Article:Sharpe, G.J. (2009)
Abstract.A linear stability of freely propagating, adiabatic premixed flames is investigated in the context of a thermal-diffusive or constant density model, together with a simple two-step chainbranching model of the chemistry. This study considers the case of realistic, finite activation energy of the chain-branching step, and emphasis is on comparing with previous high activation energy asymptotic results. It is found that for realistic activation energies, a pulsating instability is absent in regimes predicted to be unstable by the asymptotic analysis. For the cellular instability, however, the finite activation energy results are in qualitative agreement with the asymptotic results, in that the flame is unstable only below a critical Lewis number of the fuel and becomes more unstable as the Lewis number is decreased. However, it is shown that very high activation energies would be required for the asymptotic analysis to be quantitatively predictive. The flame is less unstable for finite activation energies than predicted by the asymptotic analysis, in that a lower fuel Lewis number is required for instability. It is also shown that the flame structure and stability can have nonmonotonic dependencies on the activation energy. 1. Introduction. A premixed flame is a subsonic combustion wave which propagates via diffusion of chemical species and conduction of heat between the hot burnt chemical products and the cold unburnt fuel. While, in theory, unconfined premixed flames can propagate steadily as planar waves, this may not be realized in practice due to cellular and pulsating instabilities induced by thermal-diffusive and/or hydrodynamic effects [1]. A linear stability analysis of the underlying planar wave is a first step towards understanding the origins of such instabilities and predicting the conditions under which they occur.The majority of flame stability developments have employed a standard one-step chemistry combustion model, with a single, exothermic reaction step F→P, where F denotes fuel and P products, together with an Arrhenius form of the reaction rate. In this context, Sivashinsky [2] used a constant density model (CDM) and employed a high activation energy asymptotic (HAEA) limit in order to investigate the linear stability of one-step chemistry flames. The CDM ignores hydrodynamic effects but has extensively been demonstrated to correctly capture thermal-diffusive effects on flame dynamics [3]. Sivashinsky [2] showed that below a critical Lewis number, Le (the ratio of temperature conductivity to molecular diffusivity), the flame is unstable to a cellular instability. This cellular mode corresponds to a positive real linear eigenvalue in the stability problem. This critical value of Le is less than one, but tends to unity from below as the activation energy of the reaction is increased. Sivashinsky further showed that, above a second critical Lewis number, the flame is unstable to a