Characterisation of the Occurrence of the Precessing Vortex Core in Partially
Premixed and Non-Premixed Swirling Flow
By
Syred, N.*, Wong, C. +, Rodriquez-Martinez, V.*, Dawson, J.*, Kelso, R.+
* School of Engineering, Cardiff University, Wales, UK
+
School of Mechanical Engineering, University of Adelaide, Adelaide SA5005,
Australia
Abstract
This paper will present new information on the occurrence and suppression of the precessing vortex
core and its influence on the combustion process and, in particular, the stimulation of acoustic
instability in closed systems. A swirl burner which has been extensively characterised over many years
is operated in a premixed/partially-premixed mode of operation. Previously the occurrence of the
precessing vortex core (PVC) has been demonstrated under both isothermal and combustion conditions
using a number of techniques. Detailed phase-locked velocity measurements have been used to
characterise the associated flow patterns. The work has been recently extended to include fundamental
studies of enclosed oscillating swirl flow combustion systems under acoustic excitation. Here the role
of the coupling via the Rayleigh criterion of acoustic excitation and periodic heat release is clear, but
the role of the central recirculation zone and external recirculation zone is not clear in the instability
process. Recent work has shown that reducing or eliminating the size of an external recirculation zone
can substantially reduce the amplitude of oscillation in the burner. Phase-locked velocity measurements
have not given the fine details of the flow in and around the central recirculation zone and its influence
on system instability.
In this work we also examine the flow structure in the same design of swirl burner, fired into an open
environment by examining the structure of the flame and its reaction using high-speed video for
visualisation. The flame is seeded with sodium chloride (NaCl) that dissociates into sodium and
chloride radicals at high temperatures. The fluorescence of the sodium in the high temperature region
enables the visualisation of the temperature gradient and fluctuating temperatures in the flame to be
made.
The phase-locked data show that a major effect of a regular oscillation is to produce cyclic changes in
the size and shape of the central recirculation zone corresponding to changes in swirl number in the
system. The precessing vortex core thus follows the changing shape of the recirculation zone probably
via changes in precessional radius and the number of vortices present. Associated with this are changes
in the position and intensity of the reaction zone which can provide feedback to the system via the
Rayleigh criteria in a self-sustaining oscillation. Another important feature observed in the experiments
is the wobble of the swirling flow indicated by high levels of tangential directional intermittency both
on the centre line of the flow and near to the outer wall. A quarl was found to reduce this effect for high
frequency oscillations.
The open flame measurements, taken for flows near the weak extinction limit with a very much
damped precessing vortex core, showed time-dependent intermittency on the outer edges, especially in
the first 1 to 1.5 burner exit diameters. Examination of individual flame images separated by 1
millisecond frames shows a central reaction zone with a projected area fluctuating between 20% to
30% compared with the mean projected area. A significant irregular wobble of the flame was detected
close to the burner exit. This coupled with the irregularity in the reaction zone leads to the conclusion
that the central reverse flow zone is fluctuating very significantly in size and is capable of locking onto
major acoustic frequencies of a system.
It is concluded that stability can be enhanced by reducing the wobble of swirl flames by introducing
quarls instead of sudden expansions at burner exhausts and using higher levels of swirl to produce
more uniform and regular central reverse flow zones.
