C. R.

Mecanique 340 (2012) 758–768

Contents lists available at SciVerse ScienceDirect

Comptes Rendus Mecanique

www.sciencedirect.com

Out of Equilibrium Dynamics

Progress and challenges in swirling flame dynamics

Sébastien Candel a,b,∗ , Daniel Durox a,b , Thierry Schuller a,b , Paul Palies a,b ,
Jean-François Bourgouin a,b,c , Jonas P. Moeck d
a
CNRS, UPR 288, Laboratoire d’énergétique moléculaire et macroscopique combustion (EM2C), 92295 Châtenay-Malabry, France
b
Ecole Centrale Paris, 92295 Châtenay-Malabry, France
c
SNECMA (Safran Group), Centre de Villaroche, 77550 Moissy-Cramayel, France
d
Institut für Strömungsmechanik und Technische Akustik, Technische Universität Berlin, 10623 Berlin, Germany

a r t i c l e i n f o a b s t r a c t

Article history: In many continuous combustion processes the flame is stabilized by swirling the injected
Available online 22 November 2012 flow. This is the case for example in aeroengine combustors or in gas turbines where
aerodynamic injectors impart a rotating component to the flow to create a central
Keywords:
recirculation zone which anchors the flame. Swirling flame dynamics is of technical interest
Combustion dynamics
Swirling flames
and also gives rise to interesting scientific issues. Some of the recent progress in this
Flame describing function field will be reviewed. It is first shown that the swirler response to incident acoustic
Swirl fluctuations perturbations generates a vorticity wave which is convected by the flow. A result of this
process is that the swirl number fluctuates. It is then shown that the flame response
is defined by a combination of heat release rate fluctuations induced by the incoming
acoustic and convective perturbations. This is confirmed by experimental measurements
and by large eddy simulations of the reactive flow. Measured flame describing functions
(FDFs) are then used to characterize the nonlinear response of swirling flames to incident
perturbations and determine the regimes of instability of a generic system comprising
an upstream manifold, an injector equipped with a swirler and a combustion chamber
confining the flame. The last part of this article is concerned with interactions of the
precessing vortex core (PVC) with incoming acoustic perturbations. The PVC is formed at
high swirl number and this hydrodynamic helical instability gives rise to some interesting
nonlinear interactions between the acoustic frequency, the PVC frequency and their
difference frequency.
© 2012 Académie des sciences. Published by Elsevier Masson SAS. All rights reserved.

1. Introduction

In the 1970s Clavin was well aware of the pioneering work of Zeldovich (see Zeldovich et al. [1]) and anticipated the
revolution which was to come from activation energy asymptotics (AEA). In his work he perceptively emphasized the mul-
tiscale and nonlinear nature of combustion problems where thin reactive layers propagate in a flow featuring a broad range
of spatial scales. This was exploited in analytical investigations of combustion waves by making use of asymptotics, a fea-
ture which has been extensively advocated in studies of laminar flames. AEA takes advantage of the fact that the activation
energy of kinetic steps controlling the combustion process is quite high, so that the ratio E /( R T u ) is a large parameter. AEA
has provided a modern view of laminar flame structures as exemplified in the work of Liñán [2], Clavin (see [3] and [4,5]
for reviews), Sivashinsky [6] and monographs by Buckmaster and Ludford [7] and Kapila [8]. Many examples are treated by
Williams [9], Law [10–12], Matalon [13,14] and in a collective review of this topic [15]. Activation energy asymptotics has

* Corresponding author at: CNRS, UPR 288, Laboratoire d’énergétique moléculaire et macroscopique combustion (EM2C), 92295 Châtenay-Malabry, France.
E-mail address: sebastien.candel@ecp.fr (S. Candel).

1631-0721/$ – see front matter © 2012 Académie des sciences. Published by Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.crme.2012.10.024
 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768 759

provided considerable information on the structure and properties of many types of flames and on fundamental combustion
processes like ignition and extinction, effects of strain and curvature on the laminar burning velocity, flame instabilities and
response to various types of flows or external perturbations. Williams underlines the remarkable progress made in combus-
tion theory during the twenty years separating the two editions of his book [9], a progress in which Clavin was one of the
major actors.
Much of his theoretical work and much of the experimental work carried out by his colleagues (most notably G. Searby
and L. Boyer) has concerned flame structures and their dynamics generally in fundamental aerodynamic configurations.
While the list of issues which have been tackled is long it does not include questions raised by swirling flames which are
considered in what follows from a dynamical point of view. It is, however, worth noting that important issues may be
examined with tools derived from combustion theory. Swirling flames are by no means of minor interest. In jet engines
and gas turbines and in many other applications the flame is anchored by imparting an azimuthal component to the flow
usually by passing the air stream in a swirler or a set of swirlers. The rate of rotation defined by these devices generates
a central recirculation zone (CRZ) which is filled with hot combustion products serving to continuously ignite fresh reactants
introduced in the combustor. The swirling flow is confined by the lateral walls which in most cases has an annular cross
section and by the flows originating from lateral injectors. The presence of the side walls and of neighboring injectors
has a significant impact on the flow structure which complicates the analysis. A single injector placed in a sector or in
an axisymmetric configuration will only approximately represent the practical situation. It is, however, natural to examine
such configurations to identify the main dynamical features. It is also logical to examine annular geometries comprising
multiple injectors to examine possible coupling by azimuthal modes.
Much of the recent work on swirling flames has been carried out in relation with the design of advanced premixed
combustion technologies with the objective of reducing NOx emissions from gas turbines. These new systems have suc-
cessfully achieved low pollutant levels but their operation has been hindered by dynamical phenomena. Premixed flames
are more compact and more sensitive to external perturbations. Also damping in premixed systems is diminished because
the perforated liners found in classical designs are for the most part eliminated in modern combustors. For these reasons,
combustion dynamics has become a major issue in this field and it has been intensively investigated. This paper begins with
a brief review of the state of the art in combustion dynamics. It then focuses on selected issues in swirling flame dynamics:

