Applied Energy 179 (2016) 451–462

Contents lists available at ScienceDirect

Applied Energy

journal homepage: www.elsevier .com/locate /apenergy

Exhaust gas recirculation effects on flame structure and laminar burningspeeds of H2/CO/air flames at high pressures and temperatures

http://dx.doi.org/10.1016/j.apenergy.2016.06.1180306-2619/� 2016 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.

Omid Askari a, Kevin Vien b, Ziyu Wang b, Matteo Sirio c, Hameed Metghalchi b,⇑aMechanical Engineering Department, Mississippi State University, Starkville, MS 39762, USAbDepartment of Mechanical and Industrial Engineering, Northeastern University, Boston, MA 02115-5000, USAcDepartment of Mechanical and Industrial Engineering, Università degli Studi di Brescia, Brescia 25123, Italy

h i g h l i g h t s

� Effect of exhaust gas recirculation onflame morphology of syngas/air isinvestigated.

� Hydrodynamic and thermo-diffusiveeffects are evaluated on flamestability analysis.

� Burning speeds are calculated using anew differential based multi-shellmodel.

� Effect of exhaust gas recirculation onburning speed of syngas/air isinvestigated.

� Power law correlations are developedfor easy use in numerical simulations.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 April 2016Received in revised form 22 June 2016Accepted 25 June 2016

Keywords:SyngasExhaust gas recirculationLaminar burning speedFlame stabilitySchlieren photographyHigh pressure and temperatureMulti-shell

a b s t r a c t

Experimental studies have been performed in conjunction with a novel differential based multi-shellmodel to investigate the flame structure and measure laminar burning speeds of H2/CO/air/diluent pre-mixed flames at high pressures and temperatures. This paper focuses on synthetic gas (syngas) as the fuelblend, which is a mixture of H2 and CO, and investigates the effect of synthetic exhaust gas recirculation(SEGR) as the diluent on flame structure and laminar burning speed. SEGR is a mixture of 14% CO2 and86% N2. In these experiments two different SEGR concentrations of 5% and 10% have been used. Theexperiments were performed in two constant volume spherical and cylindrical chambers. The cylindricalchamber was set up in a schlieren system equipped with a high speed CMOS camera, capable of takingpictures up to 40,000 frames per second, which was used to study the structure and stability of the flame.The laminar burning speed of the combustion process was calculated from the pressure rise measure-ment during flame propagation in spherical chamber. Power law correlations have been developed forlaminar burning speeds of smooth H2/CO/air/SEGR flames over a wide range of temperatures (298 Kup to 450 K), pressures (from sub-atmospheric up to 5.5 atm), equivalence ratios (/ = 0.6–3) and threedifferent hydrogen concentration of 5%, 10% and 25% in the fuel mixture. SEGR lowers the laminar burn-ing speed and has significant effect on the flame stability compared to H2/CO/air, especially for very leanand very rich mixtures. Experimental burning speeds of H2/CO/air/SEGR mixtures have been comparedwith available measurements as well as computed values obtained by 1D free flame simulations using

452 O. Askari et al. / Applied Energy 179 (2016) 451–462

two chemical kinetics mechanisms. Very good agreements have been observed with the experimentaldata available in the scientific literature as well as computational burning speeds.

� 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Synthetic gas, also known as syngas, is fundamentally a mixtureof hydrogen and carbon monoxide gases along with various otherhigher-order hydrocarbons. Syngas is considered an alternativefuel since it can be created through various sources such as bio-mass gasification, reactions that involve natural gas and coal, aswell as the recycling of stationary turbine byproducts. With theadvent of integrated gasification combined cycle (IGCC) technol-ogy, syngas can be created from coal with lower emissions. Thus,the development and research pertaining to syngas fuels arebecoming more relevant amid growing concerns about pollutantsand carbon emissions.

Syngas is considered as a strong candidate to replace many fuelscurrently in use, therefore, it is imperative to fully understand andcharacterize how syngas behaves in various conditions. The lami-nar burning speed adequately characterizes a fuel and provides agood indicator of how a fuel performs. It is widely used and con-tains information about a mixture’s exothermicity, diffusivity,and reactivity. It is also important to study the laminar burningspeed [1–12] in a high pressure environment as well as with differ-ent diluents, since those are normally gas turbine and internalcombustion engines [13] relevant conditions. One typical form ofdiluent is the inert gas used in exhaust gas recirculation (EGR)technique commonly used in automobile engines [13], which isprimarily a mixture of carbon dioxide, nitrogen, water and otherproducts of combustion.

There is a wide assortment of literature on the laminar burningspeed of syngas fuels with and without diluent. Hassan et al. [14]measured the laminar burning speed of various hydrogen to carbonmonoxide ratios (3:97, 5:95, 10:90, 25:75, 50:50), sub-atmosphericto elevated pressures (0.5–4 atm), atmospheric temperature, andwide equivalence ratio (0.6–5.0) in a spherical combustion cham-ber. Sun et al. [15] used a dual-cylindrical chamber to extract lam-inar burning speed data at atmospheric temperature from manydifferent H2/CO ratios (1:99, 5:95, 25:75, 50:50), elevated pressures(up to 40 atm), and equivalence ratios (0.5–5.0). Sun et al. [15] alsoreplaced nitrogen with helium as the diluent in order to increasethe stability of flames, which allowed them to obtain data for muchhigher pressures. Natarjan et al. [16] used the burner and particlevelocimetry technique to measure laminar burning speed for5:95, 50:50, and 95:5 syngas percentages dilutedwith CO2 and laterextended measurements to higher pressures with helium substitu-tion to reduce flame instability [17]. Prathap et al. [18] used theconstant pressure method to measure the laminar burning speedof 50:50 H2/CO mixtures diluted by N2 and later CO2 [19]. Vuet al. [20] compared the effects of CO2, N2, and He as diluents onthe cellular instabilities in syngas flames in a cylindrical chamberat elevated pressures for a 50:50 hydrogen to carbon monoxideratio, and found that helium suppresses instabilities. Burbanoet al. [21] used the burner method to extend the data on the effectsof CO2 and N2 dilution on laminar burning speed and stability over awider equivalence ratio (0.6–4.3). Lapalme and Seers [22] investi-gated the effect of initial temperature (up to 450 K) and carbondioxide and methane dilution on the laminar burning speeds ofsyngas flames, as well as provided a correlation based on their data.Han et al. [23] measured laminar burning speeds for various CO2

diluent percentages (10–40%) at elevated temperatures andpressures for equivalence ratios of / ¼ 0:8 and / ¼ 1:0 using a

