Elsevier

Combustion and Flame

Volume 192, June 2018, Pages 410-425
Combustion and Flame

Dynamical response of a perfectly premixed flame and limit behavior for high power density systems

https://doi.org/10.1016/j.combustflame.2018.02.007Get rights and content

Abstract

The interaction between a perfectly premixed flame and an acoustic wave is studied in a 1D configuration. The mathematical framework to represent the flame dynamics is based on a recent work by Chen et al. that accounts for flame displacement in an acoustic field (Chen et al., 2016). An energetic approach is developed to estimate the source terms associated to flame/acoustic interactions. Three main contributions are identified: the flame velocity response, which corresponds to the flame response to upstream velocity perturbations (often modeled by an nτ model), the flame pressure response, related to the link between the reactant consumption rate and the local density, and the flame motion contribution, related to the displacement of the heat release rate in an acoustic pressure field. The energetic contributions are evaluated theoretically using different formalisms to make connections with previous work. A focus is made on the limit case corresponding to high frequency/high power density (HF/HPD) configurations. This asymptotic condition is of interest in some practical applications, and a remarkable expression is obtained for the flame contribution in that case. The acoustic source term is expressed simply by the integral of the product between the mean heat release rate and the acoustic energy. A parametric study is carried out on a 1D duct of constant section incorporating a perfectly premixed flame. The spatial structure and frequency of the first mode (which is studied exclusively here) are obtained using a classical decomposition of pressure, velocity and entropy in plane waves and expressing jump and boundary conditions on the amplitudes. The flame energetic contributions are compared to one another and to the growth rate obtained using the classical wave method. The different methods are validated on a wide range of parameters. The particular behavior of systems featuring a high power density is presented.

Introduction

Combustion instabilities are an important technical concern when designing industrial systems which incorporate combustion chambers. They often result from the coupling between the unsteady heat released by the flame and the acoustic modes of the combustor. Understanding the phenomena driving the dynamics and response of a flame submitted to acoustic perturbations is of theoretical and practical interest.

Chen et al. [1] provided recently an interesting description of the dynamical behavior of lean premixed flames. They formulated the discontinuity between upstream and downstream of the flame using the Rankine–Hugoniot relationships, introducing the flame velocity in the laboratory frame of reference us. This mathematical description provides a way to study two apparent problems that appear when the flame is considered immobile. The first one is that the mass flow rate is not conserved through an acoustic discontinuity. The second one is that velocity fluctuations upstream of the flame translates in significant entropy fluctuations downstream even for a perfectly premixed flame, which is not observed in experiments. Considering the fact that the flame position changes with time to adapt to upstream conditions allows to solve elegantly these conceptual issues. The impact of this displacement on the stability analysis of an actual system is not investigated in their study.

In parallel, an analysis of the impact of the motion of a flame on the acoustic source term is presented in [2]. An energetic approach is derived to estimate analytically the acoustic source term related exclusively to the flame displacement. One strong assumption of this study is that the flame follows exactly the acoustic velocity, which is expressed by the balance equation: q˙t+·q˙uf=0where q˙(x,t) is the local heat release rate and uf is the velocity of the flame in the laboratory frame of reference and chosen equal to the acoustic velocity. It is not true in general. This assumption was brought forth to isolate the contribution of the motion on the acoustic source term. In that case, it was proven that the contribution of the flame displacement to the acoustic balance equation was strictly positive and could become significant for systems with high power density. One reaches significant growth rate by rising the global heat release rate or by decreasing the dimensions of the system. This generally corresponds to an increase in the resonant frequencies of the system.

The first idea of the present work is to use the dynamical representation of the premixed flame provided in [1], apply it on a concrete case and estimate the different contributions, including the one due to the flame displacement discussed in [2]. As mentioned above, the displacement contribution becomes significant when the power density of the system increases. A particular focus is made to estimate the system behavior in the asymptotic condition corresponding to very high power density.

An energetic approach is chosen to account for the different contributions independently. There are two major contributors to the system stability: the coupling between flame and acoustics and the acoustic fluxes at the boundary conditions. In this study, this problem has been considered with much scrutiny because it appeared that some interesting features arise when the Mach number starts to grow. In addition, the formulation of the jump conditions by Chen et al. introduces an entropy term that is convected in the burnt gases. Therefore, an energetic formulation for the boundary contribution that includes entropy effect has been used [3].

The complex way premixed flames respond geometrically to acoustic perturbations is out of the scope of this paper. The G-Equation was used successfully in the past to model such phenomena [4], [5], [6], [7]. Instead, the flame response is captured here using an nτ model. The parameters n and τ are varied to study the influence on linear stability. Luzzato and Morgans [8] recently considered the effect of the flame motion on stability. The mathematical solution was obtained using the characteristic method in the temporal domain. This methodology provides interesting results, in particular because it gives access to non-linear features of the instability. Nevertheless the physical mechanisms which drive the impact on stability are not clearly identified and discussed. The study does not address the limit cycle for high-frequency modes and the role of entropy convection, which is investigated here. Using an energetic approach allows to provide a deeper physical understanding of why and in which context the implication of the motion is important.

The objectives of this article are the following:

  • Use the dynamical description of the flame proposed by Chen et al. [1] to model the displacement contribution in an actual simple system.

  • Introduce an energetic approach representing independently the different contributions from the flame and boundaries.

