Sound generation by premixed flame annihilation with full and simple chemistry

https://doi.org/10.1016/j.proci.2014.05.140Get rights and content

Abstract

This paper presents a numerical study of sound generation by premixed laminar flame annihilation with full and simple chemistry. Planar annihilation is examined from lean (ϕ=0.5) to rich (ϕ=3) conditions, which correspond to (effective) Lewis numbers ranging from 0.5 to 2.0 respectively. The full chemistry simulations use a well known detailed mechanism for hydrogen-air mixtures [1,2]. The simple chemistry simulations are tuned such that the freely propagating laminar flame speed matched those from the detailed chemistry simulations at the same equivalence ratio.

These simulations examine two, related issues. First, they are used to confirm whether the Lewis number plays a significant role in the sound generation process, as our previous, simple chemistry studies have found [3,4]. More generally, these simulations seek to determine whether the use of simple chemistry to model sound generation is reasonable. The importance of the Lewis number is confirmed, and the use of simple chemistry is found to be reasonable in some cases.

Introduction

The modelling of combustion generated sound continues to be an active topic of research in our community. This is primarily because the lean, premixed combustion of natural gas is now commonly used in industry and power generation to achieve acceptably low emissions of the oxides of nitrogen (NOx). This, in turn, makes the stable operation of the combustor challenging, in large part due to the sensitivity of premixed flames to acoustic disturbances. This sensitivity is manifest by several mechanisms [5], [6], [7], [8], [9]. As emissions regulations tighten, thermoacoustic stability becomes a greater challenge since further reductions in equivalence ratio reduce flame stability in step with reduced NOx emissions.

Such issues are not solely the concern of natural gas fuelled devices. For example, proposed gas turbine systems designed for carbon capture and storage (CCS) typically run on either hydrogen or hydrogen-rich synthesis gas [10]. Once again, most are proposed to operate in premixed mode in order for these systems to achieve acceptable NOx emissions. Thus, the importance of modelling combustion generated sound is likely to remain if CCS becomes implemented.

The reliable modelling of sound generation by premixed flames nonetheless remains a challenge. We have argued in previous works that this is in part due to an incomplete understanding of the physics involved, as well as the computational difficulties in simultaneously modelling the flame motion and acoustic field, the latter particularly for problems of practical interest [3], [4], [11], [12]. As such, we have focused our efforts on simplified one and two dimensional problems in which both the flame and the sound it generates are resolved, thereby permitting closer interrogation.

Our results from these studies were consistent with other work in the literature in identifying the importance of so-called ‘annihilation events’  as sources of sound, e.g. [13], [14], [15], [16]. These events occur when two flame surfaces interact, resulting in a rapid consumption of the unburnt premixture between them and then a rapid change in overall heat release as the flames finally extinguish.

In studying these annihilation events, we assumed simple chemistry and, amongst other things, found that the Lewis number had a significant effect on the amplitude of the sound generated [3], [4]. These effects were explained by reference to variations in the fuel consumption and displacement speeds during the annihilation event; variations also observed by others for non-acoustic studies of premixed flame annihilation, e.g. [17]. However, the observed importance of the Lewis number perhaps calls into question the validity of using simple chemistry for studying sound generation by flame annihilation. Since acoustic energy is so much smaller than the chemical energy released by combustion, hydrodynamically insignificant errors in modelling the annihilation event might be acoustically significant.

This paper therefore presents a comparison of the sound generated by premixed flame annihilation with full and simple chemistry. Since the Lewis number was observed previously to have a quantitatively similar effect in planar, axisymmetric and spherically symmetric annihilation events, we consider only planar annihilation in this paper. These events are examined from lean (ϕ=0.5) to rich (ϕ=3) conditions, which correspond to (effective) Lewis numbers ranging from 0.5 to 2.0 respectively.

Section snippets

Numerical set-up

The configuration studied consists of two planar, identical flames travelling towards one other. These flames annihilate after consuming the reactants in between them. The problem of sound generation in this configuration was already studied in [3], [11] using one-step chemical kinetics. In the present work we simulate the problem using the code NTMIX-CHEMKIN, a high-order accurate flow solver that is designed to perform direct simulations of flames with detailed chemistry [18], [19].

