Sooting tendencies of primary reference fuels in atmospheric laminar diffusion flames burning into vitiated
air
Muhammad Kashif,Philippe Guibert,Jérôme Bonnety,Guillaume Legros ⇑
Sorbonne Universités,UPMC Univ Paris 06,CNRS,UMR 7190Institut Jean le Rond d’Alembert,F-750
05Paris,France
a r t i c l e i n f o Article history:
Received 18October 2013
Received in revid form 11December 2013Accepted 11December 2013
Available online 15January 2014Keywords:
Yield sooting index
Light extinction method
Soot volume fraction measurement Primary reference fuels
a b s t r a c t
In this study,the sooting tendencies of primary reference fuels (PRFs)are measured in term of yield soot-ing indices (YSIs)in methane diffusion flames doped with the vapors of PRFs.The prent paper repre-nts an incremental advance complementing the original methodology prescribed by M
cEnally and Pfefferle.The influence of both PRF formulation and CO 2dilution of the coflowing air on the YSIs is also assd.The diffusion flames burning in a coflowing oxidizer stream are established over the Santoro’s burner and vapor of the liquid fuel to be investigated is injected into the fuel stream.Lar extinction measurements are performed to map the two-dimensional field of soot volume fraction in the flame.For the pure liquid hydrocarbons ,n -hexane,n -heptane,isooctane,and benzene,the YSI reported in the original paper by McEnally and Pfefferle quantitatively predict the sooting propensi-ties,measured here at much higher dopant concentrations.The prent study therefore extends the con-sistency of the YSI methodology on the Santoro’s burner.For blends of n -heptane and isooctane,the sooting tendency of doped flames exhibits regular and monotonic trends and decreas with increasing n -heptane mole fraction or CO 2dilution.Interestingly,the evolution of YSI with the isooctane mole frac-tion exhibits a strong similarity for varying CO 2mole fraction.A quadratic least-squares fit is then derived,providing a phenomenological model of YSI as a function of both isooctane mole fraction in the fuel stream and CO 2mole fraction in the oxidizer.A non-negligible cross effect of PRF formulation and CO 2dilution on YSI is revealed.The method elaborated within the framework of the prent paper could be extended to surrogate fuels.This would help develop a comprehensive databa and empirical correlations that could predict the sooting propensities of different surrogate fuels,therefore their pote
n-tially mitigationed soot production through control of fuel composition and/or exhaust gas recirculation.This databa would also be uful for the validation of CFD simulations incorporating sophisticated model of soot production.
Ó2013The Combustion Institute.Published by Elvier Inc.All rights rerved.
1.Introduction
The growing concerns about increasing energy demand,energy efficiency policies,energy curity issues,hazardous effects of combustion-related emissions on human health,and global warm-ing change have led to more and more stringent regulations on ex-haust emissions from combustion devices.The recent legislation and guidelines to mitigate emissions related problems prescribe the integration of renewable fuels into the transportation fuel pool and the u of high-efficiency combustion devices to reduce CO 2emissions [1,2].The challenges can be converted into opportuni-ties through a careful lection of renewable/alternative fuel com-positions and efficient advanced combustion process.Alternative and biofuels exhibit a wide range of combustion characteristics and induce different effects on combustion pro-cess,and therefore ultimately on exhaust emissions.Hence,mix-tures compod of alternative fuels could be designed specifically to comply with a t of requirements,for instance,to reduce soot and NO x at the same time [3].
Soot particles are directly related to the reduction of combus-tion efficiency [4,5].Ideally,CO 2and H 2O are the only products of complete combustion,but in practical devices,intermediate spe-cies formed due to the pyrolysis of hydrocarbons in locally fuel-rich zones lead eventually to the production of soot particles through complex physicochemical process involving gaous species and solid particles [6].Incread risks of cardiac arrest,pul-monary dias,pregnancy complications,and asthma are now associated with the prence of particulate matter in urban air and are also linked to higher mortality rates [4,7–9].According to some estimate,soot particles may play an even more critical role in global warming than methane,which makes the cond highest
0010-2180/$-e front matter Ó2013The Combustion Institute.Published by Elvier Inc.All rights rerved.dx.doi/10.bustflame.2013.12.009
⇑Corresponding author.