dual-cylindrical setup. Wang et al. [24] reported laminar burningspeed data for awide range of equivalence ratio (0.6–5.6), hydrogenpercentages (5–75%), and N2 or CO2 percentages (0–60%) using aheat flux burner as well as a Bunsen flame in conjunction withOH-PLIF method. Askari et al. [25] measured the laminar burningspeeds of H2/CO/air flames using a new differential based multi-shell model over a wide range of temperatures (298 K up to617 K), pressures (from sub-atmospheric up to 5.5 atm), equiva-lence ratios (0.6–5) and three different hydrogen concentrationsof 5%, 10% and 25% respectively. They concluded when the initialpressure increases, the tendency for the flame to destabilize takesplace earlier due to a significant decrease of the flame thicknessand enhancement of hydrodynamic instability. They measuredthe laminar burning speeds for smooth flames using the pressurerise method and developed power law correlations [25]. There isa wide range of scientific literature that is currently available onthe laminar burning speed of syngas flames diluted with varyingpercentages of CO2 or N2, but no literature exists for syngas flamesdiluted exactly with both 14% CO2 and 86% N2 to simulate theexhaust gas recirculation in internal combustion engines.

The present study investigates the effect of synthetic exhaustgas recirculation, with the composition of 14% CO2 and 86% N2,on the stability and laminar burning speeds of H2/CO/air flames.Since creating the real EGR, which is the engine post-combustionexhaust gases in our lab is impossible, a synthetic EGR with afore-mentioned composition which has the same specific heat as realEGR is used. The effect of SEGR addition (5% and 10%) to H2/CO/air on flame morphology, flame stability and laminar burningspeed has been studied in a wide range of temperatures, pressuresand equivalence ratios for three hydrogen concentrations. In thispaper laminar burning speeds of H2/CO/air/SEGR mixtures andtheir correlations are reported over a wide range of temperatures(298 K up to 450 K), pressures (from sub-atmospheric up to5.5 atm), equivalence ratios (0.6–3) and three different hydrogenconcentration of 5%, 10% and 25% in the fuel mixture.

2. Experimental facilities

Experiments have been performed using a cylindrical chamberto study the morphology and stability of the flame and a sphericalchamber for laminar burning speeds measurement. The cylindricalchamber is 13.5 cm in diameter and 13.5 cm in length. The cylin-drical chamber is equipped with fused quartz windows that aresealed to the chamber with two high temperature elastomer O-rings. The cylindrical chamber is set up in a Z-shape schlieren sys-tem equipped with a high speed CMOS camera, capable of takingpictures up to 40,000 frames per second [26,27]. Two band heatersare installed in order to raise the initial temperature of the systemup to 500 K. Both chambers are fitted with two extended automo-tive spark plugs, and K-type thermocouples to measure the tem-perature of the inside gas mixtures. The spark energy has beentuned to be sufficiently close to the minimum ignition energy tominimize the effect of spark discharge on flame expansion [28].Fig. 1 shows the general configuration of the experimental set up.

The spherical chamber is made of stainless steel that can with-stand pressures up to 400 atm. The spherical chamber is built usingtwo hemispheres with a diameter of 15.24 cm. It is placed inside anoven to heat up the chamber up to the initial temperature of 500 K.The pressure rise inside the spherical chamber was measured using

Fig. 1. Schematic diagram of experimental facilities and Schlieren system.

O. Askari et al. / Applied Energy 179 (2016) 451–462 453

a Kistler high sensitivity pressure sensor [26,27]. The chamberswere filled by the method of partial pressures using a manifoldsupply system comprised of valves, high accuracy pressure trans-ducers, pipes connected to the respective mixture constituents,and a vacuum pump. A gas chromatography (GC) system was usedto verify the composition of premixed fuel inside the chamber. Adata acquisition system collected and synchronized pressure-time data as well as flame propagation images. A LabView programwas used to initiate the combustion process. After filling the cham-ber with fuel, air and SEGR the system was given at least 5 min tomake sure that the mixture is quiescent. At each operational con-dition experiments were done using both chambers. Cylindricalchamber experiments yielded flame images that were used tostudy the morphology and stability of the flame as well as to verifythat the flame is smooth, laminar, and spherical. The same exper-imental conditions were repeated in the spherical chamber inorder to collect the pressure rise data and check its reproducibility.In this work each experiment was carried out at least three timesat each initial condition to ensure that the confidence level of theexperiments were above 95% [29]. Only smooth, laminar, andspherical flames were used to calculate the laminar burning speed.More information about experimental facilities can be found inother studies [2,10,26,27,30–33].

3. Flame morphology study and stability analysis

The effect of synthetic EGR (SEGR) addition on the formation andgrowth of cellular instabilities over the flame surface of H2/CO/airmixtures was investigated in detail. Experiments with SEGR (14%CO2 + 86% N2) have been done with two different concentrationsof 5% and 10%. The initial conditions of the experiments were fixedat ambient temperature of 298 K, pressures of 0.5, 1, and 2 atm,equivalence ratios of 0.6, 1, 2 and 3 and three different hydrogenconcentrations of 5%, 10% and 25%. Hydrodynamic and thermo-diffusive effects are two major kinds of instabilities that occur inpremixed flames [25]. A complete flame structure and instabilityanalysis including the definitions and formulations on H2/CO/airmixtures has been explained in a previous publication [25].

Fig. 2 shows the snapshots of the expanding spherical flame fordifferent concentrations of SEGR at hydrogen concentrations of25%, initial temperature of 298 K, initial pressure of 1 atm and var-ious equivalence ratios. The combustion duration times are indi-cated under each image. As it can be seen in Fig. 2 the additionof SEGR increases the flame stability for all equivalence ratiosexcept for stoichiometric case. Both hydrodynamic and thermo-diffusive instabilities have a significant effect on flame structureas shown in Fig. 3. For lean mixtures ð/ ¼ 0:6Þ the effective Lewisnumber [25] is less than unity ðLeeff < 1Þ resulting in a negativeeffect on stability. The amount of reduction of effective Lewis num-ber by increasing the SEGR concentration is negligible as shown inFig. 3. Simultaneously, the increase of flame thickness and reduc-tion of burning speed suppress the hydrodynamic instability andpromote flame stability. Fig. 3 shows that increasing SEGR concen-tration has a positive effect on flame stability for lean mixturesð/ ¼ 0:6Þ. Increasing SEGR concentration results in an increase inflame thickness, which overcomes the thermo-diffusive instabilityand results in a more stable flame. Fig. 3 also shows that for thestoichiometric case ð/ ¼ 1:0Þ the positive effect of flame thicknessand the negative effect of thermo-diffusive instability ðLeeff < 1Þ,which are similar in magnitude, cancel each other out. For this rea-son the flame stability doesn’t drastically change with increasingSEGR concentration. In the case of rich mixtures ð/ ¼ 2:0� 3:0Þthe effect of thermo-diffusive instability will be overcome by theincrease in effective Lewis number ðLeeff > 1Þ. At the same timethe hydrodynamic instability is weakened by increasing the flamethickness which enhances the tendency of the flame to stabilizedue to increasing SEGR concentration.

