On the Flow Structure and Dynamics of Methane and Syngas Lean Flames in a Model Gas-Turbine Combustor

Abstract

The present paper compares the flow structure and flame dynamics during combustion of methane and syngas in a model gas-turbine swirl burner. The burner is based on a design by Turbomeca. The fuel is supplied through injection holes between the swirler blades to provide well-premixed combustion, or fed as a central jet from the swirler’s centerbody to increase flame stability via a pilot flame. The measurements of flow structure and flame front are performed by using the stereo particle image velocimetry and OH planar laser-induced fluorescence methods. The measurements are performed for the atmospheric pressure without preheating and for 2 atm with the air preheated up to 500 K. The flow Reynolds numbers for the non-reacting flows at these two conditions are 1.5 × 10³ and 1.0 × 10³, respectively. The flame dynamics are analyzed based on a high-speed OH* chemiluminescence imaging. It is found that the flame dynamics at elevated conditions are related with frequent events of flame lift-off and global extinction, followed by re-ignition. The analysis of flow structure via the proper orthogonal decomposition reveals the presence of two different types of coherent flow fluctuations, namely, longitudinal and transverse instability modes. The same procedure is applied to the chemiluminescence images for visualization of bulk movement of the flame front and similar spatial structures are observed. Thus, the longitudinal and transverse instability modes are found in all cases, but for the syngas at the elevated pressure and temperature the longitudinal mode is related to strong thermoacoustic fluctuations. Therefore, the present study demonstrates that a lean syngas flame can become unstable at elevated pressure and temperature conditions due to a greater flame propagation speed, which results in periodic events of flame flash-back, extinction and re-ignition. The reported data is also useful for the validation of numerical simulation codes for syngas flames.

1. Introduction

The efficient combustion of synthesis gas or ‘syngas’ in gas turbines is becoming more relevant due to the development of integrated gasification combined cycle (IGCC) power plants, where the products of coal gasification are burned in gas turbines [1]. Moreover, syngas can be obtained from biomass and even some types of waste [2]. Therefore, the studies of syngas combustion at elevated pressure and temperature conditions are important. Modern low-emission gas turbine combustors are based on a lean premixed combustion technology [3]. The optimization of the flow aerodynamics allows organizing efficient fuel combustion with low concentrations of both CO and NOx in combustion products. In particular, NOx lower than 5 ppm (corrected to 15% O₂) can be obtained for modern swirl burners [4,5,6]. However, it is well recognized that the lean combustors are prone to unstable operation and thermoacoustic pulsations [7,8], which results in reduced efficiency of the burning and leads to damage to the turbine. Therefore, it is important to study the contribution of different physical and chemical processes to the feedback mechanism of the thermoacoustic resonance [9,10,11,12]. Another important issue for syngas premixed flames in lean combustors is an increased probability of the flame flash-back due to the increased turbulent flame speed because of the hydrogen presence [13,14].

Modern optical measurement techniques provide detailed information on the local pulsations of the gas velocity, temperature, and species concentration. However, such measurements are quite expensive and difficult for real combustion chambers due to several reasons, including strong vibrations and limited optical access. Therefore, the impact of different processes on unsteady flow and flame dynamics is performed for research combustion sectors, which are equipped with observation windows, allowing for detailed optical measurements and modeling basic phenomena in real combustors [15,16,17,18,19,20,21,22]. Nowadays, planar optical methods like the particle image velocimetry (PIV) and planar laser-induced fluorescence (PLIF) have become very popular for the measurements of flow structure and dynamics and for the analysis of flow/flame interactions [23,24,25,26].

PLIF provides the detection of different types of molecules in a selected cross-section of the reacting flow studied, including OH, CO, NO, SO₂, and others [27,28,29,30,31,32,33]. Therefore, the PIV and PLIF methods are used simultaneously for the measurements of the velocity flow field and visualization of high-temperature regions, where chemical reactions take place. In particular, PLIF for HCO or simultaneously for HCHO and OH can be used to visualize the regions with intensive heat release [34,35,36,37,38]. Two-line or thermally-assisted OH PLIF can be used for the evaluation of temperature fields hot combustion products [39,40,41,42]. Since swirl-stabilized combustors for gaseous or liquid fuels are commonly used in gas turbines, the flow and flame for their models are intensively studied by using the PIV and OH PLIF methods [43,44,45]. There is also a growing interest in the optical diagnostics of syngas combustion in swirl-stabilized burners [21].