E-mail address: (M.Kashif),philippe.guibert@upmc.fr (P.Guibert),jerome.bonnety@upmc.fr (J.Bonnety),guillaume.legros@dalembert.upmc.fr (G.Legros).
contribution to the greenhou effect[10].Hence,reduction of soot emissions into the atmosphere would also be a relevant policy to reduce global warming[10],complementing the regulation of CO2e
missions.
However,many gaps have been identified in the available knowledge about the production of particulate matter during com-bustion process.For this reason,the number of studies on soot for-mation/oxidation has incread for the past two decades[10–13]. Meaningful reviews have namely been published that characterize the main formation pathways for polycyclic aromatic hydrocar-bons(PAH)and ultimately soot particles in laboratory reference flames[6,10,14–17].
The new insights into the physicochemical process govern-ing soot formation and oxidation and the influence of different compounds on the process will help design combustors and/ or fuel compositions for future applications with lower environ-mental impact.Emission of soot particles can indeed be reduced through either changes in the design of combustor or modification of fuel composition.The basic advantage of fuel modification is the immediate effect on the existing devices.Fuel modification strate-gies range from doping fuels with additives to reformulating a fuel’s bulk ,blends of conventional and renewable fuels),which may promote some chemical pathways that avoid soot formation[10].The number of possible fuel additives and reformulation strategies is unlimited,and they will become more and more attractive with the increasing price of fossil fuels.Thus, there is a need to especially characterize soot production proce
ss induced by the strategies.
Under atmospheric conditions,many practical fuels,such as tho derived from fossil oil,are liquid mixtures of hundreds of hydrocarbons that have variable and coupled effects on regulated engine emissions[3,18,19]and especially soot production [10,20].This makes modeling and simulation of the sooting behav-ior of the fuels a significant computational task.As a result,many works focus on developing a molecular-level understanding of fuel combustion process and surrogate fuels.A surrogate fuel is a mixture of lected pure chemical compounds in known concen-trations and exhibits physical(density,viscosity,surface tension, vapor pressure,etc.)and chemical(ignition delay,hydrogen/car-bon ratio,cetane or octane number,laminarflame burning rate, etc.)properties similar to tho of the fuel it reprents.The study of such surrogate fuels is required for the validation of multidi-mensional computationalfluid dynamics(CFD)simulations.As an illustration,the simplest gasoline surrogates could be binary mixtures of isooctane and n-heptane,called primary reference fuels(PRFs)[21–23].
Among other characteristics,the propensity of a surrogate fuel to produce soot should match the one exhibited by the real fuel it reprents[19].The relevance of a given surrogate fuel’s sooting propensity is then assd when the soot productions by the surrogate and the real fuels in practic
al devices are componed [11,24].But the sooting propensity of a surrogate fuel also allows the lection of potential additives that may significantly modify the sooting propensity of the real fuel without any radical change in the other regulated fuel properties.The relatively simple for-mulation of the surrogate fuel then enables the identification, and therefore possibly the control of the pathways governing the obrved trends[10].Therefore,a thorough characterization of the blending effect on the sooting propensity of surrogate fuels is crucial.
The control of soot production by the addition of gaous spe-cies into fuel and/or oxidizer is also an attractive strategy.The ef-fect of CO2addition on the sooting propensity of a given fuel has especially been investigated,becau many practical devices implement exhaust gas recirculation(EGR).As discusd by Du et al.[25],the introduction of an additive such as CO2affects soot formation through the following three paths:
1.dilution effect becau of reduction in the concentrations of the
reactive species;
2.thermal effect becau of the change inflame temperature;
3.possible direct chemical effect when the additive contributes to
the chemical reactions related to soot formation and oxidation.
