Publisher’s version / Version de l'éditeur:
Proceedings of Combustion Institute - Canadian Section Spring Technical
Meeting, 2009, 2009-05
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.
https://nrc-publications.canada.ca/eng/copyright
Vous avez des questions? Nous pouvons vous aider. Pour communiquer directement avec un auteur, consultez la première page de la revue dans laquelle son article a été publié afin de trouver ses coordonnées. Si vous n’arrivez pas à les repérer, communiquez avec nous à PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca.
Questions? Contact the NRC Publications Archive team at
PublicationsArchive-ArchivesPublications@nrc-cnrc.gc.ca. If you wish to email the authors directly, please see the first page of the publication for their contact information.
NRC Publications Archive
Archives des publications du CNRC
This publication could be one of several versions: author’s original, accepted manuscript or the publisher’s version. / La version de cette publication peut être l’une des suivantes : la version prépublication de l’auteur, la version acceptée du manuscrit ou la version de l’éditeur.
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
Equivalence ratio gradient effects in a stratified isooctane/air turbulent
flame front
Vena, P. C.; Deschamps, B.; Smallwood, G.J.; Johnson, M. R.
https://publications-cnrc.canada.ca/fra/droits
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
NRC Publications Record / Notice d'Archives des publications de CNRC:
https://nrc-publications.canada.ca/eng/view/object/?id=d325d3e2-05d9-4e28-8893-d8b11d090f65 https://publications-cnrc.canada.ca/fra/voir/objet/?id=d325d3e2-05d9-4e28-8893-d8b11d090f65
1
University of Montreal, Quebec May 11-14, 2009
Equivalence Ratio Gradient Effects in a Stratified Isooctane/Air
Turbulent Flame Front
P.C. Vena
a, B. Deschamps
a, G.J. Smallwood
b, M.R. Johnson
a*a
Mechanical and Aerospace Engineering, Carleton University, Ottawa, Canada
b Combustion Group, Institute for Chemical Process & Environmental Tech., NRC
Introduction
Despite the importance of partial premixing for practical combustion devices and potential advantages of developing successful stratified charge engines, few fundamental works have attempted to quantify the effects of stratification on flame behaviour. The difficulty lies in implementing an experiment capable of precisely reproducing partial premixing and coupling it with the necessary diagnostics to evaluate complex flames. Only recently, with the advent of novel optical diagnostics, has this been a realistic proposition. The development of practical high energy lasers, adequate data acquisition systems, and the necessary image processing tools have led to the non-intrusive visualization techniques necessary to accurately probe the physical and chemical phenomena in turbulent partially premixed flames.
In the published literature, two experimental approaches are generally considered. The first considers freely propagating flames, where the mixture is centrally ignited and propagates parallel to the mixture fraction gradient [1-5]. The second approach, used in the current study, considers rod stabilized V-flames, where the leading branches propagate transversely to a gradient in mixture fraction [6-10]. The primary emphasis of these studies is to investigate dynamic properties of flames including flame front topology and reaction rate. Limited results have shown conflicting gradient effects on curvature [4,10] and
flame surface density [6,7,9,10]. Although a recent study by Anselmo-Filho et al. [10] considered effects of stratification on flame front topology, to the authors’ knowledge, no experimental studies have yet considered gradient effects on the reaction rate in terms of heat release
This paper describes an experimental investigation on partially premixed combustion through the use of a unique burner that enables longitudinal variation of equivalence ratio along a flame front. A PLIF imaging setup that images OH, CH2O, 3-Pentanone, and biacetyl has been implemented to
correlate local equivalence ratio gradient to progress variable, flame surface density, curvature, and heat release in a stratified isooctane/air V-flame.
Experimental setup
A simplified diagram outlining the progression of the reactant mixture as it flows through the stratified burner is provided in Figure 1. Two fully premixed reactant streams of different mixture strengths φ1 and φ2 enter separate compartments within the 16 mm high lower section of the burner. The two
* Corresponding Author: Matthew _Johnson@carleton.ca
stoich streamline
φ
leanφ
rich smooth longitudinal variation in mixture strengtheven lateral dist in six channels, each with
different A/F
reactant mixtures φ1 and
φ2 at exit of lower
mixing section
Figure 1 Reactant mixture progression though the stratified burner
Laser Diagnostics
flows then pass through a 15 mm wide by 65 mm long by 12.7 mm thick section of expanded metal foam which ensures the flow is uniformly distributed as it enters the lateral mixing section of the burner. Because a sealed metal barrier separating the two flows cuts diagonally across the metal foam, the flow is partitioned into two equal sized triangular sections. A proportionate amount of each reactant flow φ1 and φ2 is distributed along the longitudinal axis of the burner, so that the net equivalence ratio varies linearly from φ1 to φ2 along the burner. As the flows exit the
triangular sections of metal foam, they enter a lateral mixing section consisting of a set of six, equally spaced, 10 mm wide, 152.4 m long channels filled with 3 mm balls. In this section of the burner, the two different mixture streams mix laterally along the short axis of the burner. Each of the six channels thus has a different global mixture strength, which is linearly proportional to φ1 and φ2. Finally, the flow is allowed to mix freely in the longitudinal
and lateral directions in the rectangular cross-section at the exit section of the burner to produce a reactant flow with a longitudinally varying gradient in equivalence ratio.