• The first problem concerns the swirler response to incident acoustic waves. It is shown that the interaction generates
a vorticity wave characterized by azimuthal velocity fluctuations convected by the flow.
• The second issue is that of the response of the flame formed by a swirling injector. This may be described by a flame
transfer function (FTF) which defines the relative heat release rate fluctuations ( Q̇  / Q̇ ) as a function of incident relative
velocity fluctuations (u  /u):

Q̇  (ω)/ Q̇
F (ω) = (1)
u  /u
It is shown that the mode conversion process taking place at the swirler has a marked influence on the flame response.
It is also important to deal with the nonlinear response of the flame when it is submitted to a broad range of oscillation
amplitudes. The response changes with the input level and one may then use an extension of the linear concept of
transfer function and define a flame describing function (FDF) which depends on frequency and on the amplitude of
the input:

   Q̇  (ω, |u  |)/ Q̇
F ω , u   = (2)
u  /u
This nonlinear dependence of the flame with respect to the input level leads in many cases to the nonlinear features
observed in practice as already shown in a range of previous studies and as will be confirmed in what follows.
• The third issue is concerned with the analytical determination of the transfer function of swirling flames. It is shown
that this can be determined by making use of a level-set framework which describes how the flame is perturbed under
the combined effect of incident acoustic waves and azimuthal velocity fluctuations induced by the swirler.
• The fourth issue is linked to the use of the FDF to predict regimes of oscillation of a generic system. It is shown that by
combining an acoustic network representation with a description of the flame response in terms of the FDF one obtains
suitable estimates of frequencies and amplitudes of limit cycles observed experimentally.
• The final issue has to do with the precessing vortex core (PVC) a helical instability of rotating flows. Perturbations
induced by the PVC have a frequency which is of the order of the rate of rotation of the swirling flow. When this
flow is submitted to incident acoustic perturbations, or when the system develops a thermo-acoustic oscillation, the
PVC frequency combines with the acoustic perturbation to produce a component at the difference frequency which has
a characteristic “yin–yang” spatial structure and rotates at a rate equal to the difference frequency.

2. Background on combustion dynamics

Combustion dynamics raises difficult practical issues and constitutes a challenging area in combustion research. Under
normal operating conditions, turbulent flames generate heat release rate fluctuations which are essentially incoherent and
760 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768

Fig. 1. Resonant interactions between acoustics and combustion lead to combustion instabilities. The flame response is conveniently represented in terms
of a describing function F (ω, |u  |).

radiate noise over a broadband of frequencies. In these cases, the sound field within the combustor spreads over a low
frequency broad band spectrum without significant coherent feedback to the combustion process. Under unstable operation,
heat release rate fluctuations radiate noise coherently, a resonant loop is established between the flow, combustion and the
acoustic modes of the system [16–18]. This feedback synchronizes heat release rate and pressure perturbations giving rise
to large pressure fluctuation levels which may have detrimental consequences on the system operation. Such a resonant
loop is presented in Fig. 1 in a case where only velocity disturbances are present when the mixture composition remains
uniform.
Enhanced heat fluxes to the combustor walls and intense structural vibrations lead to mechanical failure and in extreme
cases to destruction of the system. Such dynamical phenomena are specifically damaging in high performance devices where
the power density is large, a situation prevailing in high pressure systems like those used in gas turbines, aeroengines
or liquid rocket thrust chambers. Many of the current issues are found in modern gas turbines which rely on premixed
combustion to reduce NOx emissions but are more sensitive to resonant acoustic coupling leading to instability (see [19] for
an extensive review).
Early work was carried out during the development of liquid rocket engines (LREs). The pioneering papers by Tsien [20],
Crocco [21,22] and Marble [23] indicate that combustion instabilities result from delays in the combustion process and that
this delay is sensitive to parameters governing the combustion process. This has been formalized in the sensitive time lag
model (STL) commonly used in LRE instability analysis [24] and in many other situations.
Progress made more recently has been derived from: (1) Detailed experimental investigations of the driving and coupling
mechanisms; (2) Development of analytical and reduced order models; and (3) Exploration of large eddy simulations for
dynamical combustion processes. This is the subject of many reviews [16,18,25]. Gas turbine combustion dynamics is specif-
ically considered in [26]. Much knowledge has been accumulated from investigations of perturbed flames. This has given
access to the driving interactions leading to heat release rate disturbances such as flame surface area wrinkling documented
in [27–33] and in a collection of articles edited by Lieuwen and Yang [26], interactions with large scale coherent structures
described in [30,31,34], effects of mixture composition perturbations examined in [33,35–37], interactions with boundaries
at the flame anchoring device considered in [27,32] or near the flame tip reported in [38]. Other studies have provided in-
formation on perturbations of the local mass burning rate by stretch [4,10,39–41], by mixture composition oscillations [42]
or in response to unsteady heat transfer when the flame lies close to a solid boundary [43,44].
Elementary mechanisms are often simultaneously present and interfere positively or destructively. When the flame is
compact with respect to the wavelength, unsteady combustion feeds acoustic energy in the system only if the global heat
release rate perturbation integrated over the combustion region is positive. An examination of the local mass burning flux or
flame motion gives an indication on the degree of unsteadiness in the system, but contributions may cancel when integrated
over the flame surface yielding a globally low level of heat release rate fluctuation. It was also demonstrated in the case of
wedge flames that wrinkles convected along the flame front and perturbations conveyed by the flow positively interfere at
low frequency leading to large flame surface area oscillations [30,45].
Modeling tools developed from these data have led to a unified framework for the determination of flame transfer
functions under simple flow conditions [45]. The flame response in combination with acoustic network descriptions has
allowed the stability analysis of a variety of systems. Dynamics of swirling flames and associated instabilities are considered
in many recent studies for their important practical value [46–57]. A recent review [19] lists more than 500 references.
Finally, it is worth noting that much recent progress has been made in this field with large scale applications of large
eddy simulations. This is illustrated in [58] where self-sustained azimuthal oscillations in an annular combustion chamber
are nicely retrieved.