CO2addition then induces a significant effect on the amount of soot produced.Liu et al.[26]argued that the aforementioned three effects occur simultaneously and are intimately coupled.CO2addi-tion not only reduces the concentrations of reactive species(dilu-tion),but also lowers theflame temperature(thermal).In their numerical study of laminar jet diffusionflames[26],Liu et al.high-lighted the chemical effects of CO2on soot formation.The authors showed that the reaction CO2+H!CO+OH is primarily responsible for the chemical effects of CO2addition,as reported in earlier studies[27].This effect is especially significant for CO2 addition to the oxidizer side becau of the crucial role of hydroxyl radicals OH into the soot oxidation process[28].Thus,the effect of CO2addition on the sooting propensity of surrogate fuels also needs to be characterized.
When the sooting propensity of a surrogate fuel,is evoluated a major challenge comes from the low volatility of its components, which makes difficult the ignition and the stabilization of the stan-dard experiment conducted.To overcome this challenge,some studies have suggested that combustion of prevaporized surrogate fuels is an effectivefirst step in evaluating their sooting propensities[13,29–32].To this end,laboratoryflames burning the fuel vapor eded in a carrier gas have been extensively ud, tting up different experimental and postprocessing techniques [10,14,29,32–36].
The sooting tendencies of a large number of chemical compounds,surrogate fuels,hydrocarbon-bad com-mercial fuels,and mixture of hydrocarbons and biofuels have then been reported in terms of smoke point height and threshold soot-ing index(TSI)in laminar diffusionflames[12,33,37,38].
However,the smoke point height h globally tends to decrea with the fuel molecular weight,indicating an increasing TSI that is a linear function of1/h.Therefore,at a given spatial resolution of the detection system,the TSI nsitivity decreas with the fuel molecular weight.Further considering the uncertainty associated to any methodology,the discrimination among the TSIs exhibited by heavy fuels then becomes questionable.Moreover,in the TSI methodology,the smoke point height is inferred only from visible flame shape.For this reason,McEnally and Pfefferle[29]suggested a complementing index,called the yield sooting index(YSI).Fol-lowing this methodology,the maximum soot volume fraction f v;max is actually measured in laminar jet methane/air diffusion flame who fuel stream is doped with a small amount of vapor of the liquid fuel to be investigated.f v;max is then converted into an apparatus-independent YSI.
In the prent study,an experimental procedure that reproduces the methodology recommended by McEnally and Pfefferle isfirst introduced.The YSI of PRFs are then measured.The influence of the n-heptane and isooctane blending ratio is here assd for thefirst time.The methodology isfinally ext
ended to quantify in terms of YSI the effect of CO2addition to the coflowing air,together with the possible cross effect of PRF formulation and CO2addition.
2.Experimental tup
The schematic in Fig.1depicts the experimental tup that al-lows the methodology prescribed by McEnally and Pfefferle[29] to be reproduced.
1576M.Kashif et al./Combustion and Flame161(2014)1575–1586
2.1.Burner configuration
The diffusion flames were established over an axisymmetric co-flow burner identical to the one described by Santoro et al.[39]and ud in previous investigations [40,41].This configuration made possible the extraction of the YSIs of a wide range of fuels by McE-nally et al.[29,34].The fuel stream flows through a vertical axial brass duct,which has an 11-mm effective diameter of injection d F .The coflowing oxidizer mixture is introduced into a concentric 102-mm inner diameter brass cylinder.Further details about the burner can be found in Ref.[41].
In the following,the axis of symmetry is (O z )and its origin is located at the burner tip,defining the hei
ght above the burner (HAB).The cross-stream coordinate is r ,which is the distance from the axis of symmetry.The inner radius of the axial duct is R ¼d F =2.2.2.Flow control
The reactants came from high-purity gas cylinders (CH 4and CO 2;99.9%stated purities),reagent-grade bottles (n -heptane and isooctane;99.7%),and a compressor (air).
