вход по аккаунту


The Effect of Precession on the Ignition Region in Jet Flames.

код для вставкиСкачать
Dev. Chem. Eng. Mineral Process., 7(3/4).pp.345-359, 1999.
The Effect of Precession on the Ignition
Region in Jet Flames
N. Yousefpour, J. Reppel", Z.T. Alwahabi;
G.J. Nathan and K.D. King
Departments of Chemical and Mechanical Engineering, University of
Adelaide, Adelaide, South Australia 5005
Planar Laser-induced Fluorescence (PLIF) of the OH radical was used to study the
base of the active combustion region in the flame produced by a mechanically
precessed jet (MPJ) nozzle firing methane. Instantaneous images of the OH radical
distribution were obtained for a range of jet flow rates and frequencies of jet
precession about the central jet mis. The MPJ burner has been developed to better
understand the effects of jet precession which have been found to reduce NO,
emissions and increase emissivity in industrialflames. Broad distributions of the OH
radical with widths up to 31.3 rnm are shown to exist in the flame. Furthermore,
combustion at the base of the flame is seen to occur not only at the periphety of the
frame but also along the nozzle mis indicating the presence of a stoichiometric jkel
air mixture there.
The OH radical is a key reaction intermediate in the combustion of hydrocarbon fuels,
indicating the position of the instantaneous flame zone. Schefer et al. [ 11 have shown
that, while the presence of the CH radical indicates the flame front, the hot flame zone
is identified by the broader OH distribution in turbulent diffusion flames. Planar
Laser-Induced Fluorescence (PLIF) of the OH radical in the flame from a
Authorfor correspondence.
N. Yousefpour et al.
mechanically precessing jet (MPJ) nozzle provides information on the complex flame
structures produced by the precession of a jet.
The fluidic precessing jet (FPJ) burner was developed by the Department of
Mechanical Engineering at the University of Adelaide [2] as a result of studies into jet
excitation, and is the subject of ongoing research [3]. In turbulent jet diffusion flames,
mixing rates are much slower than reaction rates so that mixing controls the
combustion. Control of mixing is therefore a major aim in burner design. The FPJ
nozzle produces a naturally occurring flow phenomenon in which the flow field
within the nozzle precesses azimuthally around the nozzle axis quasi-periodically.
The jet which emerges from the nozzle chamber thus precesses about the nozzle axis
[4]. The scale of turbulence generated immediately downstream from the nozzle is
several times that from the mixing caused by shear alone in a simple turbulent jet.
The flame produced from the FPJ nozzle is markedly different from that of a
conventional round jet nozzle. The flame is bulbous and highly luminous and is
characterised by a stand off height reduced by an order of magnitude, shown to be
almost independent of throughput [ 5 ] . The flame is also very stable over a wide range
of operating conditions thus allowing for large turndown ratios.
Parallel commercial development of the FPJ has led to the GYROTHERM rotary
kiln burner which has already established itself in the cement, alumina, lime and
metal industries. These industries benefit from the highly luminous flame produced
by the FPJ burner which dramatically increases the radiant heat transfer, lowers
specific fuel consumption and, consistent with a lower mean flow temperature [6],
typically reduces NO, by about 50% when compared to existing burners at similar
excess air ratios [7, 81.
Although of simple mechanical design, the FPJ nozzle produces a complex
continuously unstable flow field that has proven difficult to investigate at the
fundamental level. The mechanical precessing jet nozzle (Figure I ) was developed to
better investigate the effects of precession via mechanical rotation of an inclined
simple jet nozzle with origin on the jet axis about the nozzle axis. The rotation of the
nozzle causes the inclined jet to precess about the axis of rotation. The exit
Effect of precession on ignition region in jetflames
conditions, precession frequency, f, and angle of jet inclination, Cp, from the rotation
axis of the jet which emerges from the mechanical precessing jet (MPJ) nozzle, unlike
those of the FPJ nozzle, are well defined and can be independently varied.
Figure 1. Schematic representation of the MPJ nozzle showing the fuel jet exiting the
nozzle with a velocity, Uo , at an angle, A to the nozzle axis. The nozzle is rotated
around the nozzle axis, z, at afrequency of precession,f.