Introduction
Lean pre-mixed combustion (LPC) has become the standard approach engineers employ to reduce
pollutant emissions, particularly NO x emissions. Some of the other main advantages lean pre-mixed
operations are increased combustion intensity, shorter flame lengths, better fuel burnout and lower
emissions. However these gains can be accompanied by stability problems. Premixed flames are, by
nature, more susceptible to static and dynamic instabilities due to the lack of inherent damping
mechanisms. The absence of diffusive mixing times leaves flames sensitive to acoustic excitation from
sound waves with flame response dependent upon the amplitude, frequency and nature of acoustic
wave impingement. If conditions are favourable, periodic fluctuations in the heat release will match
the natural resonant frequency of one or more of the geometrical components of the combustor
resulting in self-excited thermo -acoustic instabilities. The mechanism responsible for the maintenance
of limit-cycle heat-driven oscillations was originally proposed by Rayleigh 1 and refers to the
relationship between the pressure wave and rate of heat release. For these reasons, thermo -acoustic
instabilities in modern systems that use LPC conditions, such as gas turbine combustors, are of
increasing concern to combustion engineers and these predominantly use swirl stabilisation. A number
of studies have been carried out to identify the various effects responsible for governing combustion
oscillations. Keller et al.2, 3 showed that the limit cycle can be divided into inter-related characteristic
time-lags representative of combustion chemistry, acoustic period and flow dynamics. Experiments by
Richards and Janus4 , Straub and Richards5 and Straub et al.6 showed the effect of various inlet
parameters on the oscillations in the reactant supply/premixer in generating a fuel time -lag 7,
Recent studies by, Paschereit et al 9 , Broda et al 10 , Lee et al 11 , Venkataraman et al 12 , Pun et al 13 ,
Giezendanner et al 14 and Allgood et al 15 have examined the effect of instability on the flame
dynamics. These studies of swirl-stabilised flames revealed the fluctuations in the local reaction rates
as well as the various transport processes between the reactants, flame zones and products over the
cycle. Despite the absence of velocity information, these investigations into flame dynamics over limit
cycle oscillations via chemiluminescence measurements show strong indications of flame response to
the fluctuating structure of the recirculation zones in the combustor 9-15 which are critical to the flame
stabilization process. As such, the effect of pressure and heat release oscillations on the swirl flow
velocity fields are of significant importance. Depending on combustor design and the flow patterns
generated, the actual flame response to pressure oscillation may vary greatly.
Some studies have used the Rayleigh Index (RI) to identify the driving and damping zones in relation
to spatially -resolved or global chemiluminescence 10-13 . Studies have shown that the damping regions
are normally located in the proximity of the burner exit and at the upstream near-wall positions whilst
driving regions are located near to or just downstream of the normal stabilization flame zones. Flame
visualisation studies 10 show that under instability the swirl-stabilized flame shapes, which are usually
conical at the root of the flame and tulip-shaped further downstream, flatten and move upstream
towards the burner. As a result, combustion instability encourages the flame to interact with the
upstream flow-field, in particular the external (ERZ) and central recirculation zones (CRZ) and
coherent structures such as the well known precessing vortex core (PVC). Regular occurrence of
combustion oscillations in practical systems has led researchers to develop active control strategies
ranging from closed-loop active acoustic control of flow parameters (Yu et al 16 and Docquier et al 17 )
to passive control techniques such as variations to swirl vane or fuel injection port location 6,5 or the
modification of the acoustic characteristics of the system (Sivasegaram and Whitelaw 18 ), i.e.
Helmholtz resonators attached to the combustor geometry, orifice plates in the inlet at strategic
locations, or substantive modifications to the ERZ 27 .
This paper concentrates on the flow dynamics and presence/suppression of the PVC in the flow,
(Syred et al 19, 23, 25 and Froude el al 20 ) and explores its role in the excitation of instability. The actual
mechanism of the coupling is still unclear, but appears to arise from flow instability feeding into
unsteady heat release/combustion processes which feed instability via coupling with acoustic modes of
oscillation via the Rayleigh Criterion. Associated work has shown that cyclone dust separators can
couple with large ductwork systems in large high pressure, process plant, to produce low frequency,
large amplitude oscillations, the driving mechanism arising directly from the PVC generated in the
cyclone dust separator (Yazabadi et al 21,22 ). Without combustion the PVC and associated flow
instabilities 23 can couple with acoustic modes of oscillation, providing substantial excitation under the
appropriate conditions. It is often difficult to analyse the role of the PVC and indeed its presence in
combustion systems as this is a function of swirl number (S) 23 ( normally S>0.6 to 0.7 for vortex
breakdown and the PVC to appear), as well as the mode of fuel entry, or more appropriately, as this
paper will indicate, the radial position of the flame. It has been clearly shown in the past that axial fuel
entry normally damps the PVC amplitude substantially, whilst premixed combustion restores its
presence and indeed can considerably excite it. This is, of course, extremely important with the trend
to premixed and partially premixed combustors. This paper thus starts with a review of relevant work
in this area and then uses recent and new data to analyse the role of the PVC relative to other factors
which influence instability in swirl combustion systems.