The mixture strength and mean flow velocity are independently controlled using two pairs of digitally controlled Horiba LV-F liquid (isooctane) and Brooks Smart Series gas (air) mass flow controllers. The velocity and turbulence intensity profiles were reasonably uniform (< 20% variation), although there was some evidence of ripple associated with the six separate channels upstream in the burner. The mean flow velocity from along the longitudinal axis of the burner exit plain, ±16 mm from the centerline of the burner was 4.6 m/sec while turbulence intensities were between 4.2 – 7.2 % within the same longitudinal range.
A 1.5 mm diameter rod placed perpendicularly to the exit nozzle of the burner was positioned at x=16.5 mm from the inner edge on the rich side of the exit nozzle to stabilize the partially premixed flame in a V-configuration. A continuous stratified flame front thus intersected the flow centerline, shown in Figure 1 as the stoichiometric streamline. Six flames were considered by increasing the equivalence ratio from 1 to 2 on the “rich” end of the burner, while decreasing from 1 to 0 on the lean end, in equivalence ratio steps of 0.2.
In the current experiments, PLIF was used for flow visualization of OH (hydroxyl) radicals and CH2O
(formaldehyde) for near-instantaneous flame front topology and heat release measurements. Tracer LIF of 3-Pentanone (C2H5COC2H5) and Biacetyl (CH3COCOCH3) were also used for qualitative mixture fraction
measurements at the exit of the burner, as well as laser profile corrections. The laser diagnostic setup for
simultaneous OH and CH2O PLIF shown in Figure 2 consists of a dual head Nd:YAG laser (Quanta Ray PIV 400)
with independent crystals for doubling or tripling the fundamental frequency.
The left side IR beam, frequency tripled from its 1064 nm fundamental wavelength to the third harmonic at 355 nm (+/- 3 nm, 300 mJ per 7 nsec pulse), is used for CH2O and Biacetyl excitation. The right side IR beam is frequency
doubled to 532 nm and is subsequently used as the pump source for a Rhodamine B dye laser (Sirah Precision Scan). The tuneable output from the dye laser (282.92 nm +/- 0.005 nm, 11 mJ per 6 ns pulse) is used to excite OH and 3-Pentanone. The produced 20 mm high x 200 micron thick, 355 nm and 282 nm coincident laser sheets intersect the central plane of the burner exit nozzle.
Fluorescence images were acquired using two Princeton Instrument 384x576 pixel intensified charged coupled device (ICCD) cameras. The cameras and associate lenses were positioned to provide a projected spatial resolution in the flame ranging from 131 µm/pixel. This equated to a 50.3 mm by 75.5 mm effective field of view within the flame. Precise timing of lasers and cameras was achieved using a DG-535 Stanford pulse generator. Flashlamp/Q-switch pulses for both lasers were phase shifted so that the 282 nm (OH/3-pentanone) and 355 nm (CH2O/Biacetyl)
beams were delayed by 210 nsec. This phase delay was necessary to avoid overlap of the fluorescence signals induced by two laser sheets. Since OH has a shorter fluorescence lifetime than CH2O, OH was excited first.
3 Results and Discussion
The equivalence ratio gradient for each of the six flow settings under non-reacting conditions was characterised using 3-pentanone PLIF by seeding the isooctane with 3-pentanone at 4% by volume. As was shown in a previous paper, the equivalence ratio profiles along the longitudinal axis of the burner intersected where φ≈1, which corresponded to the stoichiometric streamline for all flame settings [13]. Thus, by focusing on a flame region in the immediate vicinity of the stoichiometric streamline, the effect of mixture fraction gradient on the local flame could be isolated. Since the mean exit velocity of the burner was fixed for all cases, the mean flow field and turbulence characteristics were consistent within the locally stoichiometric region .