The precession of the emerging jet can be characterised in terms of the precession
frequency, f, jet velocity, U, , and nozzle exit diameter, d, by the dimensionless
Strouhal number of precession, St,.
Stp =-.
The Strouhal number has been shown by Schneider et al. [9] to characterise the
flow for a fixed jet exit angle, Cp.
Nathan et ul. [ 5 ] showed that hydrocarbon flames with high Strouhal numbers of
precession (St, > 0.01) are 15% more radiant and reduce NO, emissions by 25%
relative to a conventional turbulent jet diffusion flame. When compared to high
Strouhal number flames, flames in the low Strouhal number regime (St, < 0.002) are
much shorter and bum with a blue colour. They have high CO emissions and
NO/NO,, indicating rapidly quenched reactions.
N. Yousefpour et al.
Figure 2 shows a photograph of au MPJ flame and the region of current PLIF
investigation iuto the base of the precessbg jet OH distribution.
FigrvC 2 A photogrqh of a mtud gasjIivne@m a mechaniwlprecessirsgjet
Mlale showing the region of curreniPLlF imstigation of the OH r&cal.
A schematic of the experimental arrangement is shown in Figure 3. The second
harmonic of a Nd:YAG laser at 532 nm (Surelite Contiuuum) was used to pump a
tunable dye laser (Lambda Physik SCANMATE, utilisiag DCM dye) producing short
(< 6 ns), narrow bandwidth pulses (< 0.2 cm" ). The output radiation of the dye laser,
at 16190.504 cm",was doubled in frequency by a BBO crystal. A Pellen Broca prism
was used to select the required second harmonic W radiation. The energy per pulse of
the selected W radiatation was
- 9 mJ measured at a pulse rate of 10 Hz.The W
beam was then directed to an expanding telescope arrangement producing a laser sheet
of height 78 mm and width 0.7 mm which passed above the MPJ nozzle through the
n o d e axis.
The MPJ nozzle used has an exit angle t$ = 45 degrees and an exit diameter d = 3
Effect of precession on ignition region in j e t j l m e s
mm.Rotational frequency of the nozzle was controlled to 0.01 Hz using an elmomc
controller (ABB). The temperature of the non-premixed methane flame was measured
using a bare wire themocouple, type K.
The values of the temperature measured across the flame were relatively low,
compared to those of a simple jet methane flame. Based on these measurments, the
average flame tempreture was calculated to be -900 K.This low flametemperature is
characteristic of fluidic and mechanically prectssed flames. Further systematic non-
intrusive temperature profile m m e n t in precessing Barnes is currently underway.
In OH radicals the population of the N = 6, at an average flame temperature of
K, is the least semsitive to temperature variations [lo]. S i l e calculations
showed that the variation of the N = 6 population was less than 3% over the
temperature range of 800-1400 K. Therefore the OH radicals were excited to the
AzC(v = 0, j = 5.5) spm-ro-wionical state through the Ql(6)transitiooS [l 11.
The laser-induced fluorescence &nd fiom the OH radicals was collected ushg an
f14.5 104 mm W lens (Niior) coupled to a gated intensified CCD camera (Princeton
InstrumentsICCD-576) orthogonal to the laser sheet to image 8 flame area Of k@t
of 86 mm and width of 128 mm, as shown in Figure 1. A gate width of 30 11s was
selected to minimise background radiation fiom the h e . The base region of the flame
chosen for investigation is blue so that soot int-
is negligiile [121.
NdYAG Lasa
Figure 3. Experimental arrangement used for PWF imaging of the OH radical in a
rtanualgasjbne, from a MPJ nozzie.
N. Yousefpour et al.
Table 1. Experimental conditions
(m s-')
Vdocity Prcarrion
0.61 E -02
0.74 E -02
0.84 E -02
0.96E -02
0.96 E -02
0.96E -02
1.14 E -02
1.49 E -02
1.92 E -02
Nine combinations of Reynolds and Strouhal number? shown m Table 1, were
investigated. The fuel jet Reynolds number range (2150 C & < 4310) covm the
laminar-hubulentReynolds number transition for pipe flow (131. The Strouhal number
range (0.0061 < Stp < 0,0192) covers the transition between "high" and "low"
Strouhal numbers described by Nathan et al [5]. For comparison, the flame from a
simplewbulent jetatReynoldsnumberof&= 2590 (Stp= 0) andjet diameter d = 3
mrn was included in the investigation.