The effects of initial pressure and flame radius on the destabi-lization of the flame front are shown in Fig. 4. This figure showsthe snapshots of the expanding spherical flame with increasingthe initial pressures at hydrogen concentration of 25%, initial tem-perature of 298 K, SEGR concentration of 10% and equivalence ratioof 2. As it can be seen in these snapshots, only laminar and smoothflames are observed for pi ¼ 0:5 atm, while the flame surfacesbecome cellular and unstable as initial pressure increases. As dis-cussed in detail in previous a publication [25] and shown in

= % = % = %

= .

43.04 ms 47.89 ms 59.22 ms

= .

15.21 ms 16.83 ms 20.39 ms

= .

8.74 ms 10.36 ms 12.29 ms

= .

10.03 ms 13.59 ms 16.18 ms

Fig. 2. Snapshots of the H2/CO/air/SEGR flames for various SEGR concentrations and equivalence ratios at hydrogen concentration of 25%, initial temperature of 298 K andinitial pressure of 1 atm.

0 5 10

0.5

0.7

0.9

1.1

1.3

Effe

ctiv

e Le

wis

num

ber

SEGR (%)

0.1

0.2

0.3

0.4

0.5

Flam

e th

ickn

ess

(mm

)

φ = 0.6φ = 1.0φ = 2.0φ = 3.0

Fig. 3. Effective Lewis number and flame thickness of the H2/CO/air/SEGR flamescorresponding to the snapshots of Fig. 2.

454 O. Askari et al. / Applied Energy 179 (2016) 451–462

Fig. 5 the effective Lewis number is not sensitive to pressurechanges. It means that flame cellularity due to the increase in ini-tial pressure is not related to thermo-diffusive instability. It is

solely associated to hydrodynamic instability through the reduc-tion of flame thickness as shown in Fig. 5. As pressure increases,flame thickness decreases, which reduces its resistance to pertur-bations associated with thermal expansion ratio across the flame[34]. On the other hand increase in flame radius promotes boththermo-diffusive and hydrodynamic instabilities through thereduction in effective Lewis number and flame thickness, respec-tively. As shown in Figs. 4 and 5, effect of hydrodynamic instabilityis greater than thermo-diffusive, which appears in the form ofsmall cells all over the flame surface.

Fig. 6 shows a series of snapshots of stoichiometric H2/CO/air/SEGR flames for initial pressure of 2 atm, SEGR concentration of5%, initial temperature of 298 K and various hydrogen concentra-tions of 5%, 10% and 25%. The figure shows that the tendency forinstability is increased with increasing hydrogen concentration inthe fuel blend. This can be explained by the reduction of Lewisnumber and flame thickness which promotes the impact ofthermo-diffusive and hydrodynamic instabilities, respectively, asdescribed earlier in a previous publication [25]. Fig. 7 shows thecritical Peclet number as a function of equivalence ratio for threedifferent hydrogen concentrations at initial pressure of 2 atm,initial temperature of 298 K and SEGR concentration of 5%. The

= . = =

=

=

=

=

Fig. 4. Snapshots of the H2/CO/air/SEGR flames for various initial pressures and flame radii at hydrogen concentration of 25%, initial temperature of 298 K, SEGR concentrationof 10% and equivalence ratio of 2.0.

10 20 30 40 50 60 70

0.91

1.11.21.31.4

Effe

ctiv

e Le

wis

num

ber

Flame Radius (mm)

10 20 30 40 50 60 7000.20.40.60.811.2

Flam

e th

ickn

ess

(mm

)

pinitial = 0.5 atm

pinitial = 1.0 atm

pinitial = 2.0 atm

Fig. 5. Effective Lewis number and flame thickness of the H2/CO/air/SEGR flamescorresponding to the snapshots of Fig. 4.

O. Askari et al. / Applied Energy 179 (2016) 451–462 455

critical Peclet number, a non-dimensional parameter, is defined asthe ratio of radius to flame thickness ðPecr ¼ rcr=df Þ at the onset ofcell formation or flame instability [25]. The onset of cell formationis defined as when the small cells appear on flame surface. Thisphenomenon happens simultaneously all over the spherical flamesurface and can be captured precisely by our high speed CMOScamera. Flame thickness is defined by df ¼ ðTad � TuÞ=ðdT=dxÞmax,where Tad is the adiabatic flame temperature, Tu the unburned

gas temperature and ðdT=dxÞmax the maximum rate of temperaturegradient [35]. As shown in Fig. 7, as the equivalence ratio increases,the critical Peclet number first decreases for lean mixtures, reachesa minimum, and then increases for rich mixtures. This minimumcritical Peclet number occurs at equivalence ratio of 2.0 fora ¼ 5% and a ¼ 10% and occurs at equivalence ratio of 1.0 fora ¼ 25%. The area under each curve indicates the stable regionwhich diminishes with increasing hydrogen concentration, mean-ing that the flame becomes cellular at smaller radii. In the presentstudy laminar burning speeds are only reported for smooth andlaminar spherical flames.

Upon ignition, as the flame propagates the pressure of thechamber increases until the pressure reaches the critical pressurefor cell formation. Tables 1 and 2 show the critical pressures andtheir corresponding temperatures of H2/CO/air/SEGR mixtures atdifferent initial pressures and a wide range of equivalence ratiosfor three hydrogen concentrations of 5%, 10% and 25%, initial tem-perature of 450 K and SEGR concentration of 5% and 10%. Thesevalues serve as indicators when the flame has visibly reached theonset of cellularity. An important issue in these experiments is thatcells are formed at large radii; for all of reported cases rcr=R > 0:5where rcr is the flame radius where cellularity occurs. It can be seenin Tables 1 and 2 that cell formation strongly depends on theequivalence ratio and hydrogen concentration. In the case thatwe have all smooth flames, the upper limit of pressures and tem-peratures are the pressures and temperatures when flames hitthe chamber wall and are reported in Tables 1 and 2 in shadedbackground started with letter S.

= . = . = .

= %

= %

= %

Fig. 6. Snapshots of the stoichiometric H2/CO/air/SEGR flames for various hydrogen concentrations at initial pressure of 2 atm, SEGR concentration of 5% and initialtemperature of 298 K.