Instabilities in Swirl Flow systems and influence of the PVC
Although there are considerable data available now on the variation of PVC frequency with Strouhal
Number or non-dimensionalised frequency number under isothermal conditions, 20,24,25 data are much
more limited on the combustion state 23 apart from that of Gupta el al 23 and Claypole 26 . Claypole used
a swirl burner with 4 circular inlets, whose area could be blocked off with inserts to vary the swirl
number, and axial fuel entry along the centre line with a number of different injectors. The exhaust
was a simple straight-sided nozzle.
For an equivalence ratio of 0.887 with the combustor fired on natural gas, the PVC was found to exist
for all swirl numbers characterised from 0.63 to 3.04. In the swirl number range 0.9 to 1.7, dependent
on flowrate, a double PVC was clearly discernable, although a second harmonic could always be
detected in the spectrum. In terms of the frequency parameter, the Strouhal number values of
St=fDe 3 /Q ranged from ~1 (S=0.63) to 0.41 (S=1.26), 0.37 (S=1.53), with the value for S=0.9 showing
a sudden jump as a double PVC mode dominated the flow with values of St dropping from 1.1 to 0.25
as the mode switches from single to double PVC. This behaviour is quite different from the isothermal
case 26 where there is a steady increase of St with swirl number 23,26. It was also shown that
combustion in this mode reduces the probe-measured pressure fluctuations from the PVC by a factor of
15 at S=1.98. However, at the PVC peak frequency the PVC still contributed about 60% to the rms
value (combustion) as opposed to nearly 90% (isothermal) as the level of pressure fluctuations was
substantially reduced by combustion. A radial fuel injector attenuated the coherence of the PVC
somewhat.
Recently Rodriquez-Martinez 27-29 has produced very detailed phase-locked velocity measurements on
a natural gas fired swirl burner/furnace system, with 60% of the fuel being premixed with the air and
the rest being introduced axially on the centreline. The swirl burner (S=2.18) was of similar
configuration to that of Claypole 26 . The furnace system is twice the exhaust diameter of the burner
and two very different modes of oscillation,~ 40/60 Hz,~240/260Hz, could be produced, being a
function of the furnace geometry, inlet and outlet (Fig 1). By introducing a conical quarl for the high
frequency oscillation case, the external recirculation zone is nearly eliminated. The insertion of the
quarl essentially stabilises the central recirculating zone and eliminates an outer region where
tangential velocities are low and where weak counter-rotating vortices (possible PVCs) are found (Fig
2).
The high frequency oscillations are attributed to near in-phase coupling of a natural Helmholtz
resonance of the swirl burner and furnace, with the exhaust of the swirl burner acting as the neck of the
resonator and periodic heat release via the mechanisms discussed in the introduction above 27 .
Conversely the low frequency oscillations (Fig 2) correspond to acoustic excitation via longitudinal
waves induced in the inlet pipes and could involve gross flow reversal and flow blockage in the swirl
burner for certain configurations tested 28,29 . Fig 2 shows the effect of equivalence ratio on these two
configurations with and without the quarl. Highest amplitudes of oscillation are obtained at
equivalence ratios between 0.6 and 0.9 for the high frequency 250 Hz oscillation. Sudden jumps in
oscillation frequencies occur at equivalence ratios of about 0.56 and 0.9 to the lower frequency
oscillation ~ 50 Hz. The effect of the quarl on the oscillation amplitude is only apparent for the high
frequency oscillations, with the amplitude of the low frequency oscillations being much lower for both
cases. However LPC systems often operate very weak, beyond an equivalence ratio of 0.6, thus putting
them beyond the range of the high frequency oscillation with this configuration, and thus the role of the
external recirculation zone in stimulating or suppressing oscillations is questionable at low equivalence
ratios.