One hundred instantaneoUs imap;es wete uncanditionally sampled with respect to
nozzle orientation to the light sheet, of which ten images were selected for each
condition for further analysis and corrected for non-uniform laser sheet intensity.
Figure 4 shows the dishctive OH distribution associated with the MPJ flame. The
unconditionally sampled image represents a slice through the flow field of the
mechanically precessing flame with a Reynolds number Re,, = 4310 and Strouhal
Stp = 0.0074. The asymmetric natufe of the precessiug jet produces
continuously unstable flow field which results in complex OH distriiutions. However,
these OH distriitions demonstrated common characteristics such as, lift-offheights,
total flame widths and distribution of OH region thicknesses for each
Effect ofprecession on ignition region in jet flames
StrouhaUReynolds condition. The distinctivefeature of the characteristic mechanically
precessed h
e imaged in Figure 4 is the presence of both thin and thick reaction
structures within the flameregion, especiaY.dong the the nozzle axis. Convoluted OH
distriitions, including sisnificant variation in the width of the reaction region, are
indicative of the strong mixing field introduced by the precession of the fbel jet.
Regions of OH radicals along the axis of the nozzle were found to exist in all images of
the the MPJ flame.This suggests the extensive
of surroundmg air into the
axial region. These features diffir markedly to the OH distribution in a simple lifted
turbulentjet flame. In studies of the latter, Schefer et ul. [1, 141 have shown that for a
jet exit velocity of 21 mls (R% = 7000), the OH radical is fonned in a narrow, 2 4 mm,
sharp envelope which cwves around large scale semisrganised vortical structures at
the outer fbel jet bounday. The OH radical is not present along the jet centre line
indicating no fuel combustion of fuel occurs on the jet &,since the mixture M o n
of the &el is above the rich flammability limit. This has also been v d e d by Alwahabi
et uZ.[12]
for a similar flame (Re,+
= 2735,
Stp = 0, d = 3 mm,
4 = 00)and by the
present data for Stp = 0.
The distribution of the OH radical in the flame &om a MPJ nozzle with Stp = 7.4 x
is complex, exhiiting both broad, extended OH regions and thin regions
comparable with those of Schefer [11. Cold flow investigations of the wncemtration in
the flow produced by a MPJ n o d e have demonstrated that regions exist in which the
mixture fiaction is in the flammable range for methane and that stoichiometricregions
exist along the n o d e axis[ 151. The large areas of non-reacting fluid within the region
bounded by the OH radical found in the present flames also correspond closely with
the regions of fluid in the non-reacting case in which the mixture fiaction is higher than
the rich flammability limit.
N. Yousefpour et al.
- .
25 50 75 100 125 150 175 200
OH intensity
I"igure 4. Instantanems images of the OH d c a l ciistribution, uncodtional!~
sampled with respect to the nozzle direction, in a j h e prorhrced by a mechanical
precessingjet nozzle. Note the presence of OH r d c a h along the nozzle mcis and the
presence of both thick and thinjlame zones (Condition b: R e d
4310, Stp
The thickness of the reaction zones in the present flames bave been qmntified by
using a cut-off in which the OH Concentration cOrreSpOadS to 15% of the maximum
value. F~eenpercent of the mBlLiRIum value c ~ m s p m d sto a signal strength eight
and has been found to provide a representative
edge discrimination technique. The maximum and minimum OH zone thicknesses are
presented in Figure 5 for both precessing and non-precessing flames as determined
using identical experimental techniques and cut off criterion. The average simple jet
flame thickness was measured to be 3.0 mm. Continuous broad zones in the
times that of the background noise
Effect of precession on ignition region in jetflames
instantaneous OH PLF images of up to 3 1.3 mm, approximately 10 times the average
thickness of the simple jet flame structures, have been found for the range of
conditions investigated, with the average maximum OH structure thickness of flames
produced by the MPJ nozzle being 9 18 mm depending on the experimental
conditions. The error bars represent the rms standard deviations of the measured
thicknesses determined fiom 10 images. Data points representing different Reynolds
number ranges are indicated by different point symbols.