The coflowing oxidizer stream consists of filtered laboratory compresd air that can be carbon dioxide diluted (not shown in Fig.1).Two Bronkhorst EL-FLOW mass flow controllers enable the variations of both the air and CO 2flow rates.Both flows are then introduced into the concentric cylinder through four bent ducts that promote a final mixing.A perforated brass plate,glass beads,and finally a 1.2-mm-cell-size 50-mm-high ceramic honey-comb straighten the oxidizer stream,which conquently flows upward at the burner tip.
Stored in a specific tank (1),the liquid fuel who sooting ten-dency is to be quantified is pushed away by an inert ,argon,independenceday
and flows through a Coriolis mass flow controller (6).The fuel is then vaporized and mixed with a carrier ,methane,using a Bronkhorst controlled evaporation and mixing (CEM)system (8).The gaous mixture is carried via a heated line (9)to the inner central duct of the burner.A heating wire i
s wrapped around this duct inside the burner up to the honeycomb inlet.From the evap-orator to the burner tip,all the duct connections are thermally insulated and the walls along the fuel line are maintained at a tem-perature of 150°C to keep any of the fuels investigated from con-densing.A thermocouple inrted into the central tube at the bottom end of the burner measures the temperature of the gaous mixture to make sure that no condensation can occur.Further-more,the possible condend fuel can be flushed through a valve located at the very bottom of the burner.Within the range of experimental conditions investigated here,no fuel droplet leaked through the valve during the flush purging that was performed after every modification of the fuel composition.
2.3.Diagnostics
The light extinction measurement (LEM)technique providing a two-dimensional soot volume fraction field has been described in details in Ref.[40].It has been shown to provide both fine temporal and spatial resolution.Here,a brief overview of the optical diag-nostics is given.
The arrangement of the optical diagnostics to perform LEM is shown in Fig.1.The system consists of a 100-mW continuous wave lar (10)operating at 645nm (À5/+7nm),a t of collimated beam-formi
ng optics (12–16),collection optics (17–18),and a camera (19).A digital pul generator (22)controls the occurrence and the duration of the CMOS exposure,together with the
shutter
arrangement for soot volume fraction measurements in doped methane flames:(1)liquid fuel tank,(2)safety valve,(3)(4)15l m particle filter on the liquid fuel line,(5)pressure gauge,(6)Coriolis mass flow controller,(7)methane flow mixer (CEM),(9)heated line,(10)100-mW continuous wave lar ðk ¼645nm Þ,(11)shutter,(12)10Âbeam expander,+50mm),(15)rotating diffur plate,(16)collimating achromatic lens (d =75mm,f =+750mm),(17)decollimating achromatic asmbly comprising a neutral density filter,an 800-l m diameter pinhole,and a narrow band filter centered at 645Æscan monochrome camera,(20)lar ON/OFF control,(21)shutter controller,(22)digital pul generator,and (23)oscilloscope.
opening(21).A frame grabber records the frames captured by the camera on a computer.
中括号
The Photon Focus MV112-bit progressive scan monochrome camera is mounted with a conventional lens and equipped with a narrow bandfilter centered at645nm(Æ2nm)and with a band-width at one-half the transmissivity maximum of20nm.With this optical arrangement,the matrix of1312Â1082pixels provides a spatial resolution of50l m for the LEM projected data over the 70-mm-diameter area investigated.For the current study,the frames were recorded at a frame rate of30Hz and an exposure
time of15ms,as the experimental conditions are such that the dif-fusionflames established are very stable.Even at theflame tip,the fluctuations of the raw extinction measurements with time are weak.Furthermore,under over the conditions investigated,the ttings enabled the peak raw extinction coefficients to be spanned from0.4to0.9,thus both allowing a decent discrimination among the raw measurements and standing away from the lower and upper ranges where nonlinear effects might be revealed.
As a line-of-sight technique,LEM needs to be combined with a subquent deconvolution to infer the local extinction coefficient distribution.This postprocessing is described hereafter.