Currently, the flow structure and dynamics, the flame stability and thermoacoustic pulsations, flame blow-off or flash-back are studied for different types of model and realistic gas-turbine swirl burners. The most popular ones among them are a low-swirl injector [4,5,8,14], a TECFLAM burner [46], model combustors based on a design by Turbomeca (a generic nozzle studied in Janus and others [15,47] and a model gas-turbine burner studied in Meier and others [7,24]), a G30 dry low emission burner by Siemens [48,49,50], a dual-swirler as a model of air-blast injector [16,23]. There are also a number of papers for dual swirlers of more sophisticated designs, e.g., LPP (lean premixing prevaporizing, [51,52]) and PERM (partially evaporated, rapidly mixed, [53,54]) injectors by Avio or so-called BIMER experimental setup [55,56]. There are also some studies for triple swirlers [57,58,59] and for realistic burners, like different lean premixed prevaporized injectors (see studies in cooperation with General Electric [18,20], Rolls-Royce [17,19,22,60], Jaxa [45], and other companies [61,62]) and lean premixed burners by Siemens [21,63].

There are also several detailed studies of flame dynamics and flow structure for syngas combustion under gas-turbine conditions for simple flame configurations [64] and for model [65,66,67,68] and realistic [69,70,71] swirl burners. It is generally recognized that the methane replacement by syngas leads to flame flash-back problems due to a higher flame propagation speed. It is also noted that the flame propagation speed affects the frequency (and mode) of thermo-acoustic instabilities [66,70]. Therefore, further investigations of syngas flame dynamics and stability in swirl-stabilized combustion chambers are needed.

The present paper compares the flow and flame dynamics for methane and syngas burning in a model gas-turbine combustor based on the design by Turbomeca [15] utilizing PIV, OH PLIF measurements, and high-speed imaging of OH* chemiluminescence. The data is processed by a snapshot proper orthogonal decomposition (POD) to extract large-scale coherent structures [72]. Section 2 of the paper provides details on the experimental setup and data processing. The time-averaged data are reported in Section 3.1. The coherent structures, extracted from the PIV/PLIF data OH* images, are discussed in Section 3.2 and Section 3.3, respectively. The obtained results demonstrate that at elevated pressure and temperature conditions, syngas lean flames can become unstable due to periodical events of flame flash-back, extinction, and re-ignition.

2. Experimental Setup

2.1. Combustor

The details on the combustion chamber, including a 3D model of the geometry, can be found in [47]. The combustion chamber is equipped with observation windows, which are made of fused silica and provide flame observation and optical diagnostics for the pressure up to 8 atm. A radial swirler is placed upstream of the inlet (with the inner diameter D of 37 mm) of the combustion chamber. The chamber outlet is organized by an axisymmetric contraction choked nozzle. The throat of the nozzle is cooled by water. The air is supplied through the swirler and two peripheral slots inside the chamber. The slots are used for the film-cooling of the observation windows and contraction nozzle.

The swirler provides two zones of gas combustion, viz., a pilot central jet and a main premixed zone. Therefore, the fuel gas can be supplied from a central nozzle drilled through the centerbody of the swirler, or between the vanes of the swirler. The air and fuel flow rates are controlled by mass-controllers (Bronkhorst). Methane and syngas (H₂/CO mixture) are supplied from pressurized vessels. In practice, the syngas composition differs strongly depending on a feedstock, processing technology, and gas treatment [73]. In laboratory studies, the H₂:CO ratio is often fixed as 1:2 or 1:1 (e.g., [74]) by referring to coal or oil gasification processes, respectively [1]. For simplicity, the 1:1 H₂/CO mixture is used in the present case, expecting a higher flash-back effect for the greater hydrogen content. The global equivalence ratios of the studied flames are 0.33 and 0.2 for the methane and syngas, respectively. The ratios between the flowrates of the pilot and main fuel are 0.3 and 0.17, respectively. These flowrates were selected close to lean blow-off conditions for the flames.

The swirler is mounted between the combustion chamber and a cylindrical plenum chamber, equipped with a cord-shaped electric wire heater for the air preheating and a perforated plate to produce well-determined inflow conditions. The flows were studied for the normal (293 K and 1 atm.) and elevated (500 K and 2 atm.) temperature and pressure. The Reynolds numbers for non-reacting air flows at these conditions are 1.5 × 10⁴ and 1 × 10⁴, respectively. The flame parameters are provided in

Table 1. Flame parameters.

