Experimental study of the lifting characteristics of the leading-edge of an attached non-premixed jet-flame: Air-side or fuel-side dilution

Abstract

The impact of air-side and methane-side dilution (CO₂, N₂ and Ar) on the lifting process of attached non-premixed methane/air coaxial jet flames is studied over a wide range of aerodynamic conditions. The study of the competition between aerodynamics and dilution has allowed to discriminate and quantify the different phenomena involved in the lifting process. First, two effects, only dependent on the amount of the added diluent, contribute to promoting flame detachment (liftoff): a fluid mechanical effect that causes the bulk velocity of the reactants to increase; a mixing effect that changes the mixture fraction spatial distribution. They are significant for the methane-side dilution but negligible for the air-side dilution. Then, the main mechanism and the dimensionless numbers characterizing flame liftoff are identified. The attached-flame stability is analyzed on the basis of the lifting limits measured by the critical flow rate ratios, (Q_d/Q_f)_lift and (Q_d/Q_ox)_lift when the diluent is added either to the methane or the air stream. These limits follow self-similarity relationships based on the fuel and oxidant Peclet numbers of the diluted streams, satisfied whatever the diluents. Results using PIV and CH^* measurements are interpreted through a flame-leading-edge approach, where CH₄/air/diluent are mixed locally at the flame base. The flame propagation velocity S_L which balances the incoming gas velocity, is shown to be described by self-similarity relationships based on the molar fraction at the leading-edge reduced by values at liftoff, $X^d/X_{lift}^d$. To confirm the leading-edge propagation characteristics, the flame attachment height H_a and radius R_a are investigated at the attached-flame base. R_a is representative of the mixing and mass effects induced by the pure dilution. Contrary to R_a, H_a is piloted by S_L, and evolves according to a unique law dictated by $X^d/X_{lift}^d$. $X_{lift}^d$ is a self-similar parameter highlighting the propagation nature of the leading-edge.

Introduction

Present and future technological designs commonly rely on combustion processes where oxidant and fuel are injected separately. In such a non-premixed regime, reactant mixing is a key element in several basic phenomena, among which flame stabilization (liftoff, blowout and blowoff ) and pollutant formation (CO, NO_x, soot) are important. As shown in previous studies (e.g., [1], [2], [3]), these two processes are very sensitive to reactant dilution as well as to aerodynamic conditions. To ensure flame stability, it is important to predict the operation limits.

In a co-flowing configuration, the state of the jet-flame depends on: (i) jet and co-flow conditions: fuel and oxidant velocities, reactant composition, stream temperatures (ii) burner characteristics: rim e_l, diameter D_i. Both attached and lifted states have been studied by several research groups in order to clarify the mechanisms on which they rely [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], in particular by measuring the vertical distance H between the burner exit and the flame base. For a lifted jet-flame, the distance H ≡ H_L, dependent on the jet and coflow characteristics (velocity field, species concentration), is quite large: 2 D_i < H_L < 60 D_i, before extinction occurs (see [13], [14], [15], [22], [23], [24] ). This helps to adequately mix air and fuel, at least partially. A flame attached to the burner, however, presents only a short distance, $H\equiv H_a\sim 0.1-0.7e_l$ where mixing can occur (see [1], [12], [25], [26], [27], [28], [29]). This short distance contributes to the difficulty of measuring the quantities that characterize the structure and location of the flame base.

Some experimental investigations, essential to quantify and understand how the rim influences attached flame stability, have been carried out ([12], [26], [27], [28], [29], [30], [31]). Juniper and Candel [27] carried out a numerical simulation to study the stability of flames resulting from coaxial gaseous hydrogen and liquid oxygen jets. They examined the influence of the Damköler number Da and of the step-height e_l behind which the flame stabilized. If the flame thickness δ_fl was less than e_l, the flame tucked behind the step within a slow flow region. The flame was then very little affected by Da variations. By contrast, if δ_fl was thicker than e_l, the flame was forced out and the standoff distance increased; the flame was very sensitive to Da variations. Y. Otakayema et al. [26] studied the influence of the rim on the stabilization mechanism of non-premixed N₂ – CH₄/air jet-flames issuing from rectangular slits of thickness e_l. The distance H was studied as a function of the air coflow velocity, U_air, for various rim thicknesses e_l. The methane mole fraction and fuel (${\text{CH}}_4+{\text{N}}_2$) velocity were fixed at 0.35 and 0.8 m/s. For thin rims (e_l ≤ 2.0 mm), H gradually increased from H_a to H_L with increasing U_air before blowout occurred under the same conditions ($U_{air,out}=0.28\text{m/s}$ and H_{L, out} ≈ 20 mm) regardless of e_l. For thick rims (e_l ≥ 3.0 mm), H remained at about $H_a=1\text{mm}$ as U_air was increased; flame blew off suddenly without passing via a lifted state. Finally, for “intermediate” rims (2.0 < e_l < 3.0 mm), the attached flame base suddenly jumped to H_L ∼ 7 mm once U_air reached 0.19 m/s, leading to a lifted flame. A clear difference in flame stability was then revealed between the thin and thick rims. The material of the rim was also expected to play some role in heat transfer between the burner and the flame. Depending on its thermal properties, a material can influence the temperature of the rim, but no critical T_rim was found that characterized attached flame liftoff (see [28], [32], [33]). For low thermal conductivity, a higher T_{rim, lift} was noted leading to a slight decrease in the liftoff velocity, U_{f, lift}. Furthermore, the flame stabilization location, specified by the flame attachment radius and height (R_a, H_a), presented the same evolution as the methane-jet velocity U_CH4 was increased, with no significant differences, provided that the thermal conductivity of the burner materials was high enough to avoid heat loss associated with thermal and chemical quenching [28].