3.Quantifying sooting propensities
庇护七世
3.1.Methodology
According to the method suggested by McEnally and Pfefferle [29],the apparatus-independent YSI of a fuel is calculated from the maximum soot volume fraction f v;max¼max f f vðr;z measÞj r2½0;R g measured in the fuel doped methaneflame at afixed height z meas above the burner using the equation:
Y SI¼C f
v;max
þD:ð1ÞHere,C and D are apparatus-specific parameters that are deter-mined by the arbitrary YSI values attributed to two reference fuels. In their late paper providing an exhaustive t of YSIs[34],McEnally and Pfefferle defined the YSI scale with n-hexane and benzene as reference fuels.The were given YSI values of0and100,respec-tively.z meas was lected as the height where the highest signal pro-vided by the local lar-induced incandescence technique was recorded in the benzeneflame.
Within the frame of the prent study,this method has been conducted with different experimentalflow conditions and dopant concentrations.As the original method suggested by McEnally and Pfefferle aims to quantify the sooting propensity of a given fuel in any combustion device,it is worth asssingfirst the consistency of the method conducted in the identical tup but with otherflow conditions.
3.2.Experimental procedure
For the prent study,the following procedure was conducted to establish very steady laminarflames.The carrier gasflow isfirst t through the CEM and a lightly sooting diffusionflame of
pure methane gas is established over the burner.Once the temperature of the CEM heater and the heated lines is stabilized well above the condensation temperature of the liquid fuel vapor,the liquid fuel flow is t and vaporized.Its vapor is injected into the carrier gas.Theflame then becomes brighter as soot production is clearly enhanced.A10-s-long quence of LEM(e Section3.3)is re-corded every minute,showing that any raw extinctionfield exhib-its afluctuation of its peak lower than5%over10min after the liquidflow was established.After the last recording,the liquidflow rate is t to zero.Theflame then progressively turns blue,showing that for any condition the fuel stream is compod of pure methane 1min after the liquidflow rate is t to zero.The methaneflow rate is also t to zero.Once theflame is extinguished,the liquid fuel tank isflushed and the whole line from the tank to the burner tip is purged with an inert ,argon,for5min to remove all the liquid and vapors from the system.The tank is then refilled with the liquid fuel to be investigated.
The experimental parameters that were kept constant in the prent study are reported in Table1,together with tho t by McEnally and Pfefferle.As prescribed by the authors,the values 0and100were assigned to the YSIs of n-hexane and benzene, respectively.McEnally and Pfefferle mainly ud a diffusionflame where the carrier gas is a mixture of CH4and N2,which is doped with vapor of the liquid fuel to be investigated.In the prent study,flames of pure methane doped with va
airdroppor were established. This composition of the carrier gas was chon to both enhance soot production,therefore producingfine levels of lar extinction, and provideflames who heights do not exceed the lar beam diameter.The temperature along the fuel ,150°C,is higher than that t by McEnally and Pfefferle due to the higher partial pressure of liquid fuel vapor ud in the prent study.The coflow-ing airflow ,60,000cm3/min,is also higher here,as this le-vel was required to produce stableflames within the whole range of conditions investigated.However,some experiments conducted with a temperature along the fuel line of180°C,on one hand,and with a coflowing airflow rate of80,000cm3/min,on the other hand,lead to YSI modifications lower than0.5units.Thus,the influence of the parameters is considered negligible as compared to the standard deviation among the YSIs provided by the different methods described below.
The parameters varied in the prent study are as follows:
The PRF composition,identified by the isooctane mole fraction X i C
iloveyou陶喆
8
H18
in the mixture of n-heptane and isooctane.This parame-ter is adjusted when different quantities of pure n-heptane and isooctane are mixed before the tank is refilled.
The PRF vapor mole fraction X vap;i(i=1,2).This is controlled by the liquidflow rate through the Coriolis massflow controller.
The massflow rate required to t a given X vap;i is calculated withÆ1%uncertainty from compiled liquid pha densities and species molecular weights[42].