3.1 Progress Variable
Figure 3 shows flood plots of reaction progress variable in the vicinity of the stoichiometric contour for each of the six mixture fraction gradient settings. As indicated on the intensity scale to the right of the images, dark blue corresponds to fresh reactants ( =0) and dark red corresponds to pure products ( =1). Each full image corresponds to a 26.2 mm x 19.8 mm area within the flame.
Figure 3 Progress variable maps and corresponding interrogation window for homogeneous and stratified flames
In Figure 3, it is important to note that images corresponding to stronger mixture fraction gradients contain a wider range of local equivalence ratios within the same spatial region than images under less stratified conditions. As the equivalence ratio in an octane-air flame is varied, the Lewis number varies from Le=3.21 for φ=0.8 to Le=0.98 for
Laser Diagnostics
φ=1.5 [14] and affects the topology of the flame front. Thus, to permit fair comparison among conditions and to isolate the effects of equivalence ratio gradient variations from other effects, it was also necessary to maintain a constant range of equivalence ratios within the analysis region of the flame. This was accomplished by varying the width of the region of flame being analyzed, as indicated by the superimposed dark boxes in Figure 3. These boxes indicate the analysis regions used at each setting, which were centered along the stoichiometric streamline and defined so that the mean local equivalence ratios spanned an identical range of φ≈0.94 to φ≈1.06 in all stratified flame configurations. By choosing these variable width analysis regions, it was possible to assume that flames within were propagating under essentially identical mean flow and equivalence ratio conditions, but were in the presence of widely differing gradients.
For the severely stratified cases, the maximum value of the progress variable, , within the images does not
always reach 1 (i.e. for and for ). For these
more corrugated flames, the OH distribution behind the front was not nearly as uniform. This does not necessarily imply that the burnt gases are not present past the flame front, but only that the OH fluorescence is weaker. However, because of the associated uncertainties in analysis, data for the steepest gradient were not considered in most discussion throughout the remainder of the chapter.
3.2 Flame Surface Density
The two dimensional flame surface density (FSD) data were calculated from the binary OH PLIF images. Results were plotted in terms of for cases where the analysis window spanned the entire flame front and the progress variable ranged from approximately 0 to 1.
Figure 4 shows that FSD profiles were slightly skewed so that maximum values fell consistently toward the burnt gases, in the vicinity of =0.6. The asymmetric profiles were consistent with those observed by Shepherd [16] for methane/air and ethylene/air turbulent V-flames, such that the expression, used for chemical closure models for fractal flamelets that represents the probability of finding a flamelet within the flame zone (where k is a proportionality factor) [15], does not provided a good estimate of . Shepherd [16] proposed that the ordered, cusped structure of the V-flame, confirmed by their observation of bimodal PDFs of flame orientation and caused by cusping towards the burnt gases, contributes to the asymmetry in FSD.
Although the shape of the FSD profiles did not vary among settings, comparing the homogeneous and stratified profiles, there is arguably a non-monotonic variation in FSD with gradient. In comparison to an unstratified flame, a
weak mixture fraction induces an increase in FSD, whereas progressively stronger gradients
show progressively lower FSD values. For the strongest mixture gradient case, the measured FSD was consistently lower than that of the homogeneous flame. These findings are contrary to what was expected from the qualitative visual evaluation of the fluorescence images, in which stronger mixture fraction gradients were associated with increased flame front wrinkling, corrugation and turbulent flame brush thickness.
By definition, a flame’s surface density is the ratio of its surface area to volume, or length to area for two dimensional images. It is clear that if the flame’s 2D area increases with gradient, confirmed from the observed change in brush thickness, the flame length must also change, and do so proportionally, or near proportionally, for FSD values to remain nearly constant. This implies that although the wrinkling length scales, or the size of the wrinkles, appear to vary with gradient according to the individual images, they in fact do not do so significantly, rather there are simply more of them. Because the volume in which the greater number of wrinkles are present increases nearly proportionally to the flame surface area, the density of the flame surface does not change significantly.
3.3 Curvature
The length-weighted PDFs for each flame setting were calculated by normalizing the number of observations in each bin by the total number of local curvature measurements and plotted in Figure 5. The curvature PDFs were nearly symmetrical, with a slight bias toward the reactants. This is consistent with the FSD results of Figure 4, and the generally accepted trend for cusped flame fronts, where the larger area of positive fronts results in the biased curvature plots. The shift is further emphasized by the negligible flame area of the infrequent cusps, which manifest themselves as large negative curvatures. [17]
Figure 5 Comparison of curvature PDFs for globally stoichiometric premixed and partially premixed isooctane/air V-flames
The distribution in curvature widens as the gradient increases, and this is particularly apparent for severely stratified cases where local extinctions became more prominent. This increase in variance is attributed to the greater number of wrinkling scales with larger curvatures (smaller radii) along the front, observed in flames with lower Markstein numbers [18]. This further suggests the dependence of Ma on equivalence ratio gradient in that a decrease in Ma increases the curvature distribution along the flame front [18].
