame calculations.
perature to estimate the temperature dependence of the upper (4) The application of a limiting burning velocity with a pres-
flammability limit leads to slight underestimations at elevated sure dependence Su,lim ∼ p−1/2 gives good agreement with
temperatures. Moreover, it must be emphasised that using this the experimental data.
approach to estimate the pressure dependence would lead to (5) The application of a constant limiting flame temperature
large underestimations at elevated pressures as the calculated slightly underestimates the temperature dependence of the
flame temperatures at the experimentally determined upper upper flammability limit, while it considerably underesti-
mates its pressure dependence.

Acknowledgements

A preliminary version of this paper was presented at the

12th International Symposium on Loss Prevention and Safety
Promotion in the Process Industries [24].

References

[1] F. Van den Schoor, R.T.E. Hermanns, J.A. van Oijen, F. Verplaetsen, L.P.H.
de Goey, Comparison and evaluation of methods for the determination of
flammability limits, applied to methane/hydrogen/air mixtures, J. Hazard.
Mater. 150 (2008) 573–581.
Fig. 8. Comparison of the experimentally determined upper flammability limit [2] K.N. Laksmisha, P.J. Paul, H.S. Mukunda, On the flammability limit and
(markers) [10] and the temperature dependence estimated based upon the limit- heat loss in flames with detailed chemistry, in: Proceedings of the 23rd
ing flame temperature approach (solid lines) for methane/air flames. Symposium on Combustion, The Combustion Institute, 1990, pp. 433–440.
 F. Van den Schoor et al. / Journal of Hazardous Materials 153 (2008) 1301–1307 1307

[3] J. Yiguang, G. Masuya, P.D. Ronney, Effects of radiative emission and V. Lissianski, Z. Qin, GRI 3.0. Available from: http://www.me.berkeley.
absorption on the propagation and extinction of premixed gas flames, in: edu/gri-mech.
Proceedings of the 27th Symposium on Combustion, The Combustion [15] A.A. Konnov, Detailed reaction mechanism for small hydrocarbons
Institute, 1998, pp. 2619–2626. combustion. Release 0.5. Available from: http://homepages.vub.ac.be/∼
[4] Y. Huang, C.J. Sung, J.A. Eng, Dilution limits of n-butane/air mixtures akonnov.
under conditions relevant to HCCI combustion, Combust. Flame 136 (2004) [16] J. Appel, H. Bockhorn, M. Frenklach, Kinetic modeling of soot forma-
457–466. tion with detailed chemistry and physics: laminar premixed flames of
[5] F.N. Egolfopoulos, A.T. Holley, C.K. Law, An assessment of the lean C2 hydrocarbons, Combust. Flame 121 (2000) 122–136, http://www.me.
flammability limits of CH4 /air and C3 H8 /air mixtures at engine-like con- berkeley.edu/soot/mechanisms/mechanisms.html.
ditions, in: Proceedings of the 31st Symposium on Combustion, The [17] K.J. Hughes, T. Turányi, A.R. Clague, M.J. Pilling, Development and test-
Combustion Institute, 2007, pp. 3015–3022. ing of a comprehensive chemical mechanism for the oxidation of methane,
[6] M. Sibulkin, A. Frendi, Prediction of flammability limit of an unconfined Int. J. Chem. Kinet. 33 (2001) 513–538, http://www.chem.leeds.ac.uk/
premixed gas in the absence of gravity, Combust. Flame 82 (1990) 334–345. Combustion/Combustion.html.
[7] M.N. Bui-Pham, J.S. Miller, Rich methane/air flames: burning velocities, [18] X.J. Gu, Laminar burning velocity and Markstein lengths of methane-air
extinction limits, and flammability limit, in: Proceedings of the 25th Sym- mixtures, Combust. Flame 121 (2000) 41–58.
posium on Combustion, The Combustion Institute, 1994, pp. 1309–1315. [19] K.J. Bosschaert, L.P.H. de Goey, The laminar burning velocity of flames
[8] I. Wierzba, S.O. Bade Shrestha, G.A. Karim, An approach for predicting propagating in mixtures of hydrocarbons and air measured with the heat
the flammability limits of fuel/diluent mixtures in air, J. I. Energy 69 (1996) flux method, Combust. Flame 136 (2004) 261–269.
122–130. [20] J. Buckmaster, D. Mikolaitis, A flammability limit model for upward prop-
[9] Y.N. Shebeko, W. Fan, I.A. Bolodian, V.Y. Navzenya, An analytical agation through lean methane/air mixtures in a standard flammability tube,
evaluation of flammability limits of gaseous mixtures of combustible- Combust. Flame 45 (1982) 109–119.
oxidizer-diluent, Fire Safety J. 37 (2002) 549–568. [21] P.D. Ronney, Understanding combustion processes through microgravity
[10] F. Van den Schoor, F. Verplaetsen, The upper flammability limit of research, in: Proceedings of the 27th Symposium on Combustion, The
methane/hydrogen/air mixtures at elevated pressures and temperatures, Int. Combustion Institute, 1998, pp. 2485–2506.
J. Hydrogen Energy 32 (2007) 2548–2552. [22] S. Crescitelli, G. Russo, V. Tufano, F. Napolitano, L. Tranchino, Flame
[11] CHEM1D, A One-dimensional Laminar Flame Code, Eindhoven Univer- propagation in closed vessels and flammability limits, Combust. Sci. Tech-
sity of Technology. Available from: http://www.combustion.tue.nl/chem1d. nol. 15 (1977) 201–212.
[12] Y.A. Çengel, M.A. Boles, Thermodynamics: An Engineering Approach, [23] C.K. Law, Propagation, structure, and limit phenomena of laminar flames
third ed., WCB/McGraw-Hill, Boston, 1998. at elevated pressures, Combust. Sci. Technol. 178 (2006) 335–360.
[13] B. Somers, The Simulation of Flat Flames with Detailed and Reduced [24] F. Van den Schoor, F. Verplaetsen, J. Berghmans, Calculation of the
Chemical Models, Ph.D. Thesis, Eindhoven University of Technology, upper explosion limit of methane-air mixtures at elevated pressures
1994. and temperatures, in: Proceedings of the 12th International Symposium
[14] G.P. Smith, D.M. Golden, M. Frenklach, N.W. Moriarty, B. Eiteneer, on Loss Prevention and Safety Promotion in the Process Industries,
M. Goldenberg, C.T. Bowman, R. Hanson, S. Song, W.C. Gardiner Jr., 2007.