The carbon dioxide mole fraction in the oxidizer stream, referred as X CO
2
.Actually,a massflow controller allows the car-bon dioxideflow rate to be adjusted.Thisflow is then mixed with the constant airflow rate to complement the oxidizerflow.
For every parameter,the range investigated within the frame of the prent study is specified in Table2,together with the corre-sponding estimated uncertainty.The former studies by McEnally et al.[29,34]did not ek intend to investigate the influence of blending and oxidizer dilution.Therefore,the prent investiga-tions on the influence of X i C
8
H18
and X CO
2
can be considered an ori-ginal contribution to the study of the sooting tendencies of the heavy hydrocarbons and their blends burning in air that is possibly diluted by EGR.
We investigated two reference benzene massflow , 0.56mg/s(2.00g/h)and0.69mg/s(2.5g/h),at room temperature. Theflow rates correspond to the vapor mole fractions 4.95Â10À2ðX vap;1Þand6.14Â10À2ðX vap;2Þ,respectively.The mass flow rate for a given liquid fuel to be investigated is then adjusted so that its vapor mole fraction is equal to X vap;i(i=1,2).It should be also noticed that the level of vapor mole fraction is here higher than that ud by McEnally and Pfefferle in their different studies. Hence,the sooting characteristics are governed predominantly by the liquid fuel chemistry.Injecting higher concentrations of vapor leads to relatively high levels of soot volume fraction,therefore en-abling that the two-dimensional soot volume fractionfields to be
1578M.Kashif et al./Combustion and Flame161(2014)1575–1586
mapped by LEM.Alternative evaluations of the corresponding YSIs are then allowed to complement the original evaluation prescribed by McEnally and Pfefferle[29],who specifically ud the maximum soot volume fraction at thefixed location z meas on the axis of the flame.
As detailed below,the overall soot load inside theflame can be inferred from the two-dimensional soot volume fractionfield.
3.3.Data processing
In the prent study,three methods are ud to calculate the YSI of a given fuel from Eq.(1),suggesting different alternatives to the evaluation of f v;max:
1.Thefirst method is referred as YSI Max:Ax:fv.The maximum soot
i think of youvolume fraction on the axis of theflame is ud to calculate YSI from Eq.(1).This evaluation is very similar to the one pre-scribed by McEnally and Pfefferle[29].It can also be considered an extension of their original definition,as the spatial resolution of the LEM technique is here higher than that of the discrete field obtained by local LII measurements by McEnally and Pfefferle.
2.The cond method is identified by YSI Global Max:fv.The peak soot
volume fraction f v;max in the whole distribution is ud to calcu-late YSI from Eq.(1).This definition can especially be relevant when the peak soot volume fraction appears in an annular region of theflame.
3.The third method is referred as YSI Max:Int:fv.Here,the integrated
soot volume fraction F vðzÞwasfirst computed as follows:
F vðzÞ¼1
R
Z R
2p r f vðr;zÞdr:ð2Þ
Then the maximum value of F vðzÞwas ud to evaluate YSI from Eq.(1).F vðzÞappears as an evaluation of the actual soot load at a given HAB into theflame.Its maximum value can therefore be c
onsidered a relevant characteristic of the competition between soot formation and oxidation along the globally upward trajectories that particles experience.
To compute the three kinds of YSI specified above,the two-dimensional soot volume fractionfields are required.The are in-ferred from the LEM data,following the processing elaborated in Ref.[40].For everyflame investigated,150concutivefields of raw integrated extinction coefficient arefirst procesd from a 10-s-long LEM recording.Then a deconvolution procedure bad on the onion peeling method stabilized using Tikhonov regulariza-tion[43]is applied to everyfield,providing150maps of local extinction coefficient j ext kðr;zÞ.As prescribed by Zhou et al.[32], the soot particles are here assumed to experience a relatively short residence time in theflame,therefore exhibiting a range of diame-ter that is small enough to warrant the Rayleigh assumption.Thus, the local soot volume fraction isfinally inferred from the local extinction coefficient:
recruitsf vðr;zÞ¼
k j ext kðr;zÞ
6pð1þa saÞEðmÞ
;ð3Þ
where EðmÞis a function of the complex refractive index m of soot.
a sa is the relative contribution of scattering to extinction,and there-fore is directly related to soot morphology.