3. Swirler response to incident acoustic waves

To analyze swirling flame dynamics, it is natural to first examine the interaction between the swirler and incident
acoustic perturbations. This can be done by considering that the swirler acts like a blade row which may be described
with an actuator disk theory (see Fig. 2). In this framework, the blade row is in turn replaced by a discontinuity and one
may then write jump conditions across this discontinuity to link upstream and downstream variables. This is a reasonable
approximation as long as the acoustic wavelength remains large compared to the swirler dimensions.
One finds these jump conditions in an article of Cumpsty and Marble [59] dealing with the conversion of entropy waves
impinging on a blade row into pressure waves. The finite Mach number relations can be specialized to the low Mach number
 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768 761

Fig. 2. Geometry of the problem. The swirler is represented by a blade row. This blade row is replaced by an actuator disk, a thin discontinuity separating
the upstream and downstream flows. An acoustic wave impinging on the swirler u 1 gives rise on the downstream side of the swirler to an acoustic wave u 2
and to a vorticity wave characterized by transverse velocity fluctuations v 2 . This last wave is convected by the flow.

Fig. 3. Velocity fluctuations on the downstream side of a blade row. The u 2 fluctuations correspond to the transmitted acoustic wave. The transverse velocity
fluctuations v 2 are linked to a vorticity wave generated by an incident acoustic wave interacting with the blade row. The amplitude of this wave is given
by | v 2 /u 2 | = tan θ 2 , where the trailing edge blade angle is here equal to θ2 = 25◦ . From [60].

case and used to deal with acoustic interactions. On the upstream side of the blade row, one can assume that the incident
acoustic wave is reflected as another acoustic wave propagating in the opposite direction:
     
 A x AR x
u1 = exp i ω −t − exp i ω − −t (3)
ρc c ρc c
where A, R, ω , ρ , c respectively designate the amplitude of the incident wave, the reflection coefficient, the angular fre-
quency, the density and sound speed. On the downstream side, the axial velocity fluctuation takes the form of an acoustic
wave, but it soon becomes apparent that one has to assume that a vorticity wave is generated on the downstream side of
the swirler and that this gives rise to transverse velocity fluctuations convected by the flow:
     
TA x x
u 2 = exp i ω −t , v 2 = B exp i ω −t (4)
ρc c u2
where T denotes the transmission coefficient, u 2 indicates the flow velocity behind the swirler and B is the amplitude of
the transverse velocity disturbance. By imposing the jump conditions specialized to low Mach number one finds that:
A
R = 0, T = 1, B= tan θ 2 (5)
ρc
where θ 2 designates the blade angle at the trailing edge measured with respect to the axial direction. According to this
theory, sound waves are fully transmitted by the blade row and a convective wave is produced at the trailing edge inducing
transverse velocity fluctuations v 2 having an amplitude which has the same order of magnitude as that of the incident
wave. This is an important component which will have a significant influence on the flame.
This mechanism demonstrated theoretically [56] has been confirmed by RANS [51] and direct [60] simulations of the
interaction between a blade row and incident acoustic waves (as illustrated in Fig. 3) and by experiments on a cold flow
traversing a swirler placed in a cylindrical channel [60].

4. Response of swirling flames

The response of swirling flames to incident acoustic perturbations is influenced by the mechanisms described in the
previous section. This is shown in the form of a block diagram in Fig. 4. Acoustic fluctuations induce azimuthal velocity
762 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768

Fig. 4. Block diagram representation of mechanisms generating heat release rate fluctuations in swirling flows.

Fig. 5. (a) Experimental configuration used to determine the swirling flame describing function. (b) Typical flame describing function. The bulk velocity in
the injection tube is U b = 2.87 m s−1 . Adapted from [61].

perturbations which give rise to swirl number perturbations. As a consequence two processes induce heat release fluctua-
tions, the first is through the direct effect of axial velocity fluctuations generating vortices which in turn roll-up the flame,
the second resulting from flame angle oscillations due to swirl number fluctuations. These perturbations interfere producing
large swirl number oscillations at the burner outlet for certain forcing frequencies and inducing a breathing motion and
flame angle fluctuations.
As a consequence the transfer function of swirling flames features local minimum and maximum gain values which
reflect the constructive or destructive interference of the previous mechanisms. It is found that conditions leading to large
swirling strength oscillations induce a weak flame response and conversely when swirl number fluctuations are absent
oscillations of the heat release rate reaches a maximum [56,57]. This may be verified experimentally by measuring the FDF
in the single injector configuration shown in Fig. 5.
The FDF displayed in Fig. 5(b) shows a local minimum and maximum gain values which are typical of the interference
process described previously. It is also found that the gain evolves with the amplitude while the phase only features minor
changes in this configuration. This is also confirmed [56] by examining the flame motion resulting from modulations by
incident acoustic waves at various frequencies (Fig. 6). At the frequency of 60 Hz, corresponding to a gain minimum, the
flame angle oscillates and counteracts perturbations induced by flame roll-up, while at f = 90 Hz the swirl number fluctu-
ations are low and the flame angle is essentially constant. Roll-up by vortices shed from the injector lip is maximum and
the disturbance in heat release rate is large yielding a maximum gain in the flame transfer function. Experimental data is
also retrieved by large eddy simulations [57].