The recent increasing interest in alternative fuels (such as syngas, biogas), which are composed of up to 40% of inert gases, calls for better knowledge of the gas-mix properties in order to control their use in energy systems, in particular when new modes of combustion (e.g., massive exhaust gas recirculation systems, oxy-fuel combustion, fire safety, flameless combustion regimes) are involved [34], [35], [36], [37]. To attain this aim, both air-side dilution (characterized by $Z_{st}<Z_{st}^o$ where Z_st is the mixture fraction at stoichiometry and exponent o means without dilution) and fuel-side dilution ($Z_{st}>Z_{st}^o$), must be studied when an extra inert or chemically weak diluent is added. Takahashi et al. [2] examined the extinction mechanism of attached flames by diluting the coflowing air in very low methane-stream-velocity configurations ($U_{\text{CH}4}=0.0092\text{m/s}$ and $U_{ox}=0.107\text{m/s}$) experimentally and by simulations. They concluded that the relative ranking of the oxidant’s molar heat capacity indicated the relative ability of inert gases to break the attached-flame stability. Their numerical results noted the importance of the downstream trailing diffusion flame in supporting flame stabilization, which ultimately occurred at the leading-edge [38]. Flame responses in terms of liftoff and blowout were studied by diluting the combustible (ethane, ethylene, acetylene, and propane) or the air in the low aerodynamic conditions cited above and in a low gravity environment [2], [39]. Min et al. [1], [23], [40] investigated how the dilution of air (with CO₂, N₂, and Ar) influenced lifting of attached non-premixed methane-jet/air-coflow flames for a large range of aerodynamic conditions ($U_{\text{CH}4}=0.15-15\text{m/s},$$U_{air}=0.1-0.7\text{m/s}$). Diluent addition greatly reduced flame stability through pure dilution, thermal, transport properties and chemical effects, in decreasing order of importance. Moreover, they noted that the flame propagation velocity S_L, calculated with the diluent amounts measured when the non-premixed flame was detached, was the same for the three diluents [23]. Using several experimental quantities analyzed from the evolution of S_L assessed with dilution, they were able to explain that stabilization at the flame base, where methane, air and diluent were locally mixed, was mainly piloted by the flame-leading-edge mechanism, ensuring a balance between S_L and the incoming gas velocity. Kim et al. [22] investigated laminar propane attached flames, highly diluted with nitrogen in a fixed air coflow configuration ($U_{air}=0.5\text{m/s}$). The coflow temperature was varied. At the ambient condition $T=294\text{K},$ and for a propane mole fraction of 0.2, the fuel jet velocity at liftoff U_{f, lift} was observed to be smaller than the diluted stoichiometric laminar flame propagation velocity S_{L, st} ∼ 0.27  m/s. This could be expected since U_air was higher than S_{L, st}; the flame base then had to find a new position in the shear layer between the two streams (fuel and air) to ensure the balance between S_{L, st} and the incoming gas velocity.