Air Flowrate (ln/min)	Fuel Type	Fuel Flowrate Main/Central (ln/min)	Equivalence Ratio	Air Temperature (K)	Pressure (atm.)
398	Methane	10.7/3.2	0.33	293	1
398	Methane	10.7/3.2	0.33	500	2
398	Syngas	29/4.9	0.2	293	1
398	Syngas	29/4.9	0.2	500	2

(please note that the values of normal liter per minute, ln/min, correspond to the normal temperature and pressure standard of NIST). For the PIV measurements, the air flow was seeded with TiO₂ solid particles by a mechanical mixer, placed upstream of the plenum chamber.

2.2. Measurement Equipment

The used stereo PIV system consisted of 4 Mpix CCD cameras (ImperX Bobcat IGV-B2020) and a double-head pulsed Nd:YAG laser (Beamtech Vlite 200). The cameras were equipped with optical lenses (Sigma 105 mm) with scheimpflug adapters and band-pass (532 ± 5 nm) optical filters. The PIV system was combined with a OH PLIF system (see Figure 1), which consisted of a tunable dye laser (Sirah Precision Scan), pumped by a pulsed Nd:YAG laser (QuantaRay) and an sCMOS camera equipped with an intensifier (LaVision IRO). The radiation of the dye laser passed through BBO doubling crystal to obtain the wavelength of 283 nm to excite the fluorescence of OH for the Q₁(8) line of the A²Σ⁺–X²Π (1–0) band. The fluorescence images were captured by the intensified camera, equipped with an optical lens, made of fused silica, and a band-pass optical filter (300–320 nm). The lasers and cameras were synchronized by a TTL signal generator.

The cameras were also spatially calibrated by putting a flat calibration target inside the combustion chamber, imaged in different normal-to-plane locations prior to the experiments. A part of the PLIF laser sheet was reflected into a calibration cuvette, filled with Rhodamine solution. The fluorescence inside the cuvette was captured by another CCD camera (ImperX Bobcat IGV-B4820) to compensate for a non-uniform laser sheet distribution during PLIF image processing. It was also used to account for the shot-to-shot fluctuations of PLIF laser energy. The PLIF laser pulse was arranged between the time interval of two PIV laser shots. The exposition of the PLIF camera was 200 ns.

High-speed visualization of the flame dynamics was performed separately by using a high-speed intensifier (LaVision high-speed IRO) connected to a CMOS camera (Photron SA5). The intensifier included an S20 photocathode and was equipped with another fused silica UV lens and a band-pass optical filter for OH* (300–340 nm). The acquisition frame rate was 9.9 kHz. Direct images of the flames were also captured by using a photo camera.

2.3. Data Processing

The images were stored and processed by using an in-house “ActualFlow” software. The PIV images were processed by an iterative cross-correlation algorithm with continuous window shift and deformation [75]. The stereo calibration and reconstruction are based on 3rd-order polynomial mapping functions, evaluated for each camera during the stereo calibration [76]. The same model was used for the back-projection of the PLIF images. Before that, the PLIF images were processed to remove a background signal (captured without combustion, when the PLIF laser was running), correct for the non-uniform laser sheet and energy variation, account for the laser beam attenuation during the absorption inside the combustion chamber [47]. In total, 1500 snapshots of the PIV/PLIF data were captured for each flame case.

Afterwards, the PIV/PLIF snapshots were processed by a proper orthogonal decomposition method. This was done based on the singular value decomposition (SVD) of the data matrix [77] U = [u′(x, t₁)… u′(x, t_N)], where u′(x, t_i) is the snapshot of fluctuating velocity at the time instant t_i.

$$U=W\Sigma V^T,\mathrm{or}{u}^{\prime }\left(x,t_k\right)={\displaystyle \sum _{q=1}^N{\alpha }_q\left(t_k\right){\sigma }_q}{\mathsf{\phi }}_q\left(x\right),$$

(1)