5. Flame transfer function of swirling flames

The transfer function of swirling flames may be derived by making use of a level-set formalism in which the flame is
represented as a thin discontinuity. A level-set description of the flame motion is well suited to determine perturbations
induced by disturbances in the local mass burning flux and in velocity. A first order perturbation analysis of the so-called
G-equation [9] leads to the following equation for the perturbed G 1 (x, t ) field [62]:
 
∂ G1 S d1
+ v t0 · ∇ G 1 = v 1 − v 0 · n0 |∇ G 0 | (6)
∂t S d0

where v t0 = v 0 − ( v 0 · n0 )n0 is the mean flow velocity parallel to the mean flame front, S d0 denotes the mean flame
displacement velocity, and G 0 is the mean flame level-set. This transport equation for the perturbed field G 1 shows how
 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768 763

Fig. 6. Phase average images of heat release rate in a swirling flame submitted to acoustic modulations. Left: f = 60 Hz. Right: f = 90 Hz. Bulk injection
velocity U b = 2.67 m s−1 , equivalence ratio φ0 = 0.7. Adapted from [56].

disturbances in flow and in normal displacement velocity (v 1 and S d1 ) wrinkle the flame. Disturbances of the mass burning
flux ρ0 S d1 and velocity perturbations induce small perturbations of the flame position in the normal direction, which are
then convected along the flame front by the projection of the mean local flow velocity v t0 [45]. This generalizes a result from
Boyer and Quinard [63] derived for uniform velocity modulations to any flow non-uniformities affecting flame wrinkling.
The response of turbulent flames to flow disturbances can also be deduced from Eq. (6) by considering coherent wrinkles
associated to organized flow disturbances above the background turbulence level. Effects of incoherent turbulent fluctuations
are then encompassed in Eq. (6) by replacing the flame displacement speed S d by a turbulent flame speed S T . This is used
in [64] to determine the response of a turbulent premixed swirling flame submitted to acoustic modulations. It is shown
that the flame transfer function of swirling flames takes the general form:
  
Q̇  u S T
= F A (ω) − (7)
Q̇ ū S̄ T

where F A designates the flame transfer function of a flame sheet featuring the same shape and submitted to flow dis-
turbances convected by the mean flow [45] and S T / S̄ T denotes coherent perturbations of the turbulent flame speed. The
turbulent flame speed S T is essentially a function of the swirl number and it is natural to express its relative perturbation
in terms of this quantity to represent effects of swirl oscillations [64]. By suitably adjusting the coefficients of the assumed
linear relation one obtains a fairly good representation of the flame transfer function (Fig. 7).

6. The FDF framework and its application to a swirling flame configuration

Flame transfer functions are now commonly used to predict stability maps of various types of systems, by combining
descriptions of the acoustic field with the combustion response [65–67]. Alternatively acoustic transfer matrices [68] can be
used to model the chamber response offering a natural framework for experimentally determined flame transfer functions to
describe the combustion response to downstream and upstream disturbances [69]. Transfer functions were measured in gas
turbine combustors at atmospheric [70] and under high operating pressures for conditions prevailing in real burners [71]
and used to estimate the instability frequencies of a high pressure gas turbine equipped with an annular combustor.
764 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768

Fig. 7. Experimental and theoretical flame transfer functions. Bulk injection velocity U b = 2.67 m s−1 , equivalence ratio φ0 = 0.7. From [64].

Linear analyses provide a rough estimate of the system stability, but do not give access to many phenomena observed
in practice, such as frequency shifting during transient growth of oscillations, mode switching, instability triggering or
hysteresis and the limit cycle amplitudes cannot be inferred. To account for these observations and determine limit cycle
levels of oscillation one has to represent the nonlinear response of the flame. This was considered by Dowling [72] in
a theoretical analysis of a ducted V -flame response to incident disturbances by using two different expressions for the FTF
depending on the amplitude level of velocity fluctuations at the flame anchor point.
In general, the nonlinear dynamics is not easy to capture analytically. One possibility explored by Noiray et al. [73]
consists in combining experimentally determined flame describing functions (FDFs) with an acoustic network description
of the system. This is illustrated by an analysis of the stability margin of a combustor yielding the nonlinear dynamics of
the system. In this unified framework the flame is identified as the main nonlinear element in the system and the flame
response is represented by transfer functions, which depend on frequency and on the level of incident perturbations.
A family of amplitude dependent transfer functions forms the describing function: F (ω, |u  |). In the “describing” function
framework one assumes that the fundamental frequency is predominant and that the higher harmonics generated in the
nonlinear element are relatively weak because the nonlinearity is also weak or because the higher frequencies are filtered
out by the other components of the system. The stability is then characterized by a nonlinear dispersion relation function
of frequency and perturbation level:
  