0 0.5 1 1.5 2 2.5 3 3.5

0

500

1,000

1,500

2,000

Crit

ical

Pec

let n

umbe

r

Equivalence ratio

α = 5%α = 10%α = 25%

Unstable

Stable

Fig. 7. Critical Peclet number of the H2/CO/air/SEGR flames for various hydrogenconcentrations at initial pressure of 2 atm, SEGR concentration of 5% and initialtemperature of 298 K.

456 O. Askari et al. / Applied Energy 179 (2016) 451–462

4. Burning speed measurements

4.1. Burning model

There are several methods for measuring the laminar burningspeed [36,37,31,38–40]. The numerical model used in this workto calculate the laminar burning speed from the pressure rise datais based on a newly developed differential based multi-shell modelby Askari et al. [25]. Using this model, burning speeds can be cal-culated over a wide range of temperatures and pressures withoutthe need for extrapolation methods. Pressure rise must be accu-rately and correctly measured, and flame radii must be largeenough to minimize the stretch effect on the laminar burningspeed. In this model governing equations of unknown variables

have been defined by a set of nonlinear ordinary differential equa-tions that can be solved by using the CVODE solver from sundialspackage [41]. The governing equations are derived by applyingthe differential form of mass and energy conservation equationsover all shells as well as the whole chamber.

The model divides the combustible mixture into an unburnedgas section and a burned gas section and considers the flame frontas a jump discontinuity. The burned gas core is divided into a num-ber of shells, each shell with a uniform temperature in local chem-ical equilibrium. Surrounding the outermost burned gas shell is alayer of unburned gas called the preheat zone that has a non-uniform temperature, immediately followed by the outermostunburned gas shell. Pressure is assumed to have no spatial gradientin the chamber at any particular instant in time. The gas mixture isassumed to behave like an ideal gas and to be in local thermody-namic equilibrium (LTE). Chemical equilibrium composition hasbeen evaluated using CANTERA code combined with the latestthermochemical properties of NASA correlations [42]. In order toencompass the burning of lean and rich mixtures, twenty majorspecies (H2, N2, O2, CO, H, O, OH, HCO, HO2, H2O, H2O2, CO2, NO,NO2, CH2, CH3, CH4, CH2O, CH3OH and C2H2) have been consideredas products of combustion. These twenty species are the ones withconcentrations higher than 10�12 based on local thermodynamicequilibrium assumption, other species which are less than 10�12

neglected in this study. All the energy losses to the electrodes,chamber walls, radiation from the burned gas, and energy transferbetween neighboring shells have been added to the current model.The effect of energy transfer in the chamber wall and spark elec-trodes as well as preheat zone is modeled by the thermal boundarylayer and displacement thickness concept [43].

This model simulates a one-dimensional spherical flame inradial direction. The chamber is filled with a fuel/air/diluent mix-ture and then centerally ignited using two extended spark elec-trodes. After ignition, an isotropic flame is formed and propagatesoutwardly in the radial direction. The following assumptions aremade in the analysis of the combustion inside the chamber:

Table 1Critical pressures and temperatures of H2/CO/air/SEGR mixtures at SEGR concentration of 5%, initialtemperature of 298 K, different initial pressures, wide range of equivalence ratios and three hydrogenconcentration of 5%, 10% and 25%.

Table 2Critical pressures and temperatures of H2/CO/air/ SEGR mixtures at SEGR concentration of 10%, initialtemperature of 298 K, different initial pressures, wide range of equivalence ratios and three hydrogenconcentration of 5%, 10% and 25%.

O. Askari et al. / Applied Energy 179 (2016) 451–462 457

1. The reactants are initially quiescent and have a uniform compo-sition, pressure and temperature.

2. The pressure is presumed to have a uniform spatial distributionwithin the chamber at each time step.

3. The gases are assumed to behave like an ideal gas and to be inlocal thermodynamic equilibrium (LTE).

The energy conservation equations for unburned, burned, andcurrently burning regions are respectively:

_Uu ¼ � _mbhu þ _Qu � _Wu ð1Þ

_Uj ¼ _Qj � _Wj; j ¼ b1 � bn�1 ð2Þ

_Ubn ¼ _mbhu þ _Qbn � _Wbn ð3Þwhere U is the internal energy, _mb the mass burning rate, Q theenergy transfer and W the work. In these equations the subscriptsof j refers the already burned shells, u and b denote the unburnedand burned gas conditions respectively, n and bn denote the totalnumber of shells in burned section and currently burning shell,respectively. The dot sign on top of the parameters refers to the

complete derivative with respect to time. Using concept of displace-ment thickness [43] and thermodynamic relations [44] Eqs. (1)–(3)expand into the following format:

_T1u ¼

A1u_pþ q1

u h1u1�xb

Vdisu _xb þ q1u h1

u_Vdisu þ _Qu

B1u

ð4Þ

_T1j ¼ A1

j_pþ q1

j h1j_Vdisj þ _Qj

B1j

; j ¼ b1 � bn�1 ð5Þ

_T1bn ¼

A1bn_pþ ðmðh1

u � h1bn Þ �

q1u h1u1�xb

Vdisu Þ _xb þ q1bnh1bn_Vdisbn

þ _Qbn

B1bn

ð6Þ

_xb ¼ �ðC1u þP

kC1k Þ _pþ ðD1

u þPkD

1k Þ þ ðE1

u þPkE

1k Þ

F;

k ¼ u; b1; � � � ; bn�1; bn ð7Þwhere q is the density, h the enthalpy, xb the burned gas mass frac-tion, Vdis the displacement thickness, m the total mass, p the exper-imental pressure data and the superscript 1 denotes the region far

30 40 50 60 70 80 90 1000

20

40

60

80

100

120

140

160

180

Lam

inar

bur

ning

spe

ed (c

m/s

)

Stretch rate (1/s)

φ=1, T=460K, p=4.5atm, β=0%φ=2, T=382K, p=1.2atm, β=10%φ=3, T=410K, p=3.1atm, β=0%

Fig. 8. Laminar burning speeds versus stretch rates for three different states withvarious equivalence ratios, unburned gas conditions and SEGR concentrations athydrogen concentration of 5%.