Here, the matrix of left-singular vectors W = [ϕ₁(x)…ϕ_N(x)] contains the POD modes, V^T corresponds to the normalized temporal coefficients α_q (V is the matrix of right-singular vectors), and Σ = diag[σ₁… σ_N] is the matrix of singular values, which are equal to square root values of the eigenvalues of covariation matrix. The POD modes and temporal coefficients are orthonormal:

$$W^{T}W=I_M, V^{T}V=I_N, \mathrm{or} {\displaystyle \sum _{k=1}^M{\mathsf{\phi }}_i\left(x_k\right){\mathsf{\phi }}_j\left(x_k\right)}={\delta }_{ij}, {\displaystyle \sum _{k=1}^N{\alpha }_i\left(t_k\right)}{\alpha }_j\left(t_k\right)={\delta }_{ij}$$

(2)

In Equation (2) δ_ij is the Kronecker symbol and I is the identity matrix. The concentration fluctuation fields C = [c′(x, t₁)… c′(x, t_N)] can be phase-averaged to obtain the coherent pulsations of the concentration [78,79].

$${\widehat{c}}_q\left(x\right)=\left(1/\sqrt{N}\right){\displaystyle \sum _{k=1}^N{a}_{qk}{c}^{\prime }\left(x,t_k\right)}$$

(3)

3. Results

3.1. Average Flow Structure

Figure 2 compares the mean velocity and OH PLIF intensity data for different flames. The velocity field is shown by vectors and normalized by the bulk velocity of the main air flow, which are 6.36 and 5.43 m/s for the cases of normal and elevated conditions, respectively. For each flame case, the OH PLIF data is normalized by a maximum value. The data are captured in the central (x, y) plane with the origin (0, 0) defined at the center of the pilot nozzle outlet.

In all cases, there is a central recirculation zone present. For the methane and syngas flames at normal conditions the maximum lateral cross-section of the recirculation zone is approximately of 0.76D and 0.88D, respectively. For methane, the elevated pressure and temperature do not affect strongly the velocity distribution. However, the chemical reaction zone in the inner shear layer appears to be thinner for the high pressure and temperature case. For syngas at normal conditions, the time-averaged reaction zone is even thinner. For the elevated pressure and temperature, the chemical reaction zone is stabilized close to the centerbody and is much shorter. For this flame the lateral size of the recirculation zone is also smaller, viz., approximately 0.62D. Moreover, the reverse flow magnitude in the recirculation zone is considerably higher, because the flame is shifted upstream in comparison to other flames.

Figure 3 shows the flame images captured by the photo camera. The soot is visible for the flames at normal conditions because the flame is stabilized with the aid of the central non-premixed jet. For syngas, the flame at elevated conditions is brighter and the nozzle is incandescent due to intensive heating by the closely located flame front. This is demonstrated by the average OH PLIF data, which indicates that the flame is stabilized close to the centerbody. Remarkably, a very strong soot luminosity is observed for the methane flame at the elevated pressure and temperature.

Figure 4 shows the spatial distributions of the variance of the y-component of the velocity fluctuations. There are two separate shear layers with intensive turbulent fluctuations around the annular swirling flows for the methane and syngas flames at normal conditions. For the elevated pressure and temperature, the shear layers are not distinct for the methane and could not be seen at all for the syngas flame. Moreover, it is noteworthy that the local variance of the vertical velocity component reaches 100% of the flow bulk velocity for the latter case.

3.2. Coherent Flow Structures

Figure 5 shows the instantaneous snapshots of the PIV and PLIF data. For each flame they were captured simultaneously. The velocity fields demonstrate that large-scale vortex structures are present around the annular swirling flow. Moreover, the reverse flow is not steady. The PLIF snapshot for the methane flame at normal conditions demonstrates that the flame can be detached, viz., its root on one side is located outside the swirler nozzle. For the elevated conditions, the flame is stabilized inside the nozzle almost all the time. For the syngas flame at normal conditions, the PLIF snapshot visualizes a flame front around the central jet. The central jet is not seen in the velocity fields, because it was not seeded by tracer particles. However, since the bulk velocity of the central jet is higher (more than twice) for the syngas cases, it penetrates further inside the recirculation zone. For the elevated temperature and pressure conditions a tip of the flame front is located close to the central body and most of the flame surface is inside the swirler. Thus, the present flow case is characterized by a partial flame flash-back.