D ω , u   = 0 (8)
Using this nonlinear description, it is possible to deduce growth rates depending on the incident disturbance ampli-
tude ωi (|u  |) and explain phenomena such as hysteresis, nonlinear instability triggering, and mode switching and accurately
retrieve experimental observations. The method originally developed for unconfined burners was also validated in a generic
multiple injection system with different flame tubes to confine the flame [74,75]. The system includes a resonant upstream
manifold formed by a duct having a continuously adjustable length and a combustion region in which a large number of
flames are stabilized on a multipoint injection system. The growth rates and eigenfrequencies are determined for a wide
range of duct lengths. For certain values of the length parameter one finds that a positive growth rate for vanishing small
amplitude levels indicating that the system is linearly unstable. The growth rate then changes as the amplitude is increased
and eventually vanishes for a finite amplitude level. For other values of the length, the growth rate is initially negative,
becomes positive for a finite amplitude and drops to zero for a higher level. This indicates that the system is linearly stable
but nonlinearly unstable. Using calculated growth rates it is possible to predict amplitudes of oscillation when the system
operates on a limit cycle. Mode hopping and instability triggering may also be anticipated by comparing the growth rate
curves. The case of turbulent swirling flames was recently examined in the FDF framework by including damping in the
analysis [76]. Typical calculations are displayed in Fig. 8. In general, predictions are in good agreement with measurements
indicating that the flame describing function (FDF) methodology constitutes a suitable framework for nonlinear instability
analysis.

7. Dynamical interactions with the precessing vortex core

The precessing vortex core (PVC) constitutes an intriguing feature of swirling flames. This helical instability is generally
present when the swirl number exceeds a certain value. It is mainly observed in cold flows. The persistence of the PVC
in reactive conditions is often questioned. While the role of the PVC in anchoring the flame is well established (see for
example [77]), its influence on combustion instability is less clear. Now, acoustic oscillations associated with combustion
instabilities and the PVC, are often found in swirl-stabilized flames, the interaction between these two phenomena is of
 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768 765

Fig. 8. Frequency–growth rate trajectories calculated with the FDF method for the generic configuration shown in Fig. 5. When the growth rate is below the
estimated damping rate α , the system is predicted to be stable. When the growth rate exceeds the damping rate, the system is predicted to be unstable.
Colors along the trajectory correspond to different levels of relative velocity fluctuation amplitude. When the system is unstable, the oscillation level at the
limit cycle corresponds to the value of u  /U b where ωi = α . From [76].

Fig. 9. (a) Spectral density of a photomultiplier collecting the chemiluminescence signal from the flame. One vertical half of the photomultiplier field of
view is masked, and the signal has been normalized by the mean value. (b) Spectral density of the signal recorded by a microphone located outside of the
flame tube. Bulk injection velocity U b = 9.9 m s−1 , equivalence ratio φ = 0.69. The quantities f i , f a and f h denote interaction, acoustic and helical-mode
frequency, respectively. From [81].

interest. It has been suggested that the PVC could act as driver of combustion instabilities [19,49,78] but an explanation
of the subtending processes is not available and one does not quite see how an asymmetric perturbation can couple to an
axisymmetric acoustic mode. It is of course possible to imagine that helical disturbances induced by the PVC could drive
a rotating acoustic mode in a multiple injector annular configuration, but in single-burner system, the transverse modes
typically reside at frequencies well above 1 kHz, while instabilities in these set-ups are mostly in the low frequency range,
generally of the order of a few hundred Hz. Due to its asymmetric structure, the heat release rate perturbation associated
with the helical mode cannot feed fluctuation energy into a plane acoustic mode, or more generally into a mode with
a wavelength exceeding the transverse dimension of the helical instability. This is consistent with several observations that
fluctuations in the velocity field and/or in the heat release rate that are associated with the helical mode produce no
measurable acoustic signature in the far field [79,80], as confirmed by recent experiments [81]. In these experiments, the
PVC is present in the reacting flow and its presence is characterized by recording light emission from one half of the flame
while blocking light radiated from the other half. This is accomplished by placing a screen with its boarder on the line of
sight between the recording camera and the burner axis. The PVC is also manifested in the azimuthal velocity fluctuations
and appears clearly in the power spectral densities of these signals (Fig. 9).
When the flame is modulated by an external acoustic wave or in the presence of an acoustically coupled instability,
an interaction takes place with the PVC giving rise to a difference frequency component f i = | f h − f a | formed by the PVC f h
and acoustic f a frequencies. By phase averaging emission images at this frequency, one obtains a “yin–yang” pattern which
rotates at the frequency f i . It is also worth noting that increasing the acoustic forcing amplitude resulted in a progressive
suppression of the PVC as was observed in other experiments [82] and simulations [83], leading thus also to a reduction
of the interaction component. It remains to see if such interactions may have a direct or indirect effect on thermo-acoustic
instabilities.

8. Conclusion

With the motivation of solving combustion instability issues in gas turbines, progress has been accomplished in swirling
flame dynamics. It is shown in this article that insight on this topic can be obtained by suitably combining analysis,
766 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768

modeling, experimentation and simulation. Many fundamental issues are now understood. It is well established that acoustic
waves interacting with the swirler unit produce a vorticity wave which is manifested by azimuthal velocity perturbations.
These fluctuations are convected by the flow and induce swirl number fluctuations. This in turn gives rise to an interference
between a mechanism of flame angle fluctuation and a mechanism of vortex roll-up associated with vortices shed from the
injector lip. This process defines the flame response to incident perturbations. It is also shown that the flame transfer func-
tion can be modeled on this basis. It is next shown that the FTF can be measured for different levels of input perturbations
providing a flame describing function (FDF). The FDF is then used in combination with an acoustic network description to
determine the stability of a generic system comprising a plenum, an injector equipped with a swirler and a flame tube.
Reasonable agreement is obtained between experiments and calculations relying on the FDF framework. Another intriguing
issue in swirling flames is that of the precessing vortex core (PVC). The role of this helical instability is not fully eluci-
dated, but it is shown that it can couple with acoustic waves associated with a thermo-acoustic instability. While progress
has been substantial, much remains to be uncovered with many emerging challenges from novel combustion systems and
technologies. The issue of the possible coupling between combustion and azimuthal modes in annular systems is currently
being investigated.