458 O. Askari et al. / Applied Energy 179 (2016) 451–462

from the thermal boundary layer with uniform temperature distri-bution. The parameters A1

k through E1k ðk ¼ u; b1; � � � ; bn�1; bnÞ are

defined in Askari et al. [25] in terms of thermodynamic properties[45]. Eqs. (4)–(7) form a set of nonlinear ordinary differential equa-tions which contain nþ 3 unknowns: pðtÞ; xbðtÞ; T1

u andT1biði ¼ 1 to nÞ. Given experimental pressure as a function of time,

they can be solved numerically using CVODE method to find burnedmass fraction and temperature distribution. Finally, the laminarburning speed is calculated as:

Su ¼ m _xbq1

u Afð8Þ

where Af is the flame front area. The initial mixture composition isdefined as:

/ðaH2 þ ð1� aÞCOÞ þ 5021

ð0:21O2 þ 0:79N2Þ

þ bð/þ 5021Þ

1� bð0:14CO2 þ 0:86N2Þ ð9Þ

where a is the hydrogen concentration of the fuel mixture, / theequivalence ratio and b the SEGR concentration. Combination ofthe systematic (Bias) and random (Precision) uncertainty [46] inconjunction with the root-sum-square method give us the overalluncertainty for calculated laminar burning speed data [27]. Theoverall uncertainty varys between ±1% and ±3% with the averageof ±1.51%. This range of uncertainty is only valid for conditionswhich are covered in our experiments and numerical analysis. Forother conditions which are estimated by power law correlationssuch as atmospheric conditions (except sub-atmospheric pressuretests) the mean uncertainty is ±2.76%.

A steady one-dimensional laminar premixed free flame codefrom the CANTERA package [47] in conjunction with detailedkinetics mechanisms has been also used to solve the conservationequations of mass, energy and species in order to calculate thelaminar burning speed. The multi-component diffusion modelhas been used for evaluation of transport properties. In this paper,two detailed chemical kinetics mechanisms for syngas combustion[48,49] have been selected and compared to the experimental lam-inar burning speed data.

4.2. Stretch effect investigation

Stretched laminar burning speed is a function of the geometryof the flame and is not a fundamental thermo-physical property.Stretched laminar burning speed calculated from spherical flamesare usually corrected by means of extrapolation to zero stretch.Various linear and nonlinear extrapolations have been developedto obtain the zero stretch laminar burning speed[14,15,50,51,52,23]. However, a recent paper by Wu et al. [53]has reviewed different methods of extrapolation and has shownthat there is great uncertainty in these methods. Flame stretch inspherically expanding flames is defined as:

j ¼ 1Af

dAf

dt¼ 2

rf

drfdt

ð10Þ

where j is the stretch rate, Af the flame front area, t the time, and rfthe flame radius. It is seen from Eq. (10) that as the radius of flameincreases the stretch rate decreases. Therefore, it is advantageous toconsider data where flame radius is large, minimizing stretch andthus circumventing the challenges associated with extrapolationto zero stretch.

The procedure used to investigate the stretch effects is brieflyoutlined below. Laminar burning speeds have been measured atdifferent stretch rates in order to study the effect of stretch onH2/CO/air/SEGR mixtures. Given an isentrope that begins at an ini-tial pressure and temperature, different tests are selected from that

isentrope such that the only difference in the initial conditions ofthe subsequent tests are the initial pressure and temperature.The initial pressures and temperatures selected must be con-strained to lie on the isentrope of the first test. Finally, a thermody-namic state (same pressure, temperature and equivalence ratio)that is shared by all tests is chosen and the laminar burning speedsare compared at that chosen state for different stretch rates. Moreinformation about this method can be found in previous publica-tions [1,10]. Fig. 8 shows the variation of laminar burning speedversus stretch rate for three different states with various equiva-lence ratios, unburned gas conditions and SEGR percentages athydrogen concentration of 5%. As it can be seen the change of lam-inar burning speeds with respect to stretch rate for flame radiigreater than 4 cm and stretch rates lower than 90 s�1 is negligiblewhich is in agreement with Chen et al. [39].

4.3. Laminar burning speed

To reduce the effects of stretch and spark energy discharge alllaminar burning speed data are reported in regions whererf > 4 cm and stretch rate of less than 90 s�1. The calculated lami-nar burning speeds have been used to develop a correlation interms of equivalence ratio, temperature, pressure and SEGR con-centration as shown in Eq. (11):

Su ¼ Su0ð1þ að/� 1Þ þ bð/� 1Þ2Þ TT0

� �a pp0

� �h

1� bð Þc ð11Þ

where Su0 is the laminar burning speed at reference pointð/ ¼ 1; T0 ¼ 298 K;p0 ¼ 1atmÞ, / the equivalence ratio, T the tem-perature in K, p the mixture pressure in atm and b the SEGR volu-metric fraction. To have a good fitting the temperature, pressureand SEGR fraction exponents are considered as a linear functionof equivalence ratio as shown in Eqs. (12)–(14):

a ¼ a1 þ a2ð/� 1Þ ð12Þ

h ¼ h1 þ h2ð/� 1Þ ð13Þ

c ¼ c1 þ c2ð/� 1Þ ð14ÞCoefficients, Su0, a, b, c, a1, a2, h1, h2, c1 and c2, have been found

using fminunc function in MATLAB, an unconstrained minimizationmethod, for three different hydrogen concentrations as listed inTable 3. This correlation is only valid for smooth laminar flamesin the equivalence ratio range of 0:6 < / < 3 and SEGR fraction

Fig. 10. Laminar burning speed of stoichiometric H2/CO/air/SEGR mixture alongisentropes at different initial pressures for initial temperature of 298 K, hydrogenconcentration of 5% and SEGR concentration of 10%.

Table 3Power law fitting coefficients.

Su0 a b a1 a2 h1 h2 c1 c2

a ¼ 5% 33.351 1.313 �0.410 1.884 �0.008 �0.136 �0.003 3.130 0.818a ¼ 10% 47.433 1.337 �0.444 1.843 0.004 �0.173 0.019 3.579 0.429a ¼ 25% 76.209 1.351 �0.510 1.842 0.008 �0.175 0.035 3.631 0.347

O. Askari et al. / Applied Energy 179 (2016) 451–462 459

range of 0 < b < 0:1. The range of pressure and temperature are afunction of equivalence ratio and hydrogen concentration, whichcan be determined using Tables 1 and 2 for SEGR concentrationof 5% and 10%, respectively.