Detailed phase-locked velocity measurements for both a high and a low frequency cases (see Fig.1.) 2729
show areas of high tangential directional intermittency (?) always on the axis and close to the outer
walls of the combustor. This is illustrated in Fig 3 for the high frequency case and in Fig 1, in terms of
phase angle over an oscillation cycle which is of near sinusoidal form at high frequency. Similar
results for ? are found for the low frequency case. The results are taken just above the exhaust nozzle
of the swirl burner in the furnace in Fig.1. with the position of the nozzle being illustrated in the bottom
of that figure. The oscillation frequency is, of course, dominated by the acoustics or bulk-mode
oscillation and thus the phase-locking does not pick up the PVC frequency, which will fluctuate across
the limit cycle of the oscillation anyway. This occurs because of changes in volume flow-rate through
the system caused by the limit cycle oscillation and combustion which primarily affect the axial flow
(Fig. 3), causing substantial changes in swirl number. This can be clearly seen from Fig 3 where the
central recirculation zone has just lifted out of the swirl burner exhaust at a phase angle, f, of 240o ,
whilst at 0o it is plainly established over a large cross-section of the burner exhaust. Other work 23
show that it will extend to the back plate of the burner. It is estimated that the swirl number is varying
from ~ 0.8 (close to vortex breakdown) at a phase angle of 240o to probably 4 at a phase angle of 0o .
Thus working from non-dimensional frequency parameters of 24, 26 it is possible to obtain PVC
frequencies ranging fro m 18Hz at f~240o to 83 Hz at f~0o . This gives rise to the high levels of
tangential directional intermittency shown on the centreline and close to the outer wall as the flow is
effectively wobbling about the central axis. In practice this means that the precessional radius of the
vortex core fluctuates over the limit cycle following the boundary of the CRZ and indeed there are
probably jumps between one PVC and two PVC states. The introduction of a quarl, besides virtually
eliminating the ERZ, also substantially reduces the levels of tangential directional intermittency both
on the central axis and by the walls, 27-29 , thus reducing the wobble of the flow and the flame, whilst
also constraining the motion of any PVCs.
Separate work on an identical burner firing in the open environment has shed some light on this
phenomenon (Fig 4). Here the burner is operated with axial fuel entry alone at f=0.456, well into the
LPC region, whilst the PVC amplitude is substantially damped so little residual trace can be found. A
salt solution (sodium chloride) is injected with the fuel and volatilises early in the high temperature
regions inside the burner so that the high temperature regions of the flame are well visualised by the
fluorescence of the sodium. A phantom high-speed camera with an exposure of ~0.1 millisecond
framing at 1KHz was used to obtain the natural luminescence of the flame. Successive images are
subtracted from each other to give a measure of the change of intensity especially towards the edge of
the flame. These subtracted images are then analysed to give a mean, rms and local turbulence
intensity values. Both a side and top views are obtained via the use of a stainless steel mirror. Figure 4
gives information on the fluctuation of flame intensity for frequencies up to 1 kHz for a 1 second time
frame. Since the storage limit of the camera was 1000 frames and based on LDA experience probably
10,000 images are needed to obtain better statistics. In particular, the results show that the edge of the
flame is highly intermittent with instantaneous fluctuations up to 6 times the mean and occurring within
about 1 to 1.5 burner exit diameters. Circumferentially there is also very large non-uniformities as
shown by the top view, and clearly both tangential and radial mixing appear to dominate on the edge
and top of the flame. The data can be further analysed as shown in Fig 5. Here four successive flame
images, each being separated by 1 millisecond were analysed in terms of the flame shape demarcated
by the 5% and 95% areas of maximum intensity on each image. Analysis of the behaviour of
volatilised sodium in flames indicates that the flame boundary corresponds to the outer green contour
and a temperature around 750o C. As to be expected, the downstream flame shows substantial variation
in shape but, most interestingly, the flame just leaving the burner exit shows considerable variation in
its diametric location. It also appears that the flame is physically wondering or precessing in which no
regular frequency could be detected. There are also indications of this in Fig. 4. The top view of the
flame shows that it is non-circular in shape and largely distorted. Considerable strain and distortion
thus exists on the edge of the flame with flame front displacements of 10 mm and more in 1
millisecond being common.