The POD spectra of the velocity fluctuations for the studied flames are shown in Figure 6. One most intensive mode is detected for the methane flame at normal conditions. This mode is shown by vectors in Figure 7, where the color corresponds to the phase-averaged value of the OH PLIF intensity fluctuations. This mode corresponds to a transverse hydrodynamics instability mode with almost asymmetrical pulsations of velocity components and OH PLIF intensity on opposite sides of the flow. In contrast, the second and the third POD modes are related to almost symmetrical pulsations of the velocity along the annular jet and symmetrical variation of the OH PLIF signal in the flow core. These two different types of POD modes are considered to be related to a large-scale transversal and longitudinal hydrodynamic instability modes. For the methane flame at elevated pressure and temperature, the POD spectrum corresponds to a monotonous decrease of the kinematic energy content with the mode number. As Figure 8 shows, the first mode is related to nearly symmetric OH PLIF intensity pulsations in the upper part of the recirculation bubble and strong coherent pulsations of the longitudinal flow velocity. This POD mode should correspond to a longitudinal instability mode, whereas the transverse hydrodynamics mode appears only as the fourth POD mode.

For the syngas flame at normal conditions, the first and second POD modes have considerably greater energy than the remaining modes. Their spatial distributions in Figure 9 show that the second POD mode corresponds to nearly symmetrical coherent fluctuations of OH PLIF signal around the central jet and also along the annular swirling jet. The velocity fluctuations are related to a flapping motion of the annular jet. Traces of a transversal hydrodynamic mode are seen in the first, third, and fourth POD modes with low coherent fluctuations of the OH PLIF intensity. For the syngas at elevated conditions, the first and second POD modes, depicted in Figure 10 are not perfectly symmetrical, but appear to be related to two phases (shifted by π/2) of a strong longitudinal hydrodynamics instability mode (they correspond to a much greater kinematic energy than the other modes), coupled with very strong pulsations of the OH PLIF intensity (up to 20%). The third and fourth POD modes are nearly asymmetrical and expected to be produced by a transversal mode.

3.3. High-Speed Imaging of Flame Front Dynamics

Figure 11 and Figure 12 show the time-averaged and instantaneous distributions of OH* chemiluminescence signal, respectively. Generally, except for one case, the time-averaged data of the chemiluminescence projection onto the image corresponds well to the time-averaged OH PLIF data, obtained for the central cross-section of the flame. For the syngas at normal and elevated conditions, the OH* chemiluminescence signal is the most intensive above the diffusion central flame (at y/d ≈ 1.2 and 0.4, respectively), whereas the OH fluorescence intensity is greater for the flame front near the swirler’s nozzle exit. This can be due to the fact that the OH radical in hydrocarbon flames is produced mainly via chemical reaction (4), whereas the electronically excited OH* is formed predominantly by another chemical reaction (5) [80,81]. Moreover, the chemiluminescence of CO₂* also should significantly contribute to flame radiation in the wavelength range of 300–340 nm [82].

$$•\mathrm{H}+{\mathrm{O}}_2\to •\mathrm{OH}+•\mathrm{O}$$

(4)

$$\mathrm{CH}+{\mathrm{O}}_2\to \mathrm{CO}+\mathrm{OH}*$$

(5)

The animations of the OH* high-speed imaging are provided as Supplementary Materials for this paper. The examples of the snapshots are shown in Figure 12. At normal conditions, the flame front was stabilized around the central recirculation zone and penetrated inside the nozzle. It is also observed to be partially lifted, when the flame is blown-off on one side by the flow (see the example in Figure 12a). For the elevated conditions, the flame dynamics are found to be more complex, associated with periodical lift-off of the entire flame (Figure 12c), turbulent propagation against the rotating annular flow and also frequent events of global extinction (Figure 12e,f), after the flame front penetration inside the swirler. However, the flame was not blown-out by the flow due to the presence of the central diffusion pilot flame.

The local variance for OH* image is shown in Figure 13. To visualize a bulk, coherent movement of the flame front, the SVD (1) is applied to the OH* chemiluminescence images. Whereas the POD is a common term for the velocity data analysis, the SVD decomposition for the projections OH* chemiluminescence is referred to as principle component analysis (PCA), as it is more common in literature. The PCA spectrum is shown in Figure 14. In analogy to the POD of velocity data, for the case of methane flame at normal conditions, the first PCA mode has a much greater magnitude in comparison to the remaining modes.

