Analysis of Synthetic Flames

Stephanie A. Coronel,Rémy Mével,Joseph E. Shepherd

2019-01-01

Abstract:Laminar flame properties such as the laminar burning speed and the Markstein length are important fundamental parameters for a wide number of combustion applications. The laminar burning speed is defined as the normal propagation velocity of fresh gas relative to a fixed, planar flame front; it is frequently measured experimentally using spherically expanding flames [1]. The presence of flame stretch in such experiments precludes direct measurement of the laminar flame speed [2]. Instead, the measured flame speed has to be extrapolated to conditions of zero stretch. [3] first proposed this correction to the burning speed by introducing a parameter known as the Markstein length which characterizes the response of the flame to stretch. Asymptotic theoretical analysis [4, 5, 6] performed in the limit of high activation energy and low stretch rate have related the stretched and unstretched burning speeds through a linear relationship. Further theoretical work by [4] has led to a nonlinear relationship between the stretched and the unstretched burning speed which has been used by a number of groups in the past few years to account for nonlinear effects of stretch on the flame propagation. Comparison of the results obtained through linear and nonlinear extrapolations demonstrated that both the burning speed and Markstein length can be poorly estimated by the linear method for mixtures that are away from stoichiometry [7, 8]. The present work discusses the performance of nonlinear fitting methods by extracting the laminar flame properties from synthetic data sets. The sensitivity of the results to experimental parameters like initial and final flame radius, the number of points in the data set, and measurement noise, as well as numerical parameters like the initial guess that is used to start the nonlinear fit, are investigated.