To overcome the controversial issue of the evaluation of m to-gether with the relative magnitude of scattering into the extinction phenomenon,we proceeded as Zhou et al.[32].Characterizing soot production in diffusionflames burning n-heptane vapor,the authors lected the values of EðmÞand a sa reported by Krishnan et al.[44,45]for soot particles produced in n-heptane diffusion flames.a sa is highly nsitive to the particle morphology.There-fore,m and a sa are not likely to be tup-independent,becau soot morphology is partially governed by soot characteristic residence time[46].
As an illustration of the need for different values of soot optical properties under varyingflame conditions,Zhou et al.[32]and Witkowski et al.[13]ud K e¼9:0and K e¼5:0,respectively, where6pð1þa saÞEðmÞis the dimensionless extinction coefficient. The studies followed the recommendations of Krishnan et al. [44,45]and Dalzell and Sarofim(Ref.[37]in Ref.[13]),respectively. The different values of K e can be attributed to the fact that differ-ent experimental conditions resul
t in different soot residence times,and therefore ultimately in different morphologies of soot particles.Thus,adjusting K e is potentially required to take into ac-count varying soot optical properties.
However,soot morphology depends above all on particle con-centration which is mainly governed by the fuel C/H ratio[47]. Within the framework of the prent study,soot production is mainly attributed to the vapor of the liquid fuel investigated.From pure n-hexane to pure isooctane,the C/H ratio ranges from0.429 to0.444,respectively.Except for benzene,who C/H ratio is unity, the variation of the C/H ratio among the fuels investigated here is then considered weak.Conquently,soot volume fractionfields will be inferred from Eq.(3),where the denominator is equal to 5.1(E(m)=0.27),as recommended by Krishnan et al.With this assumption,the coefficient relating f vðr;zÞto j ext kðr;zÞin Eq.(3)is fuel-independent.
As mentioned by McEnally and Pfefferle[29],this coefficient then factors out when f v;max–or its alternatives–is converted into YSI through Eq.(1).
Interestingly,McEnally and Pfefferle did not discuss the dis-crepancy that may appear among the aforementioned coefficients due to the possible variation of m with the fuel type.A discussion about t
his issue that especially aris when benzene is ud as one of the reference fuels is incorporated into Section4.1.
Table1
Experimental parameters kept constant for the evaluation of YSIs.
Parameter Prent study McEnally and Pfefferle
[34]
Ambient pressure(atm) 1.0 1.0
Ambient temperature(°C)2020
Carrier gas composition CH4(100%)CH4(55%),N2(45%) Carrier gasflow rate
(cm3/min)
200Æ1.2605
Coflowing airflow rate
thanksgiving(cm3/min)
60,000Æ36030,000
Evaporator temperature(°C)150Æ1N.A.
Heated line temperature(°C)150Æ2145月份缩写
Reference YSIs YSI(n-hexane)=0YSI(n-hexane)=0
YSI(benzene)=100YSI(benzene)=100Table2
Experimental parameters varied for the evaluation of YSIs.
Parameter Prent study McEnally and Pfefferle[34] X i C
badly
8
H18
Range[0,1]0/1
Step0.2N.A.
Uncertainty4Â10À3N.A.
X vap;i Range X
vap;1
¼4:95Â10À210À3
X vap;2¼6:14Â10À2
Step–N.A.
Uncertainty3Â10À310À5
X CO
2
Range[0,0.06]N.A.
Step0.03N.A.
Uncertainty1:5Â10À3N.A.
M.Kashif et al./Combustion and Flame161(2014)1575–15861579