The first four modes of the PCA for the OH* chemiluminescence images for the methane flame at normal conditions are shown in Figure 15. The first and second modes are clearly related to the transversal and longitudinal modes of large-scale coherent variations of the flame front pattern. For the present case, the characteristic frequency of the temporal coefficient variation for the transversal bulk mode corresponds to 26 Hz and the Strouhal number of St_T = 0.15. For the longitudinal mode, no such peak is found for the spectrum of α(t). For the other flame case, the difference in modes’ amplitude is not so strong. However, for each flame such two different types of bulk modes, viz., nearly symmetrical and asymmetrical, are detected.

Figure 16 shows the PCA modes of the OH* images for the methane flame at the elevated conditions. The first PCA mode is related to the transversal movement, whereas the longitudinal movement appears in the second and third PCA modes, which correspond to a phase shift of π/2. The longitudinal oscillations correspond to a peak near 22 Hz in the Fourier spectrum of α(t) or to St_L = 0.15. The PCA modes for the syngas flames at normal and elevated conditions are shown in Figure 17 and Figure 18, respectively. For the normal condition, the bulk movements in the transversal and longitudinal directions correspond to the first and second PCA, correspondingly. The Fourier spectra of the temporal coefficients for these modes do not have a distinguishing peak for a certain dedicated frequency. For the elevated conditions, the longitudinal bulk mode is related to the first and third PCA modes, for which a very sharp peak was found at 68 Hz (St_L = 0.46) in the spectra of the temporal correlation coefficients (see the example in Figure 14b). The transverse bulk movement mode was also detected for the present case (second PCA mode), but its spectrum does not contain any distinguished peak.

4. Conclusions

The present paper reports on the PIV and OH PLIF measurements of the flow structure during the combustion of methane and syngas in a model gas-turbine combustor. The flames are stabilized by a generic swirler, based on a design by Turbomeca, which supplies a part of the fuel as a pilot central jet and injects the remaining part between the vanes of the swirler to organize the premixed lean combustion. The flames are studied at both normal and elevated pressure and temperature for the equivalence ratios close to the lean blow-off limits. The flame dynamics are also visualized by using high-speed OH* chemiluminescence imaging. The data are processed by SVD to reveal coherent flow structures and bulk movement of the turbulent flame.

The methane flame at normal conditions is not perfectly stable with the events of partial lift-off. However, the dynamics are much more complex for the elevated pressure and temperature conditions, associated with the lift-off of the entire flame, global extinction, and re-ignition. The first few POD modes for PIV/PLIF data reveal the presence of coherent longitudinal and transversal modes. These modes are not perfectly symmetrical, or asymmetrical, since the local gas density fluctuations are not accounted for during the decomposition. Nevertheless, the traces of the longitudinal and transversal coherent modes of velocity fluctuations are detected in all cases.

The SVD analysis of OH* chemiluminescence also reveals similar spatial modes for the bulk flame movement. The Fourier analysis of the temporal coefficients for these modes show that for atmospheric methane flame, which is less stable than the syngas flame, the transversal mode corresponds to the Strouhal number St_T = 0.15. For the elevated pressure, the unsteady flame dynamics is associated with longitudinal oscillating bulk movement of the turbulent flame pattern. For the syngas flame with strong pulsations, the Fourier spectrum features several peaks at different frequencies with the most intensive for the Strouhal number of St_L = 0.46.

In summary, the dynamics of the lean methane flame close to the lean blow-off limit is found to be related with the events of asymmetrical flame lift-off, accumulation of the me-thane inside the recirculation zone (due to the presence of the pilot jet), and flame re-entrance into the swirler nozzle. Therefore, detected transversal and longitudinal spatial modes for this flame are associated with these events of the flame lift-off and re-entrance. For the elevated conditions, the lift-off of the entire methane flame is observed, followed by the flame re-entrance into the swirler with global extinction inside, followed by re-ignition inside the combustion chamber by the diffusion central jet. The unsteady dynamics of the syngas flame at elevated conditions, coupled with strong thermoacoustic pulsations, is also related with periodic events of the flame flash-back, extinction inside the nozzle, and re-ignition, but without flame lift-off due to a higher flame propagation speed.

Therefore, the present study shows that the scenario of unsteady dynamics lean syngas flames can be sufficiently different from that for methane flames. Thus, the combustion of lean syngas flames in model gas-turbine combustion chambers deserves additional studies for a wider range of parameters to develop efficient methods for the suppression of thermoacoustic pulsations in lean combustors.