Therefore, in order to understand the mechanism of a non-premixed flame’s leading-edge for which the local mixing is characterized by a (partially) premixed condition, it is essential to take into account how dilution acts on the flame propagation velocity, S_L. In premixing configurations, the addition of an inert or a chemically weak diluent (N₂, CO₂, He and Ar), induces a decrease in S_L, as was shown numerically and experimentally by [41], [42], [43], [44]. Dilution diminishes S_L with the following effectiveness, ordered from the weakest to the strongest diluent: He < Ar < N₂ < CO₂. Qiao et al. [41] and B. Galmiche et al. [42] indicated that N₂, Ar and He reduced S_L mainly by a thermal effect. But for CO₂, in addition to physical effects, they noted that some chemical effects were also involved, and the greater the CO₂ concentration, the greater the effects. The decrease of the flame speed by the chemical action was explained by Qiao et al. [41] via the reaction ${\text{CO}}_2+\text{H}=\text{CO}+\text{OH},$ which consumed the radical H necessary for the main chain-branching reaction $\text{H}+{\text{O}}_2=\text{O}+\text{OH}$ to dissociate O₂. The concentrations of radicals O, H and OH were reduced, leading to a lower reaction rate and flame speed. This is consistent with other studies that focused on the chemical effects of CO₂ addition that affected phenomena such as flame stability [40], soot formation [45], [46], flame structure and NO emission [47]. The decrease of S_L after a diluent addition evolved in the same way as the diluent’s ability to break the anchored non-premixed flame stability. This agreed with results previously obtained by Guo et al. [40] in the case of an air-coflow dilution. There, N₂ addition was shown to affect non-premixed flame behavior through pure dilution only; with Ar the pure dilution effect was countered by thermal and some non-negligible transport effects; CO₂ addition induced some supplementary chemical effects, but they had less impact than the pure dilution and thermal effects.

In such a framework, while most studies have involved a small range of aerodynamic conditions, where only one of the two reactant streams was diluted [2], [13], [48], [49], [50], [51], the present work analyzes the lifting process of an attached flame induced by diluting the two reactant streams (air coflow or methane jet), where the burner rim is an “intermediate” rim according to the Otakayema’s classification. The investigation is based on numerous experiments chosen by combining the following conditions: (i) CO₂, N₂ or Ar used as a diluent d; (ii) air-stream or methane-stream diluted by one of the three diluents; (iii) a couple of air and methane velocities (U_air, U_CH4) with 0.5 m/s ≤ U_CH4 < 13.3 m/s and 0.1 m/s ≤ U_air ≤ 0.4 m/s, covering the entire zone of the flame hysteresis domain [15] from the laminar to the turbulent regime. The experimental methodology allows to systematically analyze the influence of both phenomena, dilution and aerodynamics of the reactants (fuel and oxidant), to produce understanding of how they compete or combine in destabilizing the attached flame. Results obtained in these two dilution configurations are compared in order to identify the main mechanisms and dimensionless numbers able to describe this process and its limit (liftoff) when an inert or chemically weak diluent¹d is added. Here, “liftoff” means “lifting limit”, characterizing the jump of the flame position between its attached state, for which the flame stabilizes in the near vicinity of the rim, and its lifted state situated (relatively) far above the burner or its blowoff [30]. It corresponds to the last location where the flame is considered as attached to the burner. A fluid mechanical impact, due to the augmentation of the fuel, f, or oxidant, ox, velocities after dilution, is identified and quantified. It participates in stabilization modifications and stability loss when methane is diluted, but remains negligible with air-side dilution. A unified behavior describing flame liftoff is proposed for all the diluents through self-similarity laws which depend on two crucial dimensionless parameters: the fuel Peclet number Pe_f and the K_d parameter. Pe_f, usually used in diffusion flame studies [52], [53], is shown, for a given diluent, to be the dimensionless number characterizing flame stability at liftoff, as was suggested by Lawn [17]. K_d is the self-similar dilution quantity by means of which data, obtained with the different diluents, merge on a single curve. By introducing K_d, the flame-lifting limits for the diluents can be simply deduced from those defined for CO₂. The laws, satisfied for the “intermediate rims” , are valid whatever the aerodynamic conditions. Subsequently, a flame-leading-edge approach is proposed, where gases CH₄/air/diluent are locally mixed at the stoichiometric ratios. Particle image velocimetry (PIV) and CH^* emission images are used to characterize the velocity field and the attached-flame location (attachment height H_a and radius R_a) at the flame base in the vicinity of the leading-edge. The results prove the importance of the flame-leading-edge mechanism in supporting flame stability. In particular, they highlight how H_a is driven by the diluted flame propagation velocity S_L, while R_a is controlled by pure dilution. The fluid mechanical impacts which affect flame stability when dilution is performed in CH₄, are also taken into account in the law describing H_a as a function of S_L, while their introduction in the law governing the diluted air configuration is not required. Therefore the present work brings an original contribution able to characterize some of the crucial mechanisms involved in the new modes of combustion mentioned above for which dilution plays an essential role.