Fipm 5. Minimum and d m u m OH zone thickneses for the Reynokh number
ranges of Table I plotted against wdtion snarhal number. i’he error bars
represent the rms standardWations of the dbta sets.
The broad flame regions indicated by OH PLIF images of flames fiom a MPJ nozzle
suggest that broad regions of fluid exist within the Elammabity limits in conditions
capable of supporting combustion. Since reaction can be extinguished either by high
scalar dissipation [161 or by high local velocity [17], it suggests that neither of these
conditions apply. The breadth ofthe reaction zone itself also implies directly that low
N. Yousefpour et al.
concatration gradients, and hence low scalar dissipation, exist there. The presence of
ofsclmeider et al.
[9]. These deductionsare consisteat with the high flame stability that c8n be achieved
with the MPJ nozzle and are supported by the increased blow-ou! velocjties when
compared to simple turbulent jet flames [5]. This, as well as the presence of
combustion along the nozzle axis at the base of the flame, implies that flame
stabilisation is controlled by the turbulent mixing characteristics generated by the
nozzle upstream fiom the reaction zone.
low velocity fluid can be deduced from the cold flows -
n mmumIo
n m a*~muw
figute ti Instanmous images of the OH rdcial dim'htions u n c d t i d b
sampled with respect to the nozzle direction, showing shuclural similmitiesforflames
with a constant Saouhal mmberfbr increasing Reynokh number. Codtions: Stp =
0.96 E02, (a) Red
= 2710, @) Red = 3340, (c) R e d = 4310.
A comparison of the characteristic structure of the OH distribution, as a bction of
Reynolds and Strouhal number, also shows that the flame structure is dependent on the
flow produced by the nozzle. However the flame structure is innuenced more strongly
by Strouhal number than by Reynolds number m the present range. Instantaneous
images (Figure 6), of flames d, e, and f (Table l), demonstrate the similarity in flame
structure for increasing Reynolds number conditions (2710 < RQ < 4300) at constant
The distribution of OH within the flame found to be a strong function of Strouhal
number within the range of conditions investigated (Figure 7). In the lower
Effect of precession on ignition region in jetflames
experimental Strouhal number jet conditions, the OH radical is found to exist
predominantly in a single broad diffise structure. As the Strouhal number is increased,
these broad diffuse forms tend to split into two to threesmaller, sharper, narrower and
betterdefined structures. There also appears to be a shift in the OH distribution
toward the periphery of the flame. The “pockets” of non-reacting fluid that are
surrounded by regions containing OH,signifjllng reaction, can be deduced to be fuelrich and seem to occupy more of the cross section of the flame as the Strouhal number
is increased.These trends are evident in Figure 7.
Along with the change in the OH distribution with Strouhal number, the shape of
the flame also changes fiom being short and pale blue, with the occasional yellow
flicker, at low Strouhal numbers to a taller, narrower, more luminous orange-yellow
flame at the highest Strouhal number. This trend is collsisteM with the presence of
larger “pockets” of hel-rich fluid being formed as Strouhal number is increased,
providing more time for soot hrmatim. This deduction is supported by the
measurements of Nathan et al. [5], who found that the global residence time in flames
increeses with Strouhal number and linked this to greater time fbr soot formation.
Despite the di8Ferences in local h e structure, the presence of broad reaction regions
across all Strouhal numbersis clearly ewident m Figure 7.
The increased size of the pockets of unburned fluid with increased Strouhal number
is fiutherhighlighted h F w 8. Although theReynolds number h these images
= 2 150) is lower than that in Figure 8 (RQ = 4 130), the Strouhal numbers are higher.