Acknowledgements

It is a pleasure to dedicate this article to Paul Clavin and to acknowledge the generous support provided by Snecma,
CNRS, DGA and ANR.

References

[1] Y.B. Zeldovich, G.I. Barenblatt, V.B. Librovich, G.M. Makhviladze, The Mathematical Theory of Combustion and Explosions, Plenum Press, New York,
1985 (English translation).
[2] A. Liñán, The asymptotic structure of counterflow diffusion flames for large activation energies, Acta Astronautica 1 (1974) 1007.
[3] P. Pelcé, P. Clavin, Influence of hydrodynamics and diffusion upon the stability limits of laminar premixed flames, Journal of Fluid Mechanics 124
(1982) 219–237.
[4] P. Clavin, Dynamic behavior of premixed flame fronts in laminar and turbulent flows, Progress in Energy and Combustion Science 11 (1985) 1–59.
[5] P. Clavin, Dynamics of combustion fronts in premixed gases: from flames to detonations, Proceedings of the Combustion Institute 28 (2000) 569–586.
[6] G.I. Sivashinsky, Some developments in premixed combustion modeling, Proceedings of the Combustion Institute 29 (2002) 1737–1761.
[7] J.D. Buckmaster, G.S.S. Ludford, Theory of Laminar Flames, Cambridge Univ. Press, 1982.
[8] A.K. Kapila, Asymptotic Treatment of Chemically Reacting Systems, Pitman, Boston, 1983.
[9] F.A. Williams, Combustion Theory, Benjamin Cummings, Menlo Park, CA, 1985.
[10] C.K. Law, Dynamics of stretched flames, Proceedings of the Combustion Institute 22 (1988) 1381–1402.
[11] C.K. Law, C.J. Sung, Structure, aerodynamics and geometry of premixed flamelets, Progress in Energy and Combustion Science 26 (2000) 459–505.
[12] C.K. Law, Combustion Physics, Cambridge Univ. Press, 2006.
[13] M. Matalon, B.J. Matkowsky, Flames as gasdynamic discontinuities, Journal of Fluid Mechanics 124 (1982) 239.
[14] M. Matalon, Flame dynamics, Proceedings of the Combustion Institute 32 (2009) 57–82.
[15] J. Buckmaster, P. Clavin, A. Liñán, M. Matalon, N. Peters, G. Sivashinsky, F.A. Williams, Combustion theory and modeling, Proceedings of the Combustion
Institute 30 (2005) 1–19.
[16] S. Candel, Combustion dynamics and control: progress and challenges, Proceedings of the Combustion Institute 29 (2002) 1–28.
[17] F.E.C. Culick, V. Burnley, G. Swenson, Pulsed instabilities in solid-propellant rockets, Journal of Propulsion and Power 11 (4) (1995) 657–665.
[18] F.E.C. Culick, Unsteady motions in combustion chambers for propulsion systems, AGARDograph, NATO/RTO-AG-AVT-039, 2006.
[19] Y. Huang, V. Yang, Dynamics and stability of lean-premixed swirl-stabilized combustion, Progress in Energy and Combustion Science 35 (4) (2009)
293–384.
[20] H. Tsien, Servo-stabilization of combustion in rocket motors, Journal of the American Rocket Society 22 (1952) 256–263.
[21] L. Crocco, Aspects of combustion instability in liquid propellant rocket motors. Part I, Journal of the American Rocket Society 21 (1951) 163–178.
[22] L. Crocco, S.I. Cheng, Theory of Combustion Instability in Liquid Propellant Rocket Motors, AGARDograph, vol. 8, Butterworths Science, 1956.
[23] F.E. Marble, D.W. Cox, Servo-stabilization of low-frequency oscillations in a liquid bipropellant rocket motor, Journal of the American Rocket Soci-
ety 23 (2) (1953) 63.
[24] D.J. Harrje, F.H. Reardon, Liquid propellant rocket instability, Tech. Rep. SP-194, NASA, 1972.
[25] K. McManus, T. Poinsot, S. Candel, A review of active control of combustion instabilities, Progress in Energy and Combustion Science 19 (1993) 1–29.
[26] T.C. Lieuwen, V. Yang (Eds.), Combustion Instabilities in Gas Turbines, Operational Experience, Fundamental Mechanisms, and Modeling, Progress in
Astronautics and Aeronautics, vol. 210, American Institute of Aeronautics and Astronautics, Inc., 2005.
[27] R. Pertersen, H. Emmons, Stability of laminar flames, Physics of Fluids 4 (1961) 456–464.
[28] S. Ducruix, D. Durox, S. Candel, Theoretical and experimental determination of the transfer function of a laminar premixed flame, Proceedings of the
Combustion Institute 28 (2000) 765–773.
[29] S. Ducruix, T. Schuller, D. Durox, S. Candel, Combustion dynamics and instabilities: Elementary coupling and driving mechanisms, Journal of Propulsion
and Power 19 (5) (2003) 722–734.
[30] D. Durox, T. Schuller, S. Candel, Combustion dynamics of inverted conical flames, Proceedings of the Combustion Institute 30 (2005) 1717–1724.
[31] R. Balachandran, B. Ayoola, C. Kaminski, A. Dowling, E. Mastorakos, Experimental investigation of the nonlinear response of turbulent premixed flames
to imposed inlet velocity oscillations, Combustion and Flame 143 (1–2) (2005) 37–55.
[32] V.N. Kornilov, K.R.A.M. Schreel, L.P.H. de Goey, Experimental assessment of the acoustic response of laminar premixed Bunsen flames, Proceedings of
the Combustion Institute 31 (2007) 1239–1246.
[33] A. Birbaud, S. Ducruix, D. Durox, S. Candel, The nonlinear response of inverted “V” flames to equivalence ratio non-uniformities, Combustion and
Flame 154 (3) (2008) 356–367.
[34] T. Poinsot, A. Trouvé, D. Veynante, S. Candel, E. Esposito, Vortex-driven acoustically coupled combustion instabilities, Journal of Fluid Mechanics 177
(1987) 265–292.
[35] T. Lieuwen, B. Zinn, The role of equivalence ratio fluctuations in driving combustion instabilities in low NOx, gas turbines, Proceedings of the Combus-
tion Institute 27 (1998) 1809–1816.
 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768 767