Examination of the cine film and still images shows that the flame is sensitive to small external
disturbances and is easily disturbed by flow or acoustic perturbations, especially downstream of the
burner exit. Clearly the presence of a quarl which guides the expansion of the flow can also serve to
damp significant eddy movements on the outside of the flame as it expands past the burner exit, but
however it does nothing to damp the irregular circumferential movements shown in the top views of
Figs 4 and 5, or the irregular end-section of the flame. The 95% intensity contour also indicates that
the boundary of the flame and central reaction zone are also varying considerably as shown in Fig. 5.
This probably also correspond to an irregular fluctuation in the size and shape of the CRZ and its
associated shear layer, akin to that shown in Fig. 3, but not with a regular motion.
Fig 6 then shows an analysis of 3 separate successive side-view flame images where the outer
boundary has been outlined. Here, it is clear that, between image 783 and 784 there has been massive
movement within 1 millisecond, corresponding to mean flame front velocities of up to 40 m/s in some
places but more commonly 20 m/s. This implies that strain rates of up to 1000 sec-1 at the top and end
of the flame may exist, although further work is needed to verify this . In terms of stability the radial
movement of the whole flame close to the burner exit is evident between image 783 and 784, with a
radial displacement in 1 millisecond of 5.2 mm, or 10% of the flame diameter at this point. This gives
a possible radial strain of up to 1000 sec-1 again, but obviously the rotation of the flame is very
important here and the effect is probably attributed to the gross movement of the whole flame. Other
work 30 has shown that in the presence of an enclosure, a small level of premixing can well induce the
presence of a strong PVC. This is so even for the case where the PVC is strongly damped without an
enclosure. The mechanism of instability and coupling appears to be caused by irregularities in the
flame boundaries and/or reaction surfaces/areas, primarily associated with distortions of the CRZ.
Associated with this CRZ distortion is the production of a PVC whose radius of precession is governed
by motion and distortion of the CRZ, and the actual instantaneous level of swirl at a given instant.
Discussion and Conclusions
This paper has reviewed recent work on oscillations in swirl burner systems, used new data on open
swirling flames, and related them to the occurrence of instability in LPC gas turbine systems. The PVC
and the occurrence of the central recirculation zone are explicitly linked in that the PVC rotates around
the boundary of the CRZ, interacts with the formation of the CRZ, whilst also exciting secondary flow
like in other planes. The phase-locked velocity characterisation of a stable oscillation swirl
burner/furnace system has provided much insight into the flow regimes and patterns which contribute
to the excitation of oscillation via the Rayleigh criterion and periodic heat release in phase with the
pressure node of an acoustic mode of the system. The measurement region is concentrated in and just
past the burner exit as it fires into the furnace. For a high frequency, acoustically driven oscillation,
the CRZ size and shape vary considerably over the limit cycle, showing characteristics of a very high
swirl flow at one phase angle, changing to a much weaker, low swirl CRZ some 200o later in the cycle.