Fig. 9 shows the laminar burning speeds of syngas/air/SEGRmixtures at various equivalence ratios and SEGR concentration of10% versus temperature along four isentropes. As the equivalenceratio increases from / ¼ 0:6 to / ¼ 2:0 laminar burning speedincreases. By increasing the equivalence ratio beyond / ¼ 2:0, thelaminar burning speed decreases. As it can be seen the pressureexponent in Eq. (11) for whole range of equivalence ratio is alwaysnegative which means the initial pressure has an inverse effect onlaminar burning speed. This behavior is shown in Fig. 10 that byincreasing the initial pressure, the laminar burning speeddecreases. Also, Fig. 10 shows that the shrinking temperature rangewhere laminar burning speed can be calculated is becoming morenarrow as initial pressure increases since flames become cellular.Hydrogen addition has a significant direct impact on laminar burn-ing speed. As shown in Fig. 11 the laminar burning speed increaseswith increasing hydrogen concentration in the fuel mixture. SEGRaddition to the syngas/air mixture has a very interesting effect onlaminar burning speed. SEGR acts as an energy sink, causing theflame temperature to decrease and consequently lower the laminarburning speed as shown in Fig. 12. More explanation using sensi-tivity analysis will come later to show how SEGR addition lowersthe laminar burning speed. In all these figures, the solid line repre-sents fitted laminar burning speed data using the power law corre-lation in Eq. (11).

Since there is no laminar burning speed data in the literature forthe exact SEGR mixture composition that has been considered inthis paper (14% CO2 and 86% N2), laminar burning speeds from thisstudy have been compared with experimental data in the literaturefor no SEGR addition at standard atmospheric temperature andpressure for hydrogen concentration of 5%, as shown in Fig. 13.The measured laminar burning speed agrees well with availableexperimental data in literature. The results were within the

Fig. 9. Laminar burning speed of H2/CO/air/SEGR mixture along isentropes atdifferent equivalence ratios for initial pressure of 0.5 atm, initial temperature of298 K, hydrogen concentration of 25% and SEGR concentration of 10%.

Fig. 11. Laminar burning speed of H2/CO/air/SEGR mixture along isentropes atdifferent hydrogen concentrations for atmospheric initial pressure, initial temper-ature of 298 K, equivalence ratio of 3.0 and SEGR concentration of 5%.

expected experimental error under the test conditions where com-parisons were possible with other studies. As described earlier inthis paper the average uncertainty of ±2.76% is associated withthe conditions which are approximated from the power law fittingmethod such as atmospheric conditions (except for sub-atmospheric initial pressure). Fig. 14 shows a comparison of lami-nar burning speed calculated in this study by power law correla-tion versus two detailed kinetics mechanisms, Davis [48] and Li[49] mechanisms, for H2/CO/air/SEGR mixture at atmospheric con-ditions, hydrogen concentration of 25% and three different SEGRconcentrations of 0%, 5% and 10%. From the selected kinetics mech-anisms, the Davis mechanism [48] is in fairly good agreement withthe obtained experimental data.

Fig. 12. Laminar burning speed of H2/CO/air/SEGR mixture along isentropes atdifferent SEGR concentrations for initial pressure of 0.5 atm, initial temperature of298 K, equivalence ratio of 0.6 and hydrogen concentration of 10%.

0 0.5 1 1.5 2 2.5 3 3.50

10

20

30

40

50

60

70

80

Equivalence ratio

Lam

inar

bur

ning

spe

ed (c

m/s

)

Present studyMClean et al [50]Hassan et al [14]Natarajan et al [16,17]Sun et al [15]Bouvet et al [54,55]Singh et al [56]

Fig. 13. Comparison of present experimental laminar burning speed data versuspublished experimental data [14–17,50,54–56] for H2/CO/air at atmospherictemperature and pressure at hydrogen concentration of 5% and SEGR concentrationof 0%.

Fig. 14. Comparison of calculated laminar burning speeds versus two kineticssimulations (Davis [48] and Li [49] mechanisms) for H2/CO/air/SEGR mixture atatmospheric pressure and temperature, hydrogen concentration of 25% and threedifferent SEGR concentrations.

Fig. 15. Normalized sensitivity coefficients of H2/CO/air/SEGR mixture at differentSEGR concentrations, initial pressure of 0.5 atm, initial temperature of 298 K,equivalence ratio of 0.6 and hydrogen concentration of 25%.

460 O. Askari et al. / Applied Energy 179 (2016) 451–462

To show the effect of SEGR addition on the normalized sensitiv-ity coefficients, computations were performed for three differentvalues of SEGR concentrations of 0%, 5% and 10%. Davis et al. mech-anism [48], was employed for sensitivity analysis due to its goodaccuracy as shown in Fig. 14. The sensitivity of the predicted lam-inar burning speed with respect to each reaction rate constant iscalculated using CANTERA [47] code. Fig. 15 gives the normalizedsensitivity coefficient of H2/CO/air/SEGR mixture for the most sen-sitive reactions in the mechanism of Davis et al. [48] at three differ-ent SEGR concentrations of 0%, 5% and 10%.

Chain propagation and branching reactions have positive sensi-tivity coefficients while recombination and termination reactionshave negative sensitivity coefficients. As shown in Fig. 15 the mostdominant chain branching reactions are R1: H + O2 () O + OH,R2: O + H2 () H + OH and R16: HO2 + H () OH + OH, throughincreasing the concentrations of highly reactive radical species, Hand OH which promote the combustion. Reaction R28: CO +OH () CO2 + H is identified as the most significant chain propa-gation reaction for syngas mixture due to high concentration of CO

in the reactants. The most dominant chain termination reaction isthe third body Reaction R12: H + O2(+M) () HO2(+M) wherereactive radical, H, is converted into stable species, H2O. As shownin Fig. 15, the sensitivity coefficient of the chain termination reac-tion R12 and the chain branching and propagation reactions R1, R2,R16 and R28 are increased with the increase of SEGR concentra-tions in the mixture. It means that these five aforementioned reac-tions are playing significant roles as SEGR concentration increases.

As mentioned earlier, with the increase of CO2 concentration,through addition of SEGR, flame temperature and subsequentlythe H radical production via chain branching reaction R2 decreases.On one hand, there is a competition between chain branching reac-tion R1 and chain termination reaction R12 on the consumption ofH radical. On the other hand, CO2 can act as a third body in thechain termination reaction R12 and consumes the H radical pro-duced from chain branching reaction R2. By increasing the CO2

fraction, chain termination reaction R12 wins the competitionand suppresses the chain branching reaction R1. Through this com-petitive process, the concentration of H and OH radicals in thereaction zone are reduced drastically and therefore the other sig-nificant chain branching and propagation reactions, R16 and R28,are suppressed. As a result, the whole reaction process decreasesfollowed by an attendant decrease in laminar burning speed.