[36] T. Lieuwen, Modeling premixed combustion-acoustic wave interactions: A review, Journal of Propulsion and Power 19 (5) (2003) 765–781.
[37] H. Schwarz, L. Zimmer, D. Durox, S. Candel, Detailed measurements of equivalence ratio modulations in premixed flames using laser Rayleigh scattering
and absorption spectroscopy, Experiments in Fluids 49 (2010) 809–821.
[38] T. Schuller, D. Durox, S. Candel, Dynamics of and noise radiated by a perturbed impinging premixed jet flame, Combustion and Flame 128 (2002)
88–110.
[39] G.H. Markstein (Ed.), Nonsteady Flame Propagation, Pergamon Press, Elmsford, NY, 1964.
[40] C.K. Law, Propagation, structure, and limit phenomena of laminar flames at elevated pressures, Combustion Science and Technology 178 (2006) 335–
360.
[41] H. Wang, C. Law, T. Lieuwen, Linear response of stretch-affected premixed flames to flow oscillations, Combustion and Flame 156 (2009) 889–895.
[42] R. Lauvergne, F. Egolfopoulos, Unsteady response of C3 H8 –air laminar premixed flame submitted to mixture composition oscillations, Proceedings of
the Combustion Institute 28 (2000) 1841–1850.
[43] K. Schreel, R. Rook, L. de Goey, The acoustic response of burner stabilized premixed flat flames, Proceedings of the Combustion Institute 29 (2002)
115–122.
[44] L. de Goey, J. van Oijen, J. ten Thije Bookkamp, Propagation, dynamics and control of laminar premixed flames, Proceedings of the Combustion Insti-
tute 33 (2011) 863–886.
[45] T. Schuller, D. Durox, S. Candel, A unified model for the prediction of laminar flame transfer functions: comparison between conical and V-flame
dynamics, Combustion and Flame 134 (2003) 21–34.
[46] C. Paschereit, E. Gutmark, W. Weisenstein, Coherent structures in swirling flows and their role in acoustic combustion control, Physics of Fluids 11 (9)
(1999) 2667–2678.
[47] Y. Huang, V. Yang, Bifurcation of flame structure in a lean-premixed swirl-stabilized combustor: transition from stable to unstable flame, Combustion
and Flame 136 (2004) 383–389.
[48] C.O. Paschereit, B. Schuermans, V. Bellucci, P. Flohr, Combustion Instabilities in Gas Turbine Engines, Progress in Astronautics and Aeronautics, vol. 210,
American Institute of Aeronautics and Astronautics, Inc., 2005, Ch. 15.
[49] N. Syred, A review of oscillation mechanisms and the role of the precessing vortex core (PVC) in swirl combustion systems, Progress in Energy and
Combustion Science 32 (2006) 93–161.
[50] S. Thumuluru, T. Lieuwen, Characterization of acoustically forced swirl flame dynamics, Proceedings of the Combustion Institute 32 (2009) 2893–2900.
[51] T. Komarek, W. Polifke, Impact of swirl fluctuations on the flame response of a perfectly premixed swirl burner, Journal of Engineering for Gas Turbines
and Power 132 (6) (2010) 061503.
[52] A. Steinberg, I. Boxx, M. Stöhr, C. Carter, W. Meier, Flow–flame interactions causing acoustically coupled heat release fluctuations in a thermo-
acoustically unstable gas turbine model combustor, Combustion and Flame 157 (2010) 2250–2266.
[53] D. Kim, J.G. Lee, B.D. Quay, D. Santavicca, K. Kim, S. Srinivasan, Effect of flame structure on the flame transfer function in a premixed gas turbine
combustor, Journal of Engineering for Gas Turbines and Power 132 (2010) 021502.
[54] K. Kim, J. Lee, H. Lee, B. Quay, D. Santavicca, Characterization of forced flame response of swirl-stabilized turbulent lean-premixed flames in a gas
turbine combustor, Journal of Engineering for Gas Turbines and Power 132 (2010) 041502.
[55] P. Palies, D. Durox, T. Schuller, S. Candel, Dynamics of premixed confined swirling flames, Comptes Rendus Mécanique 337 (2009) 395–405.
[56] P. Palies, D. Durox, T. Schuller, S. Candel, The combined dynamics of swirler and turbulent premixed swirling flames, Combustion and Flame 157 (2010)
1698–1717.
[57] P. Palies, T. Schuller, D. Durox, L. Gicquel, S. Candel, Acoustically perturbed turbulent premixed swirling flames, Physics of Fluids 23 (2011) 037101 (15
pages).
[58] G. Staffelbach, L. Gicquel, G. Boudier, T. Poinsot, Large eddy simulation of self excited azimuthal modes in annular combustors, Proceedings of the
Combustion Institute 32 (2009) 2909–2916.
[59] N. Cumpsty, F. Marble, The interaction of entropy fluctuations with turbine blade rows; a mechanism of turbojet engine noise, Proceedings of the Royal
Society of London, Series A 357 (1977) 323–344.