  • Propose an asymptotic modeling of the flame contributions when the resonant frequencies and power density of the system increase.

  • Present a parametric study to demonstrate the validity of the models and illustrate the influence of some key parameters: power density, Mach number, flame response, size of the cavity among others.

In Section 2, the formulation of the flame dynamics introduced in [1] is reviewed. Section 3 provides the various energetic contributions that will be compared in the parametric study. Section 3.1 introduces αRay, which accounts for the complete flame contribution, motion and flame response. Section 3.2 reminds the E-FAME model of [2]. In Section 3.3, the flame response is decomposed in two terms, a flame velocity response linked to velocity perturbations upstream of the flame and a flame pressure response. In Section 3.4, an asymptotic model is proposed for high frequency/high power density conditions. The boundary conditions contributions are estimated in Section 3.5. Section 4 describes the classical method to obtain the wave structure in the system. It also provides a reference growth rate which is used to compare to the energetic approach. Finally, Section 5 presents the results of the parametric study.

Section snippets

Formulation of the perfectly premixed flame problem

The idea of a moving discontinuity to represent a heat source has been introduced by Chu [9], [10]. Subsequently, this idea was used in various studies ([11], [12], [13] among others). The particular feature of Chen et al. approach [1] is to link the entropy creation at the flame location and its velocity in the laboratory frame of reference. Using this formalism allows to provide meaningful explanations for apparent issues discussed previously in the literature, as already mentioned in the

Energetic approach

The energetic approach was used quite early in the development of methodologies to study the acoustic stability of systems [17]. For example, Cantrell and Hart [18] provided a rigorous formalism to estimate the impact of the boundary fluxes on the acoustic stability of the system (see Section 3.5 for more details on this subject). A useful introduction to this method, as well as the basis relationships between acoustic fluxes, combustion/acoustic coupling terms and growth of the acoustic energy

Derivation of the plane wave solution

The geometry is represented in Fig. 2. Two cavities are considered, with a flame located at the interface.

In this section, one assumes that the acoustic and entropic variables behave like 1D plane waves, which is a common assumption when studying simple geometries. The principle is to use the boundary conditions and jump conditions at the flame location to express linear relationships between the waves amplitudes. This provides two important results. First, the complex frequency, which

Results of the parametric study

The purpose of this section is to implement the different growth rates expressed so far and to calculate them in a representative case which can be related to practical applications. Various parameters are modified to illustrate the impact on the stability of the system.

A few growth rate expressions have been introduced along this document. Their purpose is to identify accurately the contributions and their physical origin. The scheme presented in Fig. 3 aims at showing the link between the

Conclusion

This article focuses on the modeling of the response of a perfectly premixed flame in a 1D cavity. The modeling is based on the set of equations provided by Chen et al. [1] who described the response and dynamics of a lean premixed infinitely thin flame in a quasi-1D flow configuration. The flame described by these equations is included in a simple acoustic element and the different acoustic contributors are carefully evaluated.

The flame contribution to the acoustic behavior of the system is

Acknowledgments

We thank the reviewers of this article for the fruitful discussions throughout the reviewing process, which have significantly contributed to increasing the quality of the paper.

References (26)

  • C.M. Luzzato et al.

    The effect of a laminar moving flame front on thermoacoustic oscillations of an anchored ducted v-flame

    Combust. Sci. Technol.

    (2015)
  • B.-T. Chu, Stability of systems containing a heat source: the Rayleigh criterion...
  • A. Dowling et al.

    Reheat buzz – an acoustically driven combusion instability

    AIAA 9th Aeroacoustics Conference

    (1984)
  • Cited by (10)

    • Effect of the flame motion on azimuthal combustion instabilities

      2022, Aerospace Science and Technology
      Citation Excerpt :

      Based on the research of Chen et al. [22], the mechanism of entropy pulsation produced by the flame motion with the change of cross section was further clarified by Chen et al. [23]. Méry [24] found that the increase of power density would lead to the increase of contribution of flame pressure and flame motion. In addition, based on linearized Navier-Stokes equation and local flame transfer function, Meindl et al. [25] proposed a method for analyzing combustion instability, which overcomed the difficulty of directly coupling the flame velocity of two- or three-dimensional flame with Navier-Stokes equations.

    • Effect of flame motion on longitudinal combustion instabilities

      2022, Aerospace Science and Technology
      Citation Excerpt :

      Based on the research of Chen et al. [23], the mechanism of entropy pulsation produced by the flame motion was explored by a number of researchers. Méry [24] categorized the influence of flame on system acoustics to three parts: the flame motion, the flame velocity and the flame pressure. Méry also constructed a high frequency model in which only the flame pressure and flame motion responses were considered.

    • Thermoacoustic analysis of a laminar premixed flame using a linearized reactive flow solver

      2019, Proceedings of the Combustion Institute
      Citation Excerpt :

      On the other hand, the LRF approach does not require an external FTF, so an unsteady CFD simulation for the identification of the FTF is not required. The local flow-flame interaction resolved by the LRF is particularly important for the investigation of non-compact flames and/or non-plane acoustics, i.e. at higher frequencies, where a standard FTF that relates upstream velocity to overall heat release is inadequate [24–26]. Moreover, configurations where fluctuations in equivalence ratio, pressure or temperature perturb the heat release require corresponding MISO or MIMO (multiple-input, single/multiple-output) formulations of hybrid models.

    View all citing articles on Scopus
    View full text