O. Askari et al. / Applied Energy 179 (2016) 451–462 461

5. Conclusions

Laminar burning speed of syngas that has been diluted withsynthetic EGR has been presented in this paper using constant vol-ume spherical and cylindrical chambers with a novel burningmodel based on the pressure rise due to combustion process. Ananalysis shows that the stretch effects on laminar burning speedare minimized for flames with radii larger than 4 cm and stretchrate of less than 90 s�1. The stability of propagating flames hasbeen analyzed using several key parameters, including the Lewisnumber, flame thickness, and Peclet number. Increasing SEGR con-centration has a stabilizing effect on flames due to the combinationeffects of hydrodynamic and thermo-diffusive instabilities and alsotends to decrease the laminar burning speed. Carbon dioxidebehaves as an inhibitor due to the competition of the reaction ofR1: H + O2 () O + OH and R12: H + O2(+M) () HO2(+M) on Hradical. Laminar burning speeds of H2/CO/air/SEGR mixtureincrease with the increase of hydrogen fraction while also promot-ing the cellular flame structure. A power law fit correlation hasbeen developed for hydrogen concentrations of 5%, 10% and 25%over equivalence ratios of 0.6 to 3.0 and SEGR concentrations of0–10% for smooth flames only. Flame critical pressures and theircorresponding temperatures can be determined using Tables 1and 2 for SEGR percentages of 5% and 10%. Laminar burning speeddata in this work agrees well with the available data in theliterature.

Acknowledgments

This paper was made possible by NPRP award [NPRP 7-1449-2-523] from Qatar National Research Fund (a member of the QatarFoundation). This statement is solely the responsibility of theauthors.

References

[1] Moghaddas A, Eisazadeh-Far K, Metghalchi H. Laminar burning speedmeasurement of premixed n-decane/air mixtures using sphericallyexpanding flames at high temperatures and pressures. Combust Flame2012;159(4):1437–43.

[2] Eisazadeh-Far K, Parsinejad F, Metghalchi H. Flame structure and laminarburning speeds of JP-8/air premixed mixtures at high temperatures andpressures. Fuel 2010;89(5):1041–9.

[3] Askari O, Janbozorgi M, Greig R, Moghaddas A, Metghalchi H. Developingalternative approaches to predicting the laminar burning speed of refrigerantsusing the minimum ignition energy. Sci Technol Built Environ 2015;21(2):220–7.

[4] Metghalchi M, Keck JC. Burning velocities of mixtures of air with methanol,isooctane, and indolene at high pressure and temperature. Combust Flame1982;48:191–210.

[5] Elia M, Ulinski M, Metghalchi M. Laminar burning velocity of methane–air–diluent mixtures. J Eng Gas Turbines Power 2001;123(1):190.

[6] Rahim F, Elia M, Ulinski M, Metghalchi M. Burning velocity measurements ofmethane-oxygen-argon mixtures and an application to extend methane-airburning velocity measurements. Int J Engine Res 2005;3(2):81–92.

[7] Parsinejad F, Matio M, Metghalchi M. A mathematical model for schlieren andshadowgraph images of transient expanding spherical thin flames. J Eng GasTurbines Power 2004;126(2):241.

[8] Parsinejad F, Keck JC, Metghalchi H. On the location of flame edge inShadowgraph pictures of spherical flames: a theoretical and experimentalstudy. Exp Fluids 2007;43(6):887–94.

[9] Eisazadeh-Far K, Parsinejad F, Metghalchi H, Keck JC. On flame kernelformation and propagation in premixed gases. Combust Flame 2010;157(12):2211–21.

[10] Eisazadeh-Far K, Moghaddas A, Al-Mulki J, Metghalchi H. Laminar burningspeeds of ethanol/air/diluent mixtures. Proc Combust Inst 2011;33(1):1021–7.

[11] Moghaddas A, Bennett C, Eisazadeh-Far K, Metghalchi H. Measurement ofLaminar burning speeds and determination of onset of auto-ignition of Jet-A/Air and Jet propellant-8/air mixtures in a constant volume spherical chamber. JEnergy Res Technol 2012;134(2):022205.

[12] Yu G, Askari O, Hadi F, Wang Z, Metghalchi H, Kannaiyan K, et al. Theoreticalprediction of Laminar burning speed and ignition delay time of gas-to-liquidfuel. J Energy Resour Technol, ASME 2016. http://dx.doi.org/10.1115/1.4033984.

[13] Askari O, Hannani SK, Ebrahimi R. Improvement and experimental validationof a multi-zone model for combustion and NO emissions in CNG fueled sparkignition engine. J Mech Sci Technol 2012;26(4):1205–12.

[14] Hassan MI, Aung KT, Faeth GM. Properties of Laminar premixed CO/H2/airflames at various pressures. J Propul Power 1997;13(2):239–45.

[15] Sun H, Yang SII, Jomaas G, Law CKK. High-pressure laminar flame speeds andkinetic modeling of carbon monoxide/hydrogen combustion. Proc CombustInst 2007;31:439–46.

[16] Natarajan J, Lieuwen T, Seitzman J. Laminar flame speeds of H2/CO mixtures:effect of CO2 dilution, preheat temperature, and pressure. Combust Flame2007;151(1–2):104–19.

[17] Natarajan J, Kochar Y, Lieuwen T, Seitzman J. Pressure and preheat dependenceof laminar flame speeds of H2/CO/CO2/O2/He mixtures. Proc Combust Inst, vol.32. p. 1261–8. I(1).

[18] Prathap C, Ray A, Ravi MR. Investigation of nitrogen dilution effects on thelaminar burning velocity and flame stability of syngas fuel at atmosphericcondition. Combust Flame 2008;155(1–2):145–60.

[19] Prathap C, Ray A, Ravi MR. Effects of dilution with carbon dioxide on thelaminar burning velocity and flame stability of H2-CO mixtures at atmosphericcondition. Combust Flame 2012;159(2):482–92.

[20] Vu TM, Park J, Kwon OB, Bae DS, Yun JH, Keel SI. Effects of diluents on cellularinstabilities in outwardly propagating spherical syngas-air premixed flames.Int J Hydrogen Energy 2010;35(8):3868–80.

[21] Burbano HJ, Pareja J, Amell AAa. Laminar burning velocities and flame stabilityanalysis of syngas mixtures at sub-atmospheric pressures. Int J HydrogenEnergy 2011;36(4):3243–52.

[22] Lapalme D, Seers P. Influence of CO2, CH4, and initial temperature on H2/COlaminar flame speed. Int J Hydrogen Energy 2014;39(7):3477–86.

[23] Han M, Ai Y, Chen Z, Kong W. Laminar flame speeds of H2/CO with CO2 dilutionat normal and elevated pressures and temperatures. Fuel 2015;148:32–8.

[24] Wang ZH, Weng WB, He Y, Li ZS, Cen KF. Effect of H2/CO ratio and N2/CO2

dilution rate on laminar burning velocity of syngas investigated by directmeasurement and simulation. Fuel 2015;141(x):285–92.