Clearly associated with this are variations in the position of the reaction zone, which surrounds the
CRZ in an annular sheath. For the low swirl, small CRZ , part of the limit cycle, the reaction zone
(now of reduced intensity) probably moves downstream into the furnace, providing the mechanism for
the in-phase heat release to sustain the oscillation. The role of the PVC is very difficult to separate
from that of the CRZ here, although the wobble of the whole flame is evident from the tangential
directional intermittency figures across the limit cycle; the reduction of this wobble via the introduction
of a quarl helps to reduce the amplitude of the high frequency oscillation. Complimentary work on a
similar burner fired into the open has provided further evidence as to the nature of the flow and the
reaction zone instability generation mechanism. Here the flame, operated weak, with a very much
damped PVC, has been seeded with sodium so as to fluoresce and enable high-speed video at 1 kHz to
be made of the flame structure. What becomes immediately apparent is that the edge of the flame is
quite irregular with instantaneous normalised fluctuations between images varying by up to a factor of
6. Compared with the mean intensity levels the central reaction zone fluctuates considerably in shape
between successive 1 millisecond images, initial estimates are of between 30% to 40% in area.
Characterisation of the edge of the flame shows very large fluctuations at the top of the flame, with
significant fluctuations of up to 10% of the diameter as the flame leaves the burner exit. This bears out
the evidence of the phase-locked measurements with an acoustically excited regular oscillation and
indicates that these flames are very susceptible to small perturbations and naturally occurring wobbles
in their whole structure, whether induced by the PVC or not.
The effect of a quarl in reducing high frequency acoustic oscillations has been demonstrated, but
further work should confirm the following:
• higher swirl levels should be used to produce more regular and stronger CRZs that are less
susceptible to deformation by pressure fluctuations. This will generate stronger PVCs, but
providing these are well controlled and regular should not cause problems unless there is a
fundamental mismatch between major acoustic modes;
• Control of the wobble of the central flow and flame appears to be important in reducing the
regular and irregular precession of the flow and flame, and a quarl can be helpful here for high
frequency oscillations
Acknowledgements
Professor N. Syred gratefully acknowledges the Royal Academy of Engineering award of a Global
Research Award, also the facilities provided by the School of Mechanical Engineering, Adelaide
University during his sabbatical leave. Appreciation is also given to Dr. Gus Nathan and Mr. A.
Langley for their assistance in the high-speed camera experiments.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Rayleigh, The Explanation of Certain Acoustical Phenomena, Nature, 188, pp.319, 1878.
Keller J.O., Bramlette T.T., Dec J.E. and Westbrook C.K., Pulse Combustion: The Importance of
Characteristic Times, Com0bust. and Flame, 75, 33, 1989.
Keller J.O., Bramlette T.T., Westbrook C.K. and Dec J.E., Pulse Combustion: The Quantification
of Characteristic Times, Combust. Flame, 79, 151, 1990.
Richards G.A. and Janus M.C., Oscillations During Premix Gas Turbine Combustion, ASME
Paper No. 97-GT-244, 1997.
Straub D. L. and Richards G. A., Effect of Fuel Nozzle Configuration on Premix Combustion
Dynamics, ASME Paper No. 98-GT-492, 1998.
Straub D. L., Richards G. A., Yip M. J., Rodgers W.A. and Robey E.H., Importance of Axial Swirl
Vane Location on Combustion Dynamics for Lean Premix Fuel Injectors, AIAA Joint Propulsion
Conference, July 13-15, Paper no. AIAA-98-3909, 1998.
Lieuwen, T. and Zinn, B.T., The Role of Equivalence Ratio Oscillations in Driving Combustion
Instabilities in Low NOx Gas Turbines, Proc. Comb. Inst. 27, pp. 1809-1816, 1998.
Mongia, R., Dibble, R. and Lovett, J., Measurement of Air-Fuel Ratio Fluctuations Caused by
Combustor Driven Oscillations, Proc. of the ASME/IGTI Turbo Expo Conference, ASME paper
No. 98-GT-304, 1998.
Paschereit, C.O., Gutmark, E., and Weisenstein, W., Excitation of thermoacoustic instabilities by
interaction of acoustics and unstable swirling flow, AIAA Journal, v 38, n 6, p 1025-1034, 2000.