[60] P. Palies, D. Durox, T. Schuller, S. Candel, Acoustic-convective mode conversion in an airfoil cascade, Journal of Fluid Mechanics 672 (2011) 545–569.
[61] P. Palies, D. Durox, T. Schuller, S. Candel, The response of swirling premixed flames to velocity perturbations, in: Proceedings of the Fourth European
Combustion Meeting, Vienna, Austria, 14–17 April 2009.
[62] T. Schuller, D. Durox, A. Cuquel, P. Palies, J. Moeck, S. Candel, Modeling the response of premixed flame transfer functions – key elements and exper-
imental proofs, in: 50th AIAA Aerospace Sciences Meeting including the New Horizons Forum and Aerospace Exposition, Nashville, Tennessee, 9–12
January 2012, AIAA 2012-0985.
[63] L. Boyer, J. Quinard, On the dynamics of anchored flames, Combustion and Flame 82 (1) (1990) 51–65.
[64] P. Palies, T. Schuller, D. Durox, S. Candel, Modeling of swirling flames transfer functions, Proceedings of the Combustion Institute 33 (2011) 2967–2974.
[65] A.P. Dowling, S.R. Stow, Acoustic analysis of gas turbine combustors, Journal of Propulsion and Power 19 (5) (2003) 751–764.
[66] T. Sattelmayer, Influence of the combustor aerodynamics on combustion instabilities from equivalence ratio fluctuations, Journal of Engineering for Gas
Turbines and Power 125 (2003) 11–19.
[67] T. Poinsot, D. Veynante, Theoretical and Numerical Combustion, 2nd edition, Edwards, 2005.
[68] K. Truffin, T. Poinsot, Comparison and extension of methods for acoustic identification of burners, Combustion and Flame 142 (4) (2005) 388–400.
[69] C.O. Paschereit, B. Schuermans, W. Polifke, O. Mattson, Measurement of transfer matrices and source terms of premixed flames, Journal of Engineering
for Gas Turbines and Power 124 (2002) 239–247.
[70] W. Krebs, P. Flohr, B. Prade, S. Hoffmann, Thermoacoustic stability chart for high intensity, Combustion Science and Technology 174 (7) (2002) 99–128.
[71] B. Schuermans, F. Guethe, D. Pennel, D. Guyot, C.O. Paschereit, Thermoacoustic modeling of a gas turbine using transfer functions measured at full
engine pressure, in: ASME Turbo Expo 2009: Power for Land, Sea, and Air, Orlando, FL, USA, June 8–12, 2009, Paper No. GT2009-59605, pp. 503–514.
[72] A. Dowling, A kinematic model of a ducted flame, Journal of Fluid Mechanics 394 (1999) 51–72.
[73] N. Noiray, D. Durox, T. Schuller, S. Candel, A unified framework for nonlinear combustion instability analysis based on the flame describing function,
Journal of Fluid Mechanics 615 (2008) 139–167.
[74] F. Boudy, D. Durox, T. Schuller, S. Candel, Nonlinear mode triggering in a multiple flame combustor, Proceedings of the Combustion Institute 33 (2011)
1121–1128.
[75] F. Boudy, D. Durox, T. Schuller, G. Jomaas, S. Candel, Describing function analysis of limit cycles in a multiple flame combustor, Journal of Engineering
for Gas Turbines and Power 133 (2011) 061502.
[76] P. Palies, D. Durox, T. Schuller, S. Candel, Nonlinear combustion instabilities analysis based on the flame describing function applied to turbulent
premixed swirling flames, Combustion and Flame 158 (2011) 1980–1991.
[77] M. Stohr, I. Boxx, C. Carter, W. Meier, Dynamics of lean blowout of a swirl-stabilized flame in a gas turbine model combustor, Proceedings of the
Combustion Institute 33 (2011) 2953–2960.
[78] A. Gupta, G. Lilley, N. Syred, Swirl Flows, Abacus Press, Kent, England, 1984.
[79] C.O. Paschereit, E. Gutmark, W. Weisenstein, Excitation of thermoacoustic instabilities by interaction of acoustics and unstable swirling flow, AIAA
Journal 38 (6) (2000) 1025–1034.
768 S. Candel et al. / C. R. Mecanique 340 (2012) 758–768

[80] I. Boxx, M. Stöhr, C. Carter, W. Meier, Temporally resolved planar measurements of transient phenomena in a partially pre-mixed swirl flame in a gas
turbine model combustor, Combustion and Flame 157 (2010) 1510–1525.
[81] J. Moeck, J. Bourgouin, D. Durox, T. Schuller, S. Candel, Nonlinear interaction between a precessing vertex core and acoustic oscillations in a turbulent
swirl flame, Combustion and Flame 159 (2012) 2650–2668.
[82] A. Lacarelle, T. Faustmann, D. Greenblatt, C. Paschereit, O. Lehmann, D. Luchtenburg, B. Noack, Spatiotemporal characterization of a conical swirler flow
field under strong forcing, Journal of Engineering for Gas Turbines and Power 131 (2009) 031504.
[83] P. Ludiciani, C. Duwig, Large eddy simulation of the sensitivity of vortex breakdown and flame stabilisation to axial forcing, Flow, Turbulence and
Combustion 86 (2011) 639–666.