[25] Askari O, Moghaddas A, Alholm A, Vein K, Alhazmi B, Metghalchi H. Laminarburning speed measurement and flame instability study of H2/CO/air mixturesat high temperatures and pressures using a novel multi- shell model. CombustFlames 2016;168:20–31.

[26] Askari O, Metghalchi H, Kazemzadeh Hannani S, Moghaddas A, Ebrahimi R,Hemmati H. Fundamental study of spray and partially premixed combustionof methane/air mixture. J Energy Res Technol 2012;135(2):021001.

[27] Askari O, Metghalchi H, Kazemzadeh Hannani S, Hemmati H, Ebrahimi R. Leanpartially premixed combustion investigation of methane direct-injectionunder different characteristic parameters. J Energy Res Technol 2014;136(2):022202.

[28] Bradley D, Gaskell PH, Gu XJ. Burning velocities, markstein lengths, and flamequenching for spherical methane-air flames: a computational study. CombustFlame 1996;104:176–98.

[29] Trochim WMK, Donnelly JP. Research methods knowledge base. Cincinnati,OH: Atomic Dog Publishing; 2000.

[30] Parsinejad F, Arcari C, Metghalchi H. Flame structure and burning speed of Jp-10 air mixtures. Combust Sci Technol 2006;178(5):975–1000.

[31] Rokni E, Moghaddas A, Askari O, Metghalchi H. Measurement of Laminarburning speeds and investigation of flame stability of acetylene (C2H2)/airmixtures. J Energy Res Technol 2014;137(1):012204.

[32] Hajialimohammadi A, Ahmadisoleymani S, Abdullah A, Askari O, Rezai F.Design and manufacturing of a constant volume test combustion chamber forJet and flame visualization of CNG direct injection. Appl Mech Mater2012;217–219:2539–45.

[33] Askari O, Elia M, Ferrari M, Metghalchi H. Auto-ignition characteristics study ofgas-to-liquid fuel at high pressures and low temperatures. J Energy ResourTechnol, ASME 2016. http://dx.doi.org/10.1115/1.4033983.

[34] Eisazadeh-Far K, Moghaddas A, Metghalchi H, Keck JC. The effect of diluent onflame structure and laminar burning speeds of JP-8/oxidizer/diluent premixedflames. Fuel 2011;90(4):1476–86.

[35] Law CK, Kwon OC. Effects of hydrocarbon substitution on atmospherichydrogen–air flame propagation. Int J Hydrogen Energy 2004;29(8):867–79.

[36] Egolfopoulos FNN, Cho P, Law CKK. Laminar flame speeds of methane-airmixtures under reduced and elevated pressures. Combust Flame 1989;76(3–4):375–91.

[37] Metghalchi M, Keck JC. Laminar burning velocity of propane-air mixtures athigh temperature and pressure. Combust Flame 1980;38:143–54.

[38] Lewis B, von Elbe G. Determination of the speed of flames and the temperaturedistribution in a spherical bomb from time-pressure explosion records. J ChemPhys 1934;2:283.

[39] Chen Z, Burke MP, Ju Y. Effects of compression and stretch on thedetermination of laminar flame speeds using propagating spherical flames.Combust Theor Model 2009;13(2):343–64.

[40] Davis SG, Law CK, Wang H. An experimental and kinetic modeling study ofpropyne oxidation. Symp (Int) Combust 1998;27(1):305–12.

[41] https://computation.llnl.gov/casc/sundials/main.html.[42] Zehe, M.J., Gordon, S., McBride, B.J., 2002. CAP: A Computer Code for

Generating Tabular Thermodynamic Functions from NASA Lewis Coefficients.NASA/TP-2001-210959/REV1 CAP.

[43] Keck JC. Thermal boundary layer in a gas subject to a time dependent pressure.Lett Heat Mass Trans 1981;8(4):313–9.

[44] Hsieh JS. Principles of thermodynamics. McGraw-Hill; 1975.

462 O. Askari et al. / Applied Energy 179 (2016) 451–462

[45] Askari O, Beretta GP, Eisazadeh-Far K, Metghalchi H. On the thermodynamicproperties of thermal plasma in the flame kernel of hydrocarbon/air premixedgases. Euro Phys J D (EPJ D) 2016(4). http://dx.doi.org/10.1140/epjd/e2016-70195-4.

[46] Coleman HW, Steele JWG. Experimentation and uncertainty analysis forengineers. John Wiley & Sons; 1989.

[47] Goodwin, D., Moffat, H., Speth, R., 2015. Cantera: An object-oriented softwaretoolkit for chemical kinetics, thermodynamics, and transport processes[Online]. Available: <http://www.cantera.org>.

[48] Davis SG, Joshi AV, Wang H, Egolfopoulos F, Troe J, Li J, Kumaran SS, Smith D,Dryer FL. An optimized kinetic model of H2/CO combustion. Proc Combust Inst,30. p. 1283–92. I.

[49] Li J, Zhao Z, Kazakov A, Chaos M, Dryer FL, Scire JamesJ, et al. A comprehensivekinetic mechanism for CO, CH2O, and CH3OH combustion. Int J Chem Kinet2007;39:109–36.

[50] MClean IC, Smith DB, Taylor SC. The use of carbon monoxide/hydrogenburning velocities to examine the rate of the CO+OH reaction. 25thSymposium on Combustion symposium on combustion, vol. 25. p. 749–57. 1.

[51] Burke MP, Chaos M, Dryer FL, Ju Y. Negative pressure dependence of massburning rates of H2/CO/O2/diluent flames at low flame temperatures. CombustFlame 2010;157:618–31.

[52] Liu CC, Shy SS, Chiu CW, Peng MW, Chung HJ. Hydrogen/carbon monoxidesyngas burning rates measurements in high-pressure quiescent and turbulentenvironment. Int J Hydrogen Energy 2011;36(14):8595–603.

[53] Wu F, Liang W, Chen Z, Ju Y, Law CK. Uncertainty in stretch extrapolation oflaminar flame speed from expanding spherical flames. Proc Combust Inst2014;35(1):663–70.

[54] Bouvet N, Chauveau C, Gökalp I, Halter F. Experimental studies of thefundamental flame speeds of syngas (H2/CO)/air mixtures. Proc Combust Inst2011;33:913–20.

[55] Bouvet N, Chauveau C, Gökalp I, Lee SY, Santoro RJ. Characterization of syngaslaminar flames using the Bunsen burner configuration. Int J Hydrogen Energy2011;36(1):992–1005.

[56] Singh D, Nishiie T, Tanvir S, Qiao L. An experimental and kinetic study ofsyngas/air combustion at elevated temperatures and the effect of wateraddition. Fuel 2012;94:448–56.