Broda, J.C., Seo, S., Santoro, R.J., Shirhattikar, G. and Yang, V., An Experimental Study of
Combustion Dynamics of a Premixed Swirl Injector, Proc. Comb. Inst. 27,1849, 1998.
Lee, S.-Y., Seo, S., Broda J.C., Pal, S. and Santoro R.J An Experimental Estimation of Mean
Reaction Rate and Flame Structure During Combustion Instability in Lean Premixed Gas Turbine
Combustor, Proc., Comb. Inst. 28, 775, 2000.
Venkataraman, K.K., Preston, L.H., Simons, D.W., Lee, B.J., Lee, J.G. and Santavicca, D.A,
Mechansim of Combustion Instability in a Lean Premixed Dump Combustor, J. Prop. Power,15, 6,
909, 1999.
Pun, W., Palm, S.L. and 0Culick, F.E.C., Combustion Dynamics of an Acoustically Forced Flame,
Combust. Sci. and Tech., 175:499-521, 2003.
Giezendanner, R., Keck, O., Weigand, P., Meier, W., Meier, U., Stricker, W. and Aigner, M.,
0Periodic Combustion Instabilities in a Swirl Burner Studied by Phase-Locked Planar LaserInduced Fluorescence, Combust. Sci. and Tech., 175, pp. 721-741, 2003.
Allgood, D., Murugappan, S., Acharya, S. and Gutmark, E., Infrared Measurements of
Thermoacoustic Instabilities in a Swirl-Stabilized Combustor, Combust. Sci. and Tech., 175:333355, 2003.
Yu, K.H., Wilson, K.J. and Schadow, K.C., Liquid -Fueled Active Instability Suppression, Proc.
Comb. Inst. 27, pp. 2039-2046, 1998.
Docquier, N., Lacas, F. and Candel, S., Closed-Loop Equivalence Ratio Control of Premixed
Combustors Using Spectrally Resolved Chemiluminescence Measurements, Proc. Comb. Inst., 29,
pp.139-145, 2002.
Sivasegram, S. and Whitelaw, J.H., Oscillations in Axisymmetric Dump Combustors, Combust.
Sci. and Tech., Vol 52, pp. 413-426, 1987.
Syred, N., Hanby, V.I. Resonant Instabilities Generated in Swirl Burners. J. Inst. Fuel, 46, ,pp 402407, 1973
Froud, D, O'Doherty, T and Syred N - Phase averaging of the precessing vortex core in a swirl
burner under piloted and premixed combustion conditions, Combustion and Flame , 100, pp 407412, 1995.
P A Yazdabadi, A J Griffiths and N Syred -Characteristics of the PVC phenomena in the exhaust
of a cyclone dust separator, Experiments in Fluids, 17, 84-95, 1994
22. P A Yazdabadi, A J Griffiths and N Syred -Investigations into precessing vortex core phenomenon
in cyclone dust separators, IMechE Proc E, J of Process Engineering, 208, pp 147-154, 1994