Detailed chemistry simulation results

Figure 1 illustrates the simulated annihilation and sound generation process for a lean (ϕ=0.5), a stoichiometric (ϕ=1) and a rich (ϕ=3) flame. Only half of the computational domain is shown, with x=0 being the symmetry axis. Moreover, only a small portion of the domain is shown, to focus on the flame behaviour near the origin. We plot in Fig. 1 the evolution of the heat release rate profile Q̇ normalized by its maximum in the freely propagating flame (Q̇max), the evolution of the reduced

Conclusions

This paper presented a numerical study of sound generation by premixed laminar flame annihilation with full and simple chemistry. Planar annihilation is examined from lean (ϕ=0.5) to rich (ϕ=3) conditions, which correspond to (effective) Lewis numbers ranging from 0.5 to 2.0 respectively. The full chemistry simulations used a well known detailed mechanism for hydrogen-air mixtures [1], [2]. The simple chemistry simulations were tuned such that the freely propagating laminar flame speed matched

Acknowledgements

The first author acknowledges the support of Spanish MCINN under projects #ENE2011-27686-715 C02-01 and #CSD2011-0001 and the Comunidad de Madrid under project #S2009ENE-1597. The other authors wish to thank the Australian Research Council (ARC). Use of the ‘Edward’ (the University of Melbourne’s ITC cluster) is also acknowledged.

References (28)

  • P. Saxena et al.

    Combust. Flame

    (2006)
  • A. Sánchez et al.

    Prog. Energy Combust. Science

    (2014)
  • M. Talei et al.

    Combust. Flame

    (2012)
  • M. Talei et al.

    Proc. Combust. Inst.

    (2013)
  • S. Candel

    Proc. Combust. Inst.

    (2002)
  • J. Gibbins et al.

    Energ. Policy

    (2008)
  • D. Durox et al.

    Combust. Flame

    (2009)
  • T. Echekki et al.

    Proc. Combust. Inst.

    (1996)
  • T. Poinsot et al.

    J. Comp. Phys.

    (1992)
  • J. Bechtold et al.

    Combust. Flame

    (2001)
  • E. Fernández-Tarrazo et al.

    Combust. Flame

    (2006)
  • I. Wichman et al.

    Combust. Flame

    (1997)
  • G. Markstein

    Non-steady Flame Propagation

    (1964)
  • G. Searby et al.

    J. Fluid Mech.

    (1991)
  • Cited by (15)

    • A comparative study of flame-wall interaction and flame-cooling air interaction

      2021, International Journal of Heat and Fluid Flow
      Citation Excerpt :

      The CO characteristics are explored in Section 3.3 and the proposed CO model is assessed in Section 3.4. The simulations were performed using the DNS code for reacting flows, NTMIX-CHEMKIN (Baum et al., 1994; Jiang et al., 2019; Jimenez et al., 2002; Jiménez et al., 2015; Jiménez and Kurdyumov, 2017; Palulli et al., 2019; Rivera et al., 2019; Palulli et al., 2021; Brouzet et al., 2021; Jiang et al., 2021). This code is a fully parallelised, high order accurate solver.

    • Reduced chemistry for sound generation by planar annihilation in premixed methane/hydrogen flames

      2021, Proceedings of the Combustion Institute
      Citation Excerpt :

      The performance of the reduced mechanism for prediction of sound generation by flame annihilation is then assessed. The same group also showed that the Lewis number can affect the time scale associated with an annihilation event and therefore the corresponding time scale of the generated pressure wave [15,16]. For Le ≥ 1, flame acceleration was observed during annihilation, which led to a pressure wave in the form of a step function.

    • Exhaust CO emissions of a laminar premixed propane–air flame interacting with cold gas jets

      2019, Combustion and Flame
      Citation Excerpt :

      It uses an 8th order explicit central differencing scheme and a 3rd order Runge–Kutta integrator for spatial and time derivatives, respectively [24]. Species production and molecular transport terms are solved using the CHEMKIN and TRANSPORT packages [25]. Diffusion velocities are modelled using mixture-based species-specific diffusivities, calculated from temperature-based binary diffusivities [26].

    • Unsteady flame–wall interaction: Impact on CO emission and wall heat flux

      2019, Combustion and Flame
      Citation Excerpt :

      The DNS code known as NTMIX-CHEMKIN was used to conduct the simulations. NTMIX-CHEMKIN [22] is a fully parallelised, high-order accurate solver that has been frequently used to perform DNS of combusting flows, e.g. [18,23–25]. It features a sixth-order compact scheme for spatial derivatives, combined with a third-order Runge–Kutta time integrator.

    View all citing articles on Scopus
    View full text