Laser Diagnostics 4
The maximum curvature was consistently around , and the probability of encountering this level of
curvature decreased for stronger gradients, consistent with observations of Anselmo-Filho et al. [6]. Although the bias of the profiles did not change with gradient like those of Anselmo-Filho et al. [6] , the curvature distribution did vary consistently and monotonically.
3.4 Heat Release
Heat release maps were obtained from the pixel by pixel product of spatially aligned OH and CH2O images. The
results are plotted as a function of in Figure 6.
Figure 6 Heat Release for premixed and stratified flames
The heat release profiles were skewed much like those of FSD in Figure 4. HR for the homogeneous flame peaked at approximately 0.009 at ≈0.65. The stratified flames all showed comparatively higher global heat release rates than the premixed reference case. This implied that that the combustion intensity increased despite the only small
and was consistent with the increased 2D flame length discussed previously. For
, the mean heat release rate began to decrease, confirming the presence of local extinctions changes in FSD
observed in individual images.
Conclusion
The effect of partial premixing on turbulent isooctane V-flames has been studied through the use of a novel stratified burner capable of producing longitudinal variations in mixture strength. PLIF imaging of OH, CH2O, biacetyl, and
3-pentanone was done in order to evaluate the variations in flame front topology and combustion intensity.
It has been shown that gradients in equivalence ratio affect flame front wrinkling for locally stoichiometric mixtures in terms of flame surface density and curvature. However, variations were small relative to those observed for lean methane/air V-flames under comparable flow conditions. The variation in heat release was more significant, with consistently greater values in stratified cases than the reference homogeneous case.
These observations suggest that equivalence ratio gradients alter the behaviour of stratified flames and further hint at the possible interaction between neighbouring flamelets. This can potentially limit the application of flamelet models for partially-premixed flames.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 REFERENCES
[ ] Y. Ra, W.K. Cheng, The Fifth Int'l Sym on Diag and Mod of Comb in Internal Comb Engines (2001), 251-257. [ ] A. Pires Da Cruz, A.M. Dean, J.M. Grenda, 28th Symp. (Int'l), The Comb. Inst. (2000), 1925-1932.
[ ] T. Kang, D.C. Kyritsis, Comb. Sci. and Tech. 177 (2005), 2191-2210.
[ ] B. Renou, E. Samson, A. Boukhalfa, Comb. Sci. and Tech. 176 (2004), 1864-1890. [ ] N. Pasquier, et al., 31st Symp. (Int'l), The Comb. Inst. (2007), 1567-1574.
[ ] O. Degardin, PhD thesis (2006), Institue National des Sciences Appliquees de Rouen. [ ] Robin V., et al., Combustion and Flame 153 (2008), 288-315.
[ ] D. Pave, PhD thesis (2002), l’Universite d’Orleans.
[ ] C. Galazzi, D. Escudie., Combustion and Flame 145 (2006), 621-634.
[ ] Anselmo-Filho P., Cant R.S., Hochgreb S., Barlow R.S., 32nd Symp. (Int'l), The Comb. Inst. (2008), 269-290.
[ ] P.C. Vena, B. Deschamps, G.J. Smallwood, and M.R. Johnson, CICS, Waterloo (2006), ON, paper G4-1.
[ ] O.L. Gülder et al., Combustion and Flame 120 (2000), 407-416.
[ ] P.C. Vena, B. Deschamps, G.J. Smallwood, and M.R. Johnson, CICS Toronto (2008), ON, 148-153.
[ ] Takashi H., Kimitoshi T., Féd Int des Soc d’Ing des Tech de l’Auto (FISITA) Student Congress (2006), SC40. [ ] Gouldin F.C., Miles P.C., Combustion and Flame 100 (1995), 202-210.
[ ] Shepherd I.G., 26th Symp. (Int'l), The Comb. Inst. (1996), 373-379.
[ ] Lee T.W., North G.L., Santavicca D.A., Combustion and Flame 93 (1993), 445-456. [ ] Haq M.Z., Journal of Engineering for Gas Turbines and Power 128 (2006), 455-462.