23. Gupta, A.K., Lilley, D.J., Syred, N., Swirl Flows, Abacus Press, Tunbridge Wells, 1984
24. O’Doherty, T., Froud, D., Bates, C.J., Syred, N., Characterisaticsa of a Power Station Boiler, Proc.
Institution Mech. Eng., Part A: Journal of Power and Energy, 208, 89-36, 1994
25. Syred, N., O’Doherty, T., Froud, D., The Interaction of the Precessing Vortex Core and reverse
Flow Zone in the Exhaust of a Swirl Burner, Proc. Institution Mech. Eng., Part A: Journal of
Power and Energy, 208, 27-36, 1994
26. Claypole, T.C., PhD thesis, University of Wales, 1980
27. Rodriguez-Martinez, V.M., Dawson, J.R. Syred, N. and O'Doherty, T., 'The Effect of Expansion
Plane Geometry on Fluid Dynamics Under Combustion Instability in a Swirl Combustor', 42nd
Aerospace Science Meeting and Exhibit, No. AIAA paper 2003-0116, Reno, Nevada, January 6 9, 2003
28. Dawson J.R., Rodriquez-Martinez, V.M., Syred, N., O’Doherty, T., ‘Low Frequency Combustion
Oscillations in a Swirl Burner Furnace’42nd Aerospace Sciences Meeting and Exhibit Conference,
AIAA, 5-9th January 2004, Nevada, USA
29. Rodriquez-Martinez, V.M., PhD thesis, Cardiff University, 2003
30. Fick, W., PhD thesis, Cardiff University, 1968
Premixer
Inlet
nozzle
Combustor
Removeable
expansion plane
(35°)
air
Two tangential inlet
swirl generator
Pilot gas
air
L nozz
Dnozz
Qair (l/min)
Qgas (l/min)
φ
f (Hz)
125(9400) 2.63(200)
3.02(230)
0.67(50)
2100
190
0.9
41
81(6100)
1.47(110)
1.00(75)
2600
210
0.81
243
Case
Linlet
1
2
Lfurn
3.16(240)
Normalized dimensions of swirl burner/furnace test rig in terms of exit burner diameter, De = 75 mm.
Actual dimensions in mm in brackets. The dimensions and operating conditions for case 1 and case 2 refer
to conditions which maximise the regularity of the oscillation and allow phase locked velocity
measurements to be made and interpreted. Both low and high frequency oscillations could be produced
with both configurations. The quarl insert around the burner nozzle could be removed to produce case 1
Fig 1. Schematic of Swirl Burner/Furnace System
Frequency [Hz]
300
250
200
150
100
50
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
equivalence ratio, φ
Case 2b
Case 2a
0.8
pressure rms [kPa]
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
F frequency response to variations in equivalence ratio for Sudden
Fig. 2. Values of rms pressure p´rms and
Expansion, Case 2a and the Quarl divergent expansion, Case 2b
0°
60 °
120 °
180 °
240 °
300 °
1.5
U/U i, W/U i
x/De
a)
1
0.5
1.5
x/De
b)
1
0.5
1.5
x/D e
c)
8
100
7
90
6
80
5
70
4
60
3
50
2
40
1
30
0
20
-1
10
-2
1
-3
0.5
-4
1.5
d)
x/De
γ (%)
1
0.5
0
0
0.5
r/De
10
0.5
r/De
10
0.5
r/De
10
0.5
r/De
10
0.5
r/De
10
0.5
1
r/De
Figure 3 Phase-averaged contour plots at 6 phase angles for Sudden Expansion Configuration
a) axial velocities,
b) directional intermittence (γ) of
axial velocities
c) tangential velocities
d) γ of tangential velocities.
γ is the phase-averaged percentage
of negative samples, directional
intermittency. contour cutoff 3%.
Results with quarl for ? tangential
velocities lower, but negative
values still occurring in similar
positions
Fig 4 . Non-Dimensionalised Intensity Variations from a Free Natural Gas Swirling
Flame, Simultaneous Horizontal and Vertical Views. Equivalence Ratio 0.36
Intensity variations obtained by subtracting successive images obtained from camera with
1millisecond interval (exposure ~0.1milisecond)
Fig 5.
Top four images show instantaneous
successive Images from Sodium
Seeded Flame, Contours show 5%
and 95% Intensity Levels (Images
move sequentially from left to right,
top to bottom. Bottom left figure
shows normalised mean intensity
from 1000 images. Equivalence
Ratio 0.33
Flame Boundary Images 783, 784, 785
160
140
120
axial distance mm
100
80
Series1
Series2
Series3
60
40
20
0
-80
-60
-40
-20
0
20
40
60
-20
radius mm
Key- series 1- image 783, Series 2 – image 784, Series 3 – image 785
Fig 6. Instantaneous Flame Boundaries as measured from 3 successive
Flame Images at 1 kHz, Equivalence Ratio 0.31
80