Search

The present article investigates the interactions between the pilot and main flames in a novel stratified swirl burner using both experimental and numerical methods. Experiments are conducted in a test rig operating at atmospheric conditions. The system is equipped with the BASIS (Beihang Axial Swirler Independently-Stratified) burner fuelled with premixed methane-air mixtures. To illustrate the interactions between the pilot and main flames, three operating modes are studied, where the burner works with: (i) only the pilot flame, (ii) only the main flame, and (iii) the stratified flame (with both the pilot and main flames). We found that: (1) in the pilot flame mode, the flame changes from V-shape to M-shape when the main stage is switched from closed to supplying a pure air stream. Strong oscillations in the M-shape flame are found due to the dilution of the main air to the pilot methane flame. (2) In the main flame mode, the main flame is lifted off from the burner if the pilot stage is supplied with air. The temperature of the primary recirculation zone drops substantially and the unsteady heat release is intensified. (3) In the stratified flame mode, unique beating oscillations exhibiting dual closely-spaced frequencies in the pressure spectrum is found. This is observed within over narrow window of equivalence ratio combinations between the pilot and main stages. Detailed analysis of the experimental data shows that flame dynamics and thermoacoustic couplings at these two frequencies are similar to those of the unstable pilot flame and the attached main flame cases, respectively. Large Eddy Simulations (LESs) are carried out with OpenFOAM to understand the mechanisms of the time-averaged flame shapes in different operating modes. Finally, a simple acoustic analysis is proposed to understand the acoustic mode nature of the beating oscillations. [ABSTRACT FROM AUTHOR]

Thermoacoustic instabilities in annular gas turbine combustors often involve modes which vary in both the longitudinal and circumferential directions. Recent experimental studies show that during limit cycle oscillations, different thermoacoustic modes may be uncoupled, as is the case in purely longitudinal or circumferential spinning modes. They may also be coupled, for example two counter-rotating circumferential modes combining to give standing or mixed modes, and coupling between circumferential and longitudinal modes giving rise to the slanted mode. Accurately predicting such modal couplings and the resulting spatial pattern of limit cycle oscillations remains an open challenge. This work uses a 2-D low-order network model based on modal expansions, validated against a full 3-D Helmholtz solver, to systematically investigate these couplings. For the first time, low-order network modelling is shown to capture limit cycle oscillations exhibiting both uncoupled and nonlinearly coupled modes, the latter including coupling between counter-rotating circumferential modes and between longitudinal and circumferential modes. It is shown that limit cycle solutions with totally different mode patterns (longitudinal, circumferential spinning, circumferential standing and slanted) can all exist in a given thermoacoustic system, with switches between modal patterns arising from slight changes in parameters such as the flame time delay. [ABSTRACT FROM AUTHOR]

The forced flame responses in a pressurized gas turbine combustor are predicted using numerical reacting flow simulations. Two incompressible1 large eddy simulation solvers are used, applying two combustion models and two reaction schemes (4-step and 15-step) at two operating pressures (3 and 6 bar). Although the combustor flow field is little affected by these factors, the flame length and heat release rate are found to depend on combustion model, reaction scheme, and combustor pressure. The flame responses to an upstream velocity perturbation are used to construct the flame describing functions (FDFs). The FDFs exhibit smaller dependence on the combustion model and reaction chemistry than the flame shape and mean heat release rate. The FDFs are validated by predicting combustor thermoacoustic stability at 3 and 6 bar and, for the unstable 6 bar case, also by predicting the frequency and oscillation amplitude of the resulting limit cycle oscillation. All of these numerical predictions are in very good agreement with experimental measurements. [ABSTRACT FROM AUTHOR]

Premixed lean hydrogen-air flames exhibit instabilities due to hydrodynamic instabilities but also thermo-diffusive instabilities. However, in the case of ammonia/hydrogen blends, the effect of ammonia addition on instabilities is still unclear. To investigate intrinsic instabilities in premixed ammonia/hydrogen-air flames, a parametric study of laminar premixed flames is performed for different fuel contents, from pure hydrogen (H 2) to a blend of 40% of H 2 and 60% of ammonia (NH 3) in volume, equivalence ratios (0.4 to 1.0) and pressures (1 and 10 bar) to investigate thermo-diffusive instabilities. Numerical simulations using detailed chemistry are performed where an initial perturbation is set to disturb the planar flame front and to compute its growth rate. During the initial linear phase, the perturbation's amplitude grows or decreases depending on the flame's mixture propensity to be unstable or stable, respectively. At small times, the linear growth rate of the perturbation can be estimated and compared to theory. As expected, the maximum growth rate obtained in the linear phase depends on the mixture's equivalence ratio, fuel ratio, and pressure. Ammonia addition leads to a reduced peak growth rate, likely due to lower reactivity leading to a higher Zel'dovich number. Very lean mixtures (equivalence ratios of 0.4 and 0.5) are thermo-diffusively unstable, regardless of the ammonia content, due to hydrogen's preferential diffusion. [ABSTRACT FROM AUTHOR]

