Investigation of the effects of the stratification on local curvature and flame surface density on grid generated turbulent flames P. Anselmo Filho ∗ , C. N. Markides and S. Hochgreb Cambridge University Engineering Department Abstract There is little understanding of the extent to which models developed for premixed and non-premixed gases are applicable to stratified situations. Recent Direct Numerical Simulation studies on the subject have started to shed light on the matter. However, there is no current experimental database against which to test models and hypotheses. One of the interesting questions is to what extent local stratification may affect flame wrinkling and curvature. This paper investigates the effects of reactant stratification on 2D flame geometric properties via simultaneous Planar Laser Induced Fluorescence (PLIF) of OH and acetone. OH PLIF is a well established technique for marking the flame front, and acetone PLIF marks the local mixture fraction at the flame. Simultaneous OH/acetone images allow the examination and comparison of both mapping techniques for local flame curvature and surface flame density as a function of stratification and turbulence properties. This study shows that acetone can be used as flame front marker for qualitative purposes under stratified conditions. The results for probability density functions of local curvature and flame surface density are analyzed using both techniques for fully premixed and stratified cases. ∗ Corresponding author: pa264@cam.ac.uk Proceedings of the European Combustion Meeting 2007 Introduction Premixed lean combustion is desirable due to low emission levels of NO x , unburned hydrocarbons, and particulate matter produced, but it can present flame instability and poor ignitability, particularly close to flammability limits. Globally lean combustion can benefit from the presence of a richer mixture zones to provide better ignitability or stability, creating a stratified mixture. In some other situations, such as aeroengine combustors, premixing is not practically achievable within the reduced mixing times available. Despite the practical applications of stratified combustion [1-4], there is little data on how stratified turbulent flames behave. The use of models for pure premixed or non-premixed cases in imperfectly mixed conditions has not been validated. Consequently, it is necessary to develop an experimental database in which the mixture fraction and turbulence, and the resulting flame propagation are well characterised [5, 6]. As a result of the growing interest in this matter, many experimental studies have started to investigate stratified combustion [7-10]. Those studies range from free flame propagation [7, 8] to stabilised V-flames [9, 10] using combined laser techniques, i.e. Planar Laser Induced Fluorescence (PLIF), for scalar and flame front tracking. In order to understand the effect of stratification on turbulent flame propagation, it is necessary to measure the stratification. A well-developed method of tracking fuel concentrations is via the use of a fluorescing tracer. The use of fluorescent tracers in reactive mixtures has a number of advantages and pitfalls, and the reader is referred to the literature for details [11]. Acetone has been extensively used to mark fuel concentrations in both non-reacting and reacting mixtures [7, 9, 11, and 12]. It is known that acetone decomposes somewhat earlier in the flame than methane [9], so its use as a reactive marker needs to be regarded with care. Acetone decomposes earlier than methane in the flame, so it has is not generally been considered suitable for marking the flame front accurately. On the other hand, OH PLIF has been used extensively for planar characterisation of flames, as well as flame front marker for turbulent flows [13-16]. OH radical when excited near 283 nm, fluoresces around 300 nm and present a strong signal- to-noise ratio (SNR) [14, 16, 17]. Here we compare the use of simultaneous OH and acetone PLIF to understand under what conditions acetone might be used as a single marker for mixture stratification and flame position. The simultaneous measurement also allows the characterization of the gap between acetone and OH decomposition, as well as two- dimensional curvature (κ) and flame surface density (Σ) in a dynamic turbulent flame. Specific Objectives The primary objective of this study is to investigate how stratification affects flame front curvature and flame surface density, in a rod-stabilised V-flame, as function of local mixture fraction. The work also compares the accuracy of acetone PLIF to OH PLIF in mapping flame geometry. Experimental Setup A two-dimensional slot burner is employed in this investigation. The burner comprises of 6 slots, where each slot is 5 mm wide by 50 mm length and walls between slots have a thickness of 0.3 mm. The mixture is independently controlled for the two inner pairs of symmetric slots, while shielding air flows through the outer slots. Air was filtered by high efficiency particulate air (HEPA) standards and injected separately with fuel in the burner’s mixing chamber (Fig. 1). The burner has two grids. One grid is located immediately THIRD EUROPEAN COMBUSTION MEETING ECM 2007