The images also suggest a trend that the diffbse regions of combustion are less
relevant, although the present range of data is too small to asceru~ll
From the conditions examined here, the Strouhal number can be seen to have a
much greater influence on the flame shape and structure than does the Reynolds
number. This is consistent with the cold flow results of Schneider et al. [9], who found
that the mixing characteristics in an MPJ flow are much more strongly influenced by
Strouhal number than Reynolds number. It further supports the conclusion that the
mixing characteristicsgenerated by the nozzle dominate the combustion.
N. Yousefpour et al.
25 50
75 100 125 150 175
Figure Z Instantcmeous images of the OH radical dstribution, unconditional&
sampled with r e p c t to the nozzle drection, ofjhnesfrom the MPJ nozzle showing
the change in the flame structure as a function of the S~ouhalnumber at constant
Reynokds number. Conbitom: Red
4310, (a) Stp
(c) Stp = 0.84E-02, (4 Stp = 0.96E-02.
= 0.6lE-02,
(b) Stp = 0.74E-02,
Effect of precession on ignition region in jetflames
figurn 8. I
~ images
m ofs the OH radical a’istribution, umnd~tionally
sampled with respect to the noale &metion, at high Strmhai number showing kger
“ p k e t s’’ of unbumedjluidand less sigm@ant difise Sauctures of OH at the flame
penpky. (a): R e d = 2150, Stp = 1.49E-02, (b): Red = 2I50, Stp = I.92E-02.
Planar laser-induced fluorescence (PLIF) has been used to study the active
combustion region, signi6ed by the presence of the OH radical, in a natural gas flame
produced a mechanically pmxssing jet (MPJ) nozzle fix a range of Strouhal numbers
and Reynolds numbers. Although the range of Reynolds numbers and Strouhal
numbersis limited,clear trendsare evident.
The instantaneous PLIF images indicate the presence of both broad and narrow
distributions of the OH radical across all Strouhal numbers although the broad regions
tend to dominate more in the low Strouhal number range. Zones of OH signal with
thicknesses as large as 33.1 mm were found, with the average mrvrimUm OH structure
thicknesses of 9
- 18 mm depending on experimental conditions. This is in marked
contrast to thicknesses of 2 4 mm typical in a non-precessingjet.
Combustion products at the base of the flame were seen to occur not only at the
periphery of the precessing flames but also in the region along the node axis. This is
consistent with the cold flow identihtion of Nobes et al. [151 who measured mixture
N. Yousefpour et al.
fractions in the region corresponding to the base of the flame along the axis of the
nozzle. Taken together these results suggest that the mixing characteristics of the
nozzle upstream from the flame zone dominate the combustion characteristics.
At low Strouhal numbers, the OH radical was found to occur predominantly in a
single broad diffuse structure surrounding small “pockets” of non-reacting fluid. As
the Strouhal number is increased at a constant Reynolds number, the size of the
pockets of unburned fluid increases and the thinner reaction zones appear to become
more dominant. This trend corresponds in a shift from blue to yellow flames,
suggesting that the pockets of unburned fluid are fuel rich and that increasing
Strouhal number results in more time for soot formation. Qualitative analysis also
suggests a shift in the OH distribution to the periphery of the flame.
Instantaneous images obtained at a constant Strouhal number and increasing
Reynolds numbers were found to exhibit similar characteristics indicating that, for the
range of Reynolds numbers investigated (2150 < Re,, < 4310), the distribution of the
OH radical within the flame is more strongly dependent on the Strouhal number.
Future experiments will investigate the behavior of the mechanical precessing jet in
the higher Reynolds and Strouhal number regimes.
Figures 2,4,6, 7 and 8 were originally presented in colour, however due to economic
constraints we have been required to reproduce them in black and white. Copies of
the original figures can be obtained from the authors.
This work is supported by the Australian Research Council through its Research
Infrastructure (Equipment and Facilities) Program and Collaborative Grants schemes,
and by Fuel and Combustion Technology. Discussions with Greg Newbold and David
Nobes have added valuable insight to the work. The technical support of Brian
Mulcahy from the Engineering Workshop is also acknowledged.
Effect ofprecession on ignition region in jet James
u o
Jet exit diameter
Precession frequency
Jet exit velocity from the nozzle
Reynolds number
Strouhal number
1. Schefer, R.W., Nmazian, M.and Kelly, J. (1990),
“CH,OH, and CH4 ConcentrationMeasurements
in a Lified Turbulent-Jet Flame ”, Twenty-third Symposium (International) on Combustion
(Pittsburgh: The Combustion Institute), pp 669-676.
2. Nathan, G.J.(1988), “The Enhanced Miring Burner”, PhD thesis, Department of Mechanical
Engineering, The University of Adelaide.
4. Nathan, G.J. and Luxton, R.E. and Hill, S.J. (1997), “AnAxi-Symmetric %luidic‘Nozzle to Generate
Jet Precessionfor Enhanced h r g e Scale Miring”, in preparation for J. Fluid Mech.
5 . Nathan, G.J., Turns, S. R. and Bandaru, R. V., (1996), “TheInfluence of Fuel Jet Precession on the
Global Properties and Emissions 01Unconjined Turbulent Flames”, Combust. Sci. and Tech. Vol.
122, pp 21 1-230.
6. Nathan, G.J.,Luxton, R.E. and Smart, J.P. (1992), “Reduced Nox Emissions and Enhanced Lurge
Scale Turbulence from a Precessing Jet Burner ”, Twenty-fourth Syposium (International) on
Combustion (The Combustion Institute), pp 1399-1405.
7. Nathan, G.J.and Luxton, R.E. (1992), “A Low NO Gas Burner with a Radiant Flame”, Energy
Efficiency in Process Technology. Commission of the European Communities, (Oct.), Vouliagmeni,
8. Manias, C.G. and Nathan, G.J. (1994), “LowNO Clinker Production”,World Cement, (May).
9. Schneider, G.M., Froud, D., Syred, N., Nathan, G.J.and Luxton, R.E. (1997) , “Veloci@
Measurements in a Precessing Jet Flow using a Three Dimensional LDA System”, Experiments in
Fluids. Vol. 23, pp 89-98.
10. Eckbreth, A.C. (1988), “ h e r Diagnostics for Combustion Temperature and Species” , Abacus
Press, U.K., pp 326-332.
1 1 . Dieke, G.H. and Crosswhite, H. M. (1961), “TheUftravioletBands of OH Fundonrental Data”, J.
Quant. Spectrosc. Radiat. Transfer., Vol2, pp 97-199.
12. Alwahabi, Z.T., King, K.D.,Nathan, G.J., N o h , D.S., Newbold, G.J.R. and Luxton, R.E. (1997),
“Planarher-induced Fluorescence Studies of Radicals in Precessing Jet Flames”, CHEMECA 97,
Roto~ii,New Zealand.
13. Warnatz, J., Maas, U., and Dibble, R.W. (1996), “Combustion - Physical and Chemical
Fundamentals, Modeling und Simuhtion, Experiments, Pollutant Formation”, Springer-Verlag,
Berlin, Germany, p 158.
14. Schefer, R.W., Nmazian, M.and Kelly, J. (1994), “TemporalEvolution of Tiubulence / Chemistry
Interactions in Lified, TurbulentJet Flames ”, Twenty-fifth Symposium (International) on
Combustion (Pittsburgh: The Combustion Institute), pp 1223-123 1.
15. Nobes, D.S., Nathan, G.J.,
Luxton, R.E., Alwahabi, Z.T. and King, K.D. (1996), “PhaseAverage
Planar Imaging of Concentrationof u Precessing Jet Flow ’*, First Australasian Conference on Laser
Diagnostics in Fluid Mechanics and Combustion, A.R. Masri and D.R. Honnery, Eds., pp 154-159.
16. Wamatz J., Maas U. and Dibble R.W. (1996), ”Combustion. Physrcul and Chemical Fundamentals,
Modeling and Simulation,Experiments, Pollutant Formation ”. Springer, pp 187- 188.
17. Muniz, L. and Mungal, M.G.(1997), Instantaneous Flame-Stabilization Velocities in Lified-Jet
DiBsion Flames”,Combustion and Flame. Vol 1 1 1, pp 16-31.
Без категории
Размер файла
1 367 Кб
flames, jet, effect, precession, regions, ignition
Пожаловаться на содержимое документа