This study investigates H 2 -air flames obtained with a laboratory scale coaxial dual-swirl injector in which fuel and oxidizer are injected separately. Two flame archetypes are observed experimentally for the same global equivalence ratio ϕ g ≈ 0.45 and two different thermal powers: a flame anchored to the injector (≈ 4 kW) and an aerodynamically stabilized flame exhibiting a characteristic V-shape (≈ 10 kW). Large Eddy Simulations (LES) allow to retrieve both regimes and are used to investigate these two stabilization modes. The numerical setup is first validated against isothermal and reactive Particle Image Velocimetry measurements and OH * chemiluminescence images. The mean velocity field of both operating conditions reveals the existence of a strong inner recirculation zone (IRZ) that, penetrating inside the injector nozzle, leads to a radial divergence of the central hydrogen jet, which ultimately favors one stabilization regime over the other. The first flame anchors on the hydrogen injector lip and it develops along the mixing layer between H 2 and air swirling jets. The lifted flame, instead, stabilizes in the inner shear layer between the IRZ and the exiting swirling jet of fresh gases, burning over a wide range of equivalence ratios. LES also unveil the flame structures typical of each flame: the anchored one is entirely controlled by diffusion, while the lifted flame is characterized by a first partially premixed branch and a second diffusion front. Finally, high-speed camera and LES are used to analyze the unsteady transition from lifted to anchored flames. [ABSTRACT FROM AUTHOR]

A methodology to build reduced schemes able to capture the main flame features and combustion mode in a sequential burner is proposed. This is here applied to derive a two-step scheme for kerosene flames at reheat conditions typical of aircraft engines for Large Eddy Simulations applications. It comprises 6 species (kerosene, O 2 , N 2 , CO 2 , CO , H 2 O) and two reactions where only two kinetic constants are adjusted: the pre-exponential constant of the reactions are changed as a function of the progress variable and equivalence ratio in order to match both auto-ignition times and premixed flame speeds in rich conditions. Vitiation levels are taken into account through the equivalence ratio ϕ p f characterizing the oxidizer stream composition in the reheat combustor. The novel two-step scheme is validated against semi-detailed chemistry for 1D laminar kerosene/vitiated air flames with decreasing inlet velocity (thereby decreasing reactant residence time) changing from auto-igniting flames at high inlet velocities to propagating flames flashing back at low inlet velocities. The scheme is proved to capture auto-igniting as well as propagating flames found in afterburner conditions at very low computational cost and with a precision comparable to complex chemical schemes. [ABSTRACT FROM AUTHOR]

• A temperature-induced supercritical bifurcation of combustion instabilities is found. • Kerosene-fuelled LPP combustor operating at elevated pressure and temperatures. • The phase shift between p′ and q′ is found to be the key reason. • Higher temperatures shorten the convection time and move the system toward stability. • Higher temperature helps fuel evaporation and reduces equivalence ratio fluctuation. The present article reports experimental observation and analyses of a supercritical bifurcation of combustion instabilities triggered by the air inlet temperature ( T a ). The studies are performed with a pressurised kerosene fuelled Lean Premixed Prevaporized (LPP) combustor operated under elevated temperature. Unlike some previous studies, starting from an unstable condition of the system, the amplitude of combustion instabilities suddenly decrease when T a exceeds a critical value of T a = 570 K. When the temperature is lowered back the system returns to being unstable without featuring any hysteresis behaviour, as expected in case of a supercritical bifurcation. The unstable flames feature a periodic axial motion of lift-off and re-ignition, characterized as Helmholtz mode. The phase difference between chemiluminescence and pressure signals is found to increase with T a , exceeding 90 degrees (out of phase) for temperatures higher than 570 K. A low order network framework is conducted, illustrating that when T a is increased a stability shift of this mode is predicted at T a near 570 K, in good agreement with the experimental observations. The impact of T a on the spray characteristics is also examined, finding that higher T a promotes fuel evaporation and reduces equivalence ratio fluctuation at the exit of the swirler. [ABSTRACT FROM AUTHOR]

An Arrhenius-based one-step scheme is derived for hydrogen-air combustion simulations. A Pre-Exponential Adjustment approach, based on an explicit analytical dependence of reaction rate parameters on equivalence ratio and pressure is adopted, together with a correction to improve the prediction of thermal flame thickness. The reduced scheme is validated by computations of one-dimensional unstrained and strained laminar premixed flames for a wide range of pressures ([1; 30] atm), unburned gas temperatures ([300; 800] K), and equivalence ratios ([0.4; 6.0]), with a good agreement of predicted main flame parameters between reduced and reference kinetic schemes. Coupled to a high-fidelity Navier–Stokes compressible solver, the reduced scheme is successfully proved for the numerical simulation of canonical configurations such as one-dimensional and two-dimensional premixed flames under several mixture conditions, with a significant improvement of computational efficiency. [Display omitted] • Derivation of a one-step scheme for hydrogen-air mixtures based on Arrhenius law. • Pre-Exponential Adjustment to match typical combustion systems' operative ranges. • Introduction of a correction to predict flame thickness in all mixture conditions. • Good agreement of reduced and reference schemes found for canonical configurations. • Sensible reduction (> 50%) of computational costs in CFD applications. [ABSTRACT FROM AUTHOR]

The Thickened Flame (TF) model is a widely used approach for Large Eddy Simulation of premixed flames. It is based on a "mapping" transformation where all diffusivities are multiplied by a thickening factor F while reaction terms are divided by F. Theory shows and 1D flame simulations confirm that this mapping preserves the unstretched laminar flame speed s L 0 while increasing its thickness δ L 0 by F , allowing to resolve the flame on a coarse grid. However this property is not satisfied anymore when the TF model is applied to stretched flames: the burning velocity s c (k) of a thickened flame submitted to a stretch k is not conserved compared to the non-thickened solution. A new diffusion-reaction transformation, named Stretched-Thickened Flame (S-TF) model, is developed here to conserve the burning velocity of stretched flames, proposing a generalization of the classical diffusion-reaction transformation which the TF model is based on. Thermal and mass diffusion zones are thickened differently and the laminar unstretched properties are preserved by modifying the chemical source terms. The S-TF model is applied on canonical 1D strained flames and validated for a cylindrical expanding flame configuration to prove its applicability to recover both strain and curvature effects. Results show that the Stretched-Thickened Flame model is an effective solution to correct the deficiency of the classical approach for stretched flames and can be easily implemented in CFD codes relying on the TF model. [ABSTRACT FROM AUTHOR]

Three turbulent premixed flames with the same unstretched laminar flame speeds and thicknesses are analyzed and compared using 3D Direct Numerical Simulation (DNS) in a slot burner configuration at atmospheric conditions: a CH 4 / air flame at ϕ = 1 , an NH 3 − H 2 / air (46% vol. H 2) at ϕ = 1 and an H 2 / air flame at ϕ = 0.45. While both stoichiometric methane and ammonia-hydrogen flames behave similarly, the lean hydrogen flame brush is twice as short with less flame surface, and exhibits significant alteration of its local flame structure: for the H 2 flame, the thickness of flamelets decreases significantly while burning rates increase drastically. This is observed for flame elements convexly curved with respect to fresh reactants (positive curvature) because of preferential diffusion, where thermo-diffusive instabilities generate long-tail structures that continuously head toward the fresh gases, and also for near-flat flame elements because of local strain effects. Opposite behaviors are observed for flame elements concavely curved (negative curvature) where fuel depletion caused by unfocusing of hydrogen and strain effects are unable to increase the consumption, even leading to near-extinction of the flame. The turbulent methane and ammonia-hydrogen flames studied in this work do not exhibit the instabilities seen in the pure hydrogen flame. However, local flame-flame interactions are observed in the cusp regions of all three flames, which significantly increase the displacement speed as well as flame surface destruction. A 1D-3D comparison indicates that strain effects in regions of low curvature and preferential diffusion effects in regions of positive curvature can be modeled using counterflow flamelets to account for stretch effects on the turbulent hydrogen flame propagation in addition to turbulent wrinkling. [ABSTRACT